Paul Boddie's Free Software-related blog


Archive for the ‘public sector’ Category

Pessimistic perspectives on technological sustainability

Tuesday, August 16th, 2022

I was recently perusing the Retro Computing Forum when I stumbled across a mention of Collapse OS. If your anxiety levels have not already been maxed out during the last few years of climate breakdown, psychological warfare, pandemic, and actual warmongering, accompanied by supply chain breakdowns, initially in technology and exacerbated by overconsumption and spivcoin, now also in commodities and exacerbated by many of those other factors (particularly the warmongering), then perhaps focusing on societal and civilisational collapse isn’t going to improve your mood or your outlook. Unusually, then, after my last, rather negative post on such topics, may I be the one to introduce some constructive input and perhaps even some slight optimism?

If I understand the motivations behind Collapse OS correctly, it is meant to provide a modest computing environment that can work on well-understood, commonplace, easily repaired and readily sourced hardware, with the software providing the environment itself being maintainable on the target hardware, as opposed to being cross-built on more powerful hardware and then deployed to simpler, less capable hardware. The envisaged scenario for its adoption is a world where powerful new hardware is no longer produced or readily available and where people must scavenge and “make do” with the hardware already produced. Although civilisation may have brought about its own collapse, the consolation is that so much hardware will have been strewn across the planet for a variety of purposes that even after semiconductor fabrication and sophisticated manufacturing have ceased, there will remain a bounty of hardware usable for people’s computational needs (whatever they may be).

I am not one to try and predict the future, and I don’t really want to imagine it as being along the same lines as the plot for one of Kevin Costner’s less successful movies, either, but I feel that Collapse OS and its peers, in considering various dystopian scenarios and strategies to mitigate their impacts, may actually offer more than just a hopefully sufficient kind of preparedness for a depressing future. In that future, without super-fast Internet, dopamine-fired social media, lifelike gaming, and streaming video services with huge catalogues of content available on demand, everyone has to accept that far less technology will be available to them: they get no choice in the matter. Investigating how they might manage is at the very least an interesting thought experiment. But we would be foolish to consider such matters as purely part of a possible future and not instructive in other ways.

An Overlap of Interests

As readers of my previous articles will be aware, I have something of an interest in older computers, open source hardware, and sustainable computing. Older computers lend themselves to analysis and enhancement even by individuals with modest capabilities and tools because they employ technologies that may have been regarded as “miniaturised” when they were new, but they were still amenable to manual assembly and repair. Similarly, open source hardware has grown to a broad phenomenon because the means to make computing systems and accessories has now become more accessible to individuals, as opposed to being the preserve of large and well-resourced businesses. Where these activities experience challenges, it is typically in the areas that have not yet become quite as democratised, such as semiconductor fabrication at the large-scale integration level, along with the development and manufacture of more advanced technology, such as components and devices that would be competitive with off-the-shelf commercial products.

Some of the angst around open source hardware concerns the lack of investment it receives from those who would benefit from it, but much of that investment would largely be concerned with establishing an ability to maintain some kind of parity with modern, proprietary hardware. Ignoring such performance-led requirements and focusing on simpler hardware projects, as many people already do, brings us a lot closer to retrocomputing and a lot closer to the constrained hardware scenario envisaged by Collapse OS. My own experiments with PIC32-based microcontrollers are not too far removed from this, and it would not be inconceivable to run a simple environment in the 64K of RAM and 256K of flash memory of the PIC32MX270, this being much more generous than many microcomputers and games consoles of the 1980s.

Although I relied on cross-compilation to build the programs that would run on the minimal hardware of the PIC32 microcontroller, Collapse OS emphasises self-hosting: that it is possible to build the software within the running software itself. After all, how sustainable would a frugal computing environment be if it needed a much more powerful development system to fix and improve it? For Collapse OS, such self-hosting is enabled by the use of the Forth programming language, as explained by the rationale for switching to Forth from a system implemented in assembly language. Such use of Forth is not particularly unusual: its frugal demands were prized in the microcomputer era and earlier, with its creator Charles Moore describing the characteristics of a computer designed to run Forth as needing around 8K of RAM and 8K of ROM, this providing a complete interactive system.

(If you are interested in self-hosting and bootstrapping, one place to start might be the bootstrapping wiki.)

For a short while, Forth was perhaps even thought to be the hot new thing in some circles within computing. One fairly famous example was the Jupiter Ace microcomputer, developed by former Sinclair Research designers, offering a machine that followed on fairly closely from Sinclair’s rudimentary ZX81. But in a high-minded way one might have expected from the Sinclair stable and the Cambridge scene, it offered Forth as its built-in language in response to all the other microcomputers offering “unstructured” BASIC dialects. Worthy as such goals might have been, the introduction of a machine with outdated hardware specifications condemned it in its target market as a home computer, with it offering primitive black-and-white display output against competitors offering multi-colour graphics, and offering limited amounts of memory as competitors launched with far more fitted as standard. Interestingly, the Z80 processor at the heart of the Ace was the primary target of Collapse OS, and one might wonder if the latter might actually be portable to the former, which would be an interesting project if any hardware collector wants to give it a try!

Other Forth-based computers were delivered such as the Canon Cat: an unusual “information appliance” that might have formed the basis of Apple’s Macintosh had that project not been diverted towards following up on the Apple Lisa. Dedicated Forth processors were even delivered, as anticipated already by Moore back in 1980, reminiscent of the Lisp machine era. However, one hardware-related legacy of Forth is that of the Open Firmware standard where a Forth environment provides an interactive command-line interface to a system’s bootloader. Collapse OS fits in pretty well with that kind of application of Forth. Curiously, someone did contact me when I first wrote about my PIC32 experiments, this person maintaining their own microcontroller Forth implementation, and in the context of this article I have re-established contact because I never managed to properly follow up on the matter.

Changing the Context

According to a broad interpretation of the Collapse OS hardware criteria, the PIC32MX270 would actually not be a bad choice. Like the AVR microcontrollers and the microprocessors of the 1980s, PIC32MX microcontrollers are available in convenient dual in-line packages, but unlike those older microprocessors they also offer the 32-bit MIPS architecture that is nicer to program than the awkward instruction sets of the likes of the Z80 and 6502, no matter how much nostalgia colours people’s preferences. However, instead of focusing on hardware suitability in a resource-constrained future, I want to consider the messages of simplicity and sustainability that underpin the Collapse OS initiative and might be relevant to the way we practise computing today.

When getting a PIC32 microcontroller to produce a video signal, part of the motivation was just to see how straightforward it might be to make a simple “single chip” microcomputer. Like many microcomputers back in the 1980s, it became tempting to consider how it might be used to deliver graphical demonstrations and games, but I also wondered what kind of role such a system might have in today’s world. Similar projects, including the first versions of the Maximite have emphasised such things as well, along with interfacing and educational applications (such as learning BASIC). Indeed, many low-end microcontroller-based computers attempt to recreate and to emphasise the sparse interfaces of 1980s microcomputers as a distraction-free experience for learning and teaching.

Eliminating distractions is a worthy goal, whether those distractions are things that we can conveniently seek out when our attention wanders, such as all our favourite, readily accessible Internet content, or whether they come in the form of the notifications that plague “modern” user interfaces. Another is simply reducing the level of consumption involved in our computational activities: civilisational collapse would certainly impose severe limits on that kind of consumption, but it would seem foolish to acknowledge that and then continue on the same path of ever-increasing consumption that also increasingly fails to deliver significant improvements in the user experience. When desktop applications, mobile “apps”, and Web sites frequently offer sluggish and yet overly-simplistic interfaces that are more infuriating than anything else, it might be wise to audit our progress and reconsider how we do certain things.

Human nature has us constantly exploring the boundaries of what is possible with technology, but some things which captivate people at any given point on the journey of technological progress may turn out to be distracting diversions from the route ultimately taken. In my trawl of microcomputing history over the last couple of years, I was reminded of an absurd but illustrative example of how certain technological exercises seem to become the all-consuming focus of several developers, almost being developed for the sake of it, before the fad in question flames out and everybody moves on. That example concerned “morphing” software, inspired by visual effects from movies such as Terminator 2, but operating on a simpler, less convincing level.

Suddenly, such effects were all over the television and for a few months in late 1993, everyone was supposedly interested in making the likeness of one famous person slowly change into the likeness of another, never mind that it really required a good library of images, this being somewhat before widespread digital imaging and widespread image availability. Fast-forward a few years, and it all seemed like a crazy mass delusion best never spoken of again. We might want to review our own time’s obsessions with performative animations and effects, along with the peculiarities of touch-based interfaces, the assumption of pervasive and fast connectivity, and how these drive hardware consumption and obsolescence.

Once again, some of this comes back to asking how people managed to do things in earlier times and why things sometimes seem so complicated now. Thinking back to the 1980s era of microcomputing, my favourite 8-bit computer of those times was the Acorn Electron, this being the one I had back then, and it was certainly possible to equip it to do word processing to a certain level. Acorn even pitched an expanded version as a messaging terminal for British Telecom, although I personally think that they could have made more of such opportunities, especially given the machine’s 80-column text capabilities being made available at such a low price. The user experience would not exactly be appealing by today’s standards, but then nor would that of Collapse OS, either.

When I got my PIC32 experiment working reasonably, I asked myself if it would be sufficient for tasks like simple messaging and writing articles like this. The answer, assuming that I would enhance that effort to use a USB keyboard and external storage, is probably the same as whether anyone might use a Maximite for such applications: it might not be as comfortable as on a modern system but it would be possible in some way. Given the tricks I used, certain things would actually be regressions from the Electron, such as the display resolution. Conversely, the performance of a 48MHz MIPS-based processor is obviously going to be superior to a 2MHz 6502, even when having to generate the video signal, thus allowing for some potential in other areas.

Reversing Technological Escalation

Using low-specification hardware for various applications today, considering even the PIC32 as low-spec and ignoring the microcomputers of the past, would also need us to pare back the demands that such applications have managed to accumulate over the years. As more powerful, higher-performance hardware has become available, software, specifications and standards have opportunistically grown to take advantage of that extra power, leaving many people bewildered by the result: their new computer being just as slow as their old one, for example.

Standards can be particularly vulnerable where entrenched interests drive hardware consumption whilst seeking to minimise the level of adaptation their own organisations will need to undertake in order to deliver solutions based on such standards. A severely constrained computing device may not have the capacity or performance to handle all the quirks of a “full fat” standard, but it might handle an essential core of that standard, ignoring all the edge cases and special treatment for certain companies’ products. Just as important, the developers of an implementation handling a standard also may not have the capacity or tenacity for a “full fat” standard, but they may do a reasonable job handling one that cuts out all the corporate cruft.

And beyond the technology needed to perform some kind of transaction as part of an activity, we might reconsider what is necessary to actually perform that activity. Here, we may consider the more blatant case of the average “modern” Web site or endpoint, where an activity may end up escalating and involving the performance of a number of transactions, many of which superfluous and, in the case of the pervasive cult of analytics, exploitative. What once may have been a simple Web form is often now an “experience” where the browser connects to dozens of sites, where all the scripts poll the client computer into oblivion, and where the functionality somehow doesn’t manage to work, anyway (as I recently experienced on one airline’s Web site).

Technologists and their employers may drive consumption, but so do their customers. Public institutions, utilities and other companies may lazily rely on easily procured products and services, these insisting (for “security” or “the best experience”) that only the latest devices or devices from named vendors may be used to gain access. Here, the opposite of standardisation occurs, where adherence to brand names dictates the provision of service, compounded by the upgrade treadmill familiar from desktop computing, bringing back memories of Microsoft and Intel ostensibly colluding to get people to replace their computer as often as possible.

A Broader Brush

We don’t need to go back to retrocomputing levels of technology to benefit from re-evaluating the prevalent technological habits of our era. I have previously discussed single-board computers like the MIPS Creator CI20 which, in comparison to contemporary boards from the Raspberry Pi series, was fairly competitive in terms of specification and performance (having twice the RAM of the Raspberry Pi Models A+, B and B+). Although hardly offering conventional desktop performance upon its introduction, the CI20 would have made a reasonable workstation in certain respects in earlier times: its 1GHz CPU and 1GB of RAM should certainly be plenty for many applications even now.

Sadly, starting up and using the main two desktop environments on the CI20 is an exercise in patience, and I recommend trying something like the MATE desktop environment just for something responsive. Using a Web browser like Firefox is a trial, and extensive site blocking is needed just to prevent the browser wanting to download things from all over the place, as it tries to do its bit in shoring up Google’s business model. My father was asking me the other day why a ten-year-old computer might be slow on a “modern” Web site but still perfectly adequate for watching video. I would love to hear the Firefox and Chrome developers, along with the “architects of the modern Web”, give any explanation for this that doesn’t sound like they are members of some kind of self-realisation cult.

If we can envisage a microcomputer, either a vintage one or a modern microcontroller-based one, performing useful computing activities, then we can most certainly envisage machines of ten or so years ago, even ones behind the performance curve, doing so as well. And by realising that, we might understand that we might even have the power to slow down the engineered obsolescence of computing hardware, bring usable hardware back into use, and since not everyone on the planet can afford the latest and greatest, we might even put usable hardware into the hands of more people who might benefit from it.

Naturally, this perspective is rather broader than one that only considers a future of hardship and scarcity, but hardship and scarcity are part of the present, just as they have always been part of the past. Applying many of the same concerns and countermeasures to today’s situation, albeit in less extreme forms, means that we have the power to mitigate today’s situation and, if we are optimistic, perhaps steer it away from becoming the extreme situation that the Collapse OS initiative seeks to prepare for.

Concrete Steps

I have, in the past, been accused of complaining about injustices too generally for my complaints to be taken seriously, never mind such injustices being blatant and increasingly obvious in our modern societies and expressed through the many crises of our times. So how might we seek to mitigate widespread hardware obsolescence and technology-driven overconsumption? Some suggestions in a concise list for those looking for actionable things:

  • Develop, popularise and mandate lightweight formats, protocols and standards
  • Encourage interoperability and tolerance for multiple user interfaces, clients and devices
  • Insist on an unlimited “right to repair” for computing devices including the software
  • Encourage long-term thinking in software and systems development

And now for some elucidation…

Mandatory Accessible Standards

This suggestion has already been described above, but where it would gain its power is in the idea of mandating that public institutions and businesses would be obliged to support lightweight formats, protocols and standards, and not simply as an implementation detail for their chosen “app”, like a REST endpoint might be, but actually as a formal mechanism providing service to those who would interact with those institutions. This would make the use of a broad range of different devices viable, and in the case of special-purpose devices for certain kinds of users, particularly those who would otherwise be handed a smartphone and told to “get with it”, it would offer a humane way of accessing services that is currently denied to them.

For simple dialogue-based interactions, existing formats such as JSON might even be sufficient as they are. I am reminded of a paper I read when putting together my degree thesis back in the 1990s, where the idea was that people would be able to run programs safely in their mail reader, with one example being that of submitting forms.

T-shirt ordering dialogues shown by Safe-Tcl

T-shirt ordering dialogues shown by Safe-Tcl running in a mail program, offering the recipient the chance to order some merchandise that might not be as popular now.

In that paper, most of the emphasis was on the safety of the execution environment as opposed to the way in which the transaction was to be encoded, but it is not implausible that one might have encoded the details of the transaction – the T-shirt size (with the recipient’s physical address presumably already being known to the sender) – in a serialised form of the programming language concerned (Safe-Tcl) as opposed to just dumping some unstructured text in the body of a mail. I would need to dig out my own thesis to see what ideas I had for serialised information. Certainly, such transactions even embellished with other details and choices and with explanatory information, prompts and questions do not require megabytes of HTML, CSS, JavaScript, images, videos and so on.

Interoperability and Device Choice

One thing that the Web was supposed to liberate us from was the insistence that to perform a particular task, we needed a particular application, and that particular application was only available on a particular platform. In the early days, HTML was deliberately simplistic in its display capabilities, and people had to put up with Web pages that looked very plain until things like font tags allowed people to go wild. With different factions stretching HTML in all sorts of directions, CSS was introduced to let people apply presentation attributes to documents, supposedly without polluting or corrupting the original HTML that would remain semantically pure. We all know how this turned out, particularly once the Web 2.0 crowd got going.

Back in the 1990s, I worked on an in-house application at my employer that used a document model inspired by SGML (as HTML had been), and the graphical user interface to the documents being exchanged initially offered a particular user interface paradigm when dealing with collections of data items, this being the one embraced by the Macintosh’s Finder when showing directory hierarchies in what we would now call a tree view. Unfortunately, users seemed to find expanding and hiding things by clicking on small triangles to be annoying, and so alternative presentation approaches were explored. Interestingly, the original paradigm would be familiar even now to those using generic XML editor software, but many people would accept that while such low-level editing capabilities are nice to have, higher-level representations of the data are usually much more preferable.

Such user preferences could quite easily be catered to through the availability of client software that works in the way they expect, rather than the providers of functionality or the operators of services trying to gauge what the latest fashions in user interfaces might be, as we have all seen when familiar Web sites change to mimic something one would expect to see on a smartphone, even with a large monitor on a desk with plenty of pixels to spare. With well-defined standards, if a client device or program were to see that it needed to allow a user to peruse a large collection of items or to choose a calendar date, it would defer to the conventions of that device or platform, giving the user the familiarity they expect.

This would also allow clients and devices with a wide range of capabilities to be used. The Web tried to deliver a reasonable text-only experience for a while, but most sites can hardly be considered usable in a textual browser these days. And although there is an “accessibility story” for the Web, it largely appears to involve retrofitting sites with semantic annotations to help users muddle through the verbose morass encoded in each page. Certainly, the Web of today does do one thing reasonably by mixing up structure and presentation: it can provide a means of specifying and navigating new kinds of data that might be unknown to the client, showing them something more than a text box. A decent way of extending the range of supported data types would be needed in any alternative, but it would need to spare everyone suddenly having scripts running all over the place.

Rights to Repair

The right to repair movement has traditionally been focused on physical repairs to commercial products, making sure that even if the manufacturer has abandoned a product and really wants you to buy something new from them, you can still choose to have the product repaired so that it can keep serving you well for some time to come. But if hardware remains capable enough to keep doing its job, and if we are able to slow down or stop the forces of enforced obsolescence, we also need to make sure that the software running on the hardware may also be repaired, maintained and updated. A right to repair very much applies to software.

Devotees of the cult of the smartphone, those who think that there is an “app” for everything, should really fall silent with shame. Not just for shoehorning every activity they can think of onto a device that is far from suitable for everyone, and not just for mandating commercial relationships with large multinational corporations for everyone, but also for the way that happily functioning smartphones have to be discarded because they run software that is too old and cannot be fixed or upgraded. Demanding the right to exercise the four freedoms of Free Software for our devices means that we get to decide when those devices are “too old” for what we want to use them for. If a device happens to be no longer usable for its original activity even after some Free Software repairs, we can repurpose it for something else, instead of having the vendor use those familiar security scare stories and pretending that they are acting in our best interests.

Long-Term Perspectives

If we are looking to preserve the viability of our computing devices by demanding interoperability to give them a chance to participate in the modern world and by demanding that they may be repaired, we also need to think about how the software we develop may itself remain viable, both in terms of the ability to run the software on available devices as well as the ability to keep maintaining, improving and repairing it. That potentially entails embracing unfashionable practices because “modern” practices do not exactly seem conducive to the kind of sustainable activities we have in mind.

I recently had the opportunity to contemplate the deployment of software in “virtual environments” containing entire software stacks, each of which running their own Web server program, that would receive their traffic from another Web server program running in the same virtual machine, all of this running in some cloud infrastructure. It was either that or using containers containing whole software distributions, these being deployed inside virtual machines containing their own software distributions. All because people like to use the latest and greatest stuff for everything, this stuff being constantly churned by fashionable development methodologies and downloaded needlessly over and over again from centralised Internet services run by monopolists.

Naturally, managing gigabytes of largely duplicated software is worlds, if not galaxies, away from the modest computing demands of things like Collapse OS, but it would be distasteful to anyone even a decade ago and shocking to anyone even a couple of decades ago. Unfashionable as it may seem now, software engineering courses once emphasised things like modularity and the need for formal interfaces between modules in systems. And a crucial benefit of separating out functionality into modules is to allow those modules to mature, do the job they were designed for, and to recede into the background and become something that can be relied upon and not need continual, intensive maintenance. There is almost nothing better than writing a library that one may use constantly but never need to touch again.

Thus, the idea that a precarious stack of precisely versioned software is required to deliver a solution is absurd, but it drives the attitude that established software distributions only deliver “old” software, and it drives the demand for wasteful container or virtual environment solutions whose advocates readily criticise traditional distributions whilst pilfering packages from them. Or as Docker users might all too easily say, “FROM debian:sid”. Part of the problem is that it is easy to rely on methods of mass consumption to solve problems with software – if something is broken, just update and see if it fixes it – but such attitudes permeate the entire development process, leading to continual instability and a software stack constantly in flux.

Dealing with a multitude of software requirements is certainly a challenging problem that established operating systems struggle to resolve convincingly, despite all the shoehorning of features into the Linux technology stack. Nevertheless, the topic of operating system design is rather outside the scope of this article. Closer to relevance is the matter of how people seem reluctant to pick a technology and stick with it, particularly in the realm of programming languages. Then again, I covered much of this before and fairly recently, too. Ultimately, we want to be establishing software stacks that people can readily acquaint themselves with decades down the line, without the modern-day caveats that “feature X changed in version Y” and that if you were not there at the time, you have quite the job to do to catch up with that and everything else that went on, including migrations to a variety of source management tools and venues, maybe even completely new programming languages.

A Different Mindset

If anything, Collapse OS makes us consider a future beyond tomorrow, next week, next year, or a few years’ time. Even if the wheels do not start falling off the vehicle of human civilisation, there are still plenty of other things that can go away without much notice. Corporations like Apple and Google might stick around, whether that is good news or not, but it does not stop them from pulling the plug on products and services. Projects and organisations do not always keep going forever, not least because they are often led by people who cannot keep going forever, either.

There are ways we can mitigate these threats to sustainability and longevity, however. We can share our software through accessible channels without insisting that others use those monopolist-run centralised hosting services. We can document our software so that others have a chance of understanding what we were thinking when we wrote it. We can try and keep the requirements for our software modest and give people a chance to deploy it on modest hardware. And we might think about what kind of world we are leaving behind and whether it is better than the world we were born into.

Public Money, Public Code, Public Control

Thursday, September 14th, 2017

An interesting article published by the UK Government Digital Service was referenced in a response to the LWN.net coverage of the recently-launched “Public Money, Public Code” campaign. Arguably, the article focuses a little too much on “in the open” and perhaps not enough on the matter of control. Transparency is a good thing, collaboration is a good thing, no-one can really argue about spending less tax money and getting more out of it, but it is the matter of control that makes this campaign and similar initiatives so important.

In one of the comments on the referenced article you can already see the kind of resistance that this worthy and overdue initiative will meet. There is this idea that the public sector should just buy stuff from companies and not be in the business of writing software. Of course, this denies the reality of delivering solutions where you have to pay attention to customer needs and not just have some package thrown through the doorway of the customer as big bucks are exchanged for the privilege. And where the public sector ends up managing its vendors, you inevitably get an under-resourced customer paying consultants to manage those vendors, maybe even their own consultant colleagues. Guess how that works out!

There is a culture of proprietary software vendors touting their wares or skills to public sector departments, undoubtedly insisting that their products are a result of their own technological excellence and that they are doing their customers a favour by merely doing business with them. But at the same time, those vendors need a steady – perhaps generous – stream of revenue consisting largely of public money. Those vendors do not want their customers to have any real control: they want their customers to be obliged to come back year after year for updates, support, further sales, and so on; they want more revenue opportunities rather than their customers empowering themselves and collaborating with each other. So who really needs whom here?

Some of these vendors undoubtedly think that the public sector is some kind of vehicle to support and fund enterprises. (Small- and medium-sized enterprises are often mentioned, but the big winners are usually the corporate giants.) Some may even think that the public sector is a vehicle for “innovation” where publicly-funded work gets siphoned off for businesses to exploit. Neither of these things cultivate a sustainable public sector, nor do they even create wealth effectively in wider society: they lock organisations into awkward, even perilous technological dependencies, and they undermine competition while inhibiting the spread of high-quality solutions and the effective delivery of services.

Unfortunately, certain flavours of government hate the idea that the state might be in a role of actually doing anything itself, preferring that its role be limited to delegating everything to “the market” where private businesses will magically do everything better and cheaper. In practice, under such conditions, some people may benefit (usually the rich and well-represented) but many others often lose out. And it is not unknown for the taxpayer to have to pick up the bill to fix the resulting mess that gets produced, anyway.

We need sustainable public services and a sustainable software-producing economy. By insisting on Free Software – public code – we can build the foundations of sustainability by promoting interoperability and sharing, maximising the opportunities for those wishing to improve public services by upholding proper competition and establishing fair relationships between customers and vendors. But this also obliges us to be vigilant to ensure that where politicians claim to support this initiative, they do not try and limit its impact by directing money away from software development to the more easily subverted process of procurement, while claiming that procured systems not be subject to the same demands.

Indeed, we should seek to expand our campaigning to cover public procurement in general. When public money is used to deliver any kind of system or service, it should not matter whether the code existed in some form already or not: it should be Free Software. Otherwise, we indulge those who put their own profits before the interests of a well-run public sector and a functioning society.

The Academic Barriers of Commercialisation

Monday, January 9th, 2017

Last year, the university through which I obtained my degree celebrated a “milestone” anniversary, meaning that I got even more announcements, notices and other such things than I was already getting from them before. Fortunately, not everything published into this deluge is bound up in proprietary formats (as one brochure was, sitting on a Web page in Flash-only form) or only reachable via a dubious “Libyan link-shortener” (as certain things were published via a social media channel that I have now quit). It is indeed infuriating to see one of the links in a recent HTML/plain text hybrid e-mail message using a redirect service hosted on the university’s own alumni sub-site, sending the reader to a bit.ly URL, which will redirect them off into the great unknown and maybe even back to the original site. But such things are what one comes to expect on today’s Internet with all the unquestioning use of random “cloud” services, each one profiling the unsuspecting visitor and betraying their privacy to make a few extra cents or pence.

But anyway, upon following a more direct – but still redirected – link to an article on the university Web site, I found myself looking around to see what gets published there these days. Personally, I find the main university Web site rather promotional and arguably only superficially informative – you can find out the required grades to take courses along with supposed student approval ratings and hypothetical salary expectations upon qualifying – but it probably takes more digging to get at the real detail than most people would be willing to do. I wouldn’t mind knowing what they teach now in their computer science courses, for instance. I guess I’ll get back to looking into that later.

Gatekeepers of Knowledge

However, one thing did catch my eye as I browsed around the different sections, encountering the “technology transfer” department with the expected rhetoric about maximising benefits to society: the inevitable “IP” policy in all its intimidating length, together with an explanatory guide to that policy. Now, I am rather familiar with such policies from my time at my last academic employer, having been obliged to sign some kind of statement of compliance at one point, but then apparently not having to do so when starting a subsequent contract. It was not as if enlightenment had come calling at the University of Oslo between these points in time such that the “IP rights” agreement now suddenly didn’t feature in the hiring paperwork; it was more likely that such obligations had presumably been baked into everybody’s terms of employment as yet another example of the university upper management’s dubious organisational reform and questionable human resources practices.

Back at Heriot-Watt University, credit is perhaps due to the authors of their explanatory guide to try and explain the larger policy document, because it is most likely that most people aren’t going to get through that much longer document and retain a clear head. But one potentially unintended reason for credit is that by being presented with a much less opaque treatment of the policy and its motivations, we are able to see with enhanced clarity many of the damaging misconceptions that have sadly become entrenched in higher education and academia, including the ways in which such policies actually do conflict with the sharing of knowledge that academic endeavour is supposed to be all about.

So, we get the sales pitch about new things needing investment…

However, often new technologies and inventions are not fully developed because development needs investment, and investment needs commercial returns, and to ensure commercial returns you need something to sell, and a freely available idea cannot be sold.

If we ignore various assumptions about investment or the precise economic mechanisms supposedly required to bring about such investment, we can immediately note that ideas on their own aren’t worth anything anyway, freely available or not. Although the Norwegian Industrial Property Office (or the Norwegian Patent Office if we use a more traditional name) uses the absurd vision slogan “turning ideas into values” (it should probably read “value”, but whatever), this perhaps says more about greedy profiteering through the sale of government-granted titles bound to arbitrary things than it does about what kinds of things have any kind of inherent value that you can take to the bank.

But assuming that we have moved beyond the realm of simple ideas and have entered the realm of non-trivial works, we find that we have also entered the realm of morality and attitude management:

That is why, in some cases, it is better for your efforts not to be published immediately, but instead to be protected and then published, for protection gives you something to sell, something to sell can bring in investment, and investment allows further development. Therefore in the interests of advancing the knowledge within the field you work in, it is important that you consider the commercial potential of your work from the outset, and if necessary ensure it is properly protected before you publish.

Once upon a time, the most noble pursuit in academic research was to freely share research with others so that societal, scientific and technological progress could be made. Now it appears that the average researcher should treat it as their responsibility to conceal their work from others, seek “protection” on it, and then release the encumbered details for mere perusal and the conditional participation of those once-valued peers. And they should, of course, be wise to the commercial potential of the work, whatever that is. Naturally, “intellectual property” offices in such institutions have an “if in doubt, see us” policy, meaning that they seek to interfere with research as soon as possible, and should someone fail to have “seen them”, that person’s loyalty may very well be called into question as if they had somehow squandered their employer’s property. In some institutions, this could very easily get people marginalised or “reorganised” if not immediately or obviously fired.

The Rewards of Labour

It is in matters of property and ownership where things get very awkward indeed. Many people would accept that employees of an organisation are producing output that becomes the property of that organisation. What fewer people might accept is that the customers of an organisation are also subject to having their own output taken to be the property of that organisation. The policy guide indicates that even undergraduate students may also be subject to an obligation to assign ownership of their work to the university: those visiting the university supposedly have to agree to this (although it doesn’t say anything about what their “home institution” might have to say about that), and things like final year projects are supposedly subject to university ownership.

So, just because you as a student have a supervisor bound by commercialisation obligations, you end up not only paying tuition fees to get your university education (directly or through taxation), but you also end up having your own work taken off you because it might be seen as some element in your supervisor’s “portfolio”. I suppose this marks a new low in workplace regulation and standards within a sector that already skirts the law with regard to how certain groups are treated by their employers.

One can justifiably argue that employees of academic institutions should not be allowed to run away with work funded by those institutions, particularly when such funding originally comes from other sources such as the general public. After all, such work is not exactly the private property of the researchers who created it, and to treat it as such would deny it to those whose resources made it possible in the first place. Any claims about “rightful rewards” needing to be given are arguably made to confuse rational thinking on the matter: after all, with appropriate salaries, the researchers are already being rewarded doing work that interests and stimulates them (unlike a lot of people in the world of work). One can argue that academics increasingly suffer from poorer salaries, working conditions and career stability, but such injustices are not properly remedied by creating other injustices to supposedly level things out.

A policy around what happens with the work done in an academic institution is important. But just as individuals should not be allowed to treat broadly-funded work as their own private property, neither should the institution itself claim complete ownership and consider itself entitled to do what it wishes with the results. It may be acting as a facilitator to allow research to happen, but by seeking to intervene in the process of research, it risks acting as an inhibitor. Consider the following note about “confidential information”:

This is, in short, anything which, if you told people about, might damage the commercial interests of the university. It specifically includes information relating to intellectual property that could be protected, but isn’t protected yet, and which if you told people about couldn’t be protected, and any special know how or clever but non patentable methods of doing things, like trade secrets. It specifically includes all laboratory notebooks, including those stored in an electronic fashion. You must be very careful with this sort of information. This is of particular relevance to something that may be patented, because if other people know about it then it can’t be.

Anyone working in even a moderately paranoid company may have read things like this. But here the context is an environment where knowledge should be shared to benefit and inform the research community. Instead, one gets the impression that the wish to control the propagation of knowledge is so great that some people would rather see the details of “clever but non patentable methods” destroyed than passed on openly for others to benefit from. Indeed, one must question whether “trade secrets” should even feature in a university environment at all.

Of course, the obsession with “laboratory notebooks”, “methods of doing things” and “trade secrets” in such policies betrays the typical origins of such drives for commercialisation: the apparently rich pickings to be had in the medical, pharmaceutical and biosciences domains. It is hardly a coincidence that the University of Oslo intensified its dubious “innovation” efforts under a figurehead with a background (or an interest) in exactly those domains: with a narrow personal focus, an apparent disdain for other disciplines, and a wider commercial atmosphere that gives such a strategy a “dead cert” air of impending fortune, we should perhaps expect no more of such a leadership creature (and his entourage) than the sum of that creature’s instincts and experiences. But then again, we should demand more from such people when their role is to cultivate an institution of learning and not to run a private research organisation at the public’s expense.

The Dirty Word

At no point in the policy guide does the word “monopoly” appear. Given that such a largely technical institution would undoubtedly be performing research where the method of “protection” would involve patents being sought, omitting the word “monopoly” might be that document’s biggest flaw. Heriot-Watt University originates from the merger of two separate institutions, one of which was founded by the well-known pioneer of steam engine technology, James Watt.

Recent discussion of Watt’s contributions to the development and proliferation of such technology has brought up claims that Watt’s own patents – the things that undoubtedly made him wealthy enough to fund an educational organisation – actually held up progress in the domain concerned for a number of decades. While he was clearly generous and sensible enough to spend his money on worthy causes, one can always challenge whether the questionable practices that resulted in the accumulation of such wealth can justify the benefits from the subsequent use of that wealth, particularly if those practices can be regarded as having had negative effects of society and may even have increased wealth inequality.

Questioning philanthropy is not a particularly fashionable thing to do. In capitalist societies, wealthy people are often seen as having made their fortunes in an honest fashion, enjoying a substantial “benefit of the doubt” that this was what really occurred. Criticising a rich person giving money to ostensibly good causes is seen as unkind to both the generous donor and to those receiving the donations. But we should question the means through which the likes of Bill Gates (in our time) and James Watt (in his own time) made their fortunes and the power that such fortunes give to such people to direct money towards causes of their own personal choosing, not to mention the way in which wealthy people also choose to influence public policy and the use of money given by significantly less wealthy individuals – the rest of us – gathered through taxation.

But back to monopolies. Can they really be compatible with the pursuit and sharing of knowledge that academia is supposed to be cultivating? Just as it should be shocking that secretive “confidentiality” rules exist in an academic context, it should appal us that researchers are encouraged to be competitively hostile towards their peers.

Removing the Barriers

It appears that some well-known institutions understand that the unhindered sharing of their work is their primary mission. MIT Media Lab now encourages the licensing of software developed under its roof as Free Software, not requiring special approval or any other kind of institutional stalling that often seems to take place as the “innovation” vultures pick over the things they think should be monetised. Although proprietary licensing still appears to be an option for those within the Media Lab organisation, at least it seems that people wanting to follow their principles and make their work available as Free Software can do so without being made to feel bad about it.

As an academic institution, we believe that in many cases we can achieve greater impact by sharing our work.

So says the director of the MIT Media Lab. It says a lot about the times we live in that this needs to be said at all. Free Software licensing is, as a mechanism to encourage sharing, a natural choice for software, but we should also expect similar measures to be adopted for other kinds of works. Papers and articles should at the very least be made available using content licences that permit sharing, even if the licence variants chosen by authors might seek to prohibit the misrepresentation of parts of their work by prohibiting remixes or derived works. (This may sound overly restrictive, but one should consider the way in which scientific articles are routinely misrepresented by climate change and climate science deniers.)

Free Software has encouraged an environment where sharing is safely and routinely done. Licences like the GNU General Public Licence seek to shield recipients from things like patent threats, particularly from organisations which might appear to want to share their works, but which might be tempted to use patents to regulate the further use of those works. Even in realms where patents have traditionally been tolerated, attempts have been made to shield others from the effects of patents, intended or otherwise: the copyleft hardware movement demands that shared hardware designs are patent-free, for instance.

In contrast, one might think that despite the best efforts of the guide’s authors, all the precautions and behavioural self-correction it encourages might just drive the average researcher to distraction. Or, just as likely, to ignoring most of the guidelines and feigning ignorance if challenged by their “innovation”-obsessed superiors. But in the drive to monetise every last ounce of effort there is one statement that is worth remembering:

If intellectual property is not assigned, this can create problems in who is allowed to exploit the work, and again work can go to waste due to a lack of clarity over who owns what.

In other words, in an environment where everybody wants a share of the riches, it helps to have everybody’s interests out in the open so that there may be no surprises later on. Now, it turns out that unclear ownership and overly casual management of contributions is something that has occasionally threatened Free Software projects, resulting in more sophisticated thinking about how contributions are managed.

And it is precisely this combination of Free Software licensing, or something analogous for other domains, with proper contribution and attribution management that will extend safe and efficient sharing of knowledge to the academic realm. Researchers just cannot have the same level of confidence when dealing with the “technology transfer” offices of their institution and of other institutions. Such offices only want to look after themselves while undermining everyone beyond the borders of their own fiefdoms.

Divide and Rule

It is unfortunate that academic institutions feel that they need to “pull their weight” and have to raise funds to make up for diminishing public funding. By turning their backs on the very reason for their own existence and seeking monopolies instead of sharing knowledge, they unwittingly participate in the “divide and rule” tactics blatantly pursued in the political arena: that everyone must fight each other for all that is left once the lion’s share of public funding has been allocated to prestige megaprojects and schemes that just happen to benefit the well-connected, the powerful and the influential people in society the most.

A properly-funded education sector is an essential component of a civilised society, and its institutions should not be obliged to “sharpen their elbows” in the scuffle for funding and thus deprive others of knowledge just to remain viable. Sadly, while austerity politics remains fashionable, it may be up to us in the Free Software realm to remind academia of its obligations and to show that sustainable ways of sharing knowledge exist and function well in the “real world”.

Indeed, it is up to us to keep such institutions honest and to prevent advocates of monopoly-driven “innovation” from being able to insist that their way is the only way, because just as “divide and rule” politics erects barriers between groups in wider society, commercialisation erects barriers that inhibit the essential functions of academic pursuit. And such barriers ultimately risk extinguishing academia altogether, along with all the benefits its institutions bring to society. If my university were not reinforcing such barriers with its “IP” policy, maybe its anniversary as a measure of how far we have progressed from monopolies and intellectual selfishness would have been worth celebrating after all.

The BBC Micro and the BBC Micro Bit

Sunday, March 22nd, 2015

At least on certain parts of the Internet as well as in other channels, there has been a degree of excitement about the announcement by the BBC of a computing device called the “Micro Bit“, with the BBC’s plan to give one of these devices to each child starting secondary school, presumably in September 2015, attracting particular attention amongst technology observers and television licence fee-payers alike. Details of the device are a little vague at the moment, but the announcement along with discussions of the role of the corporation and previous initiatives of this nature provides me with an opportunity to look back at the original BBC Microcomputer, evaluate some of the criticisms (and myths) around the associated Computer Literacy Project, and to consider the things that were done right and wrong, with the latter hopefully not about to be repeated in this latest endeavour.

As the public record reveals, at the start of the 1980s, the BBC wanted to engage its audience beyond television programmes describing the growing microcomputer revolution, and it was decided that to do this and to increase computer literacy generally, it would need to be able to demonstrate various concepts and technologies on a platform that would be able to support the range of activities to which computers were being put to use. Naturally, a demanding specification was constructed – clearly, the scope of microcomputing was increasing rapidly, and there was a lot to demonstrate – and various manufacturers were invited to deliver products that could be used as this reference platform. History indicates that a certain amount of acrimony followed – a complete description of which could fill an entire article of its own – but ultimately only Acorn Computers managed to deliver a machine that could do what the corporation was asking for.

An Ambitious Specification

It is worth considering what the BBC Micro was offering in 1981, especially when considering ill-informed criticism of the machine’s specifications by people who either prefer other systems or who felt that participating in the development of such a machine was none of the corporation’s business. The technologies to be showcased by the BBC’s programme-makers and supported by the materials and software developed for the machine included full-colour graphics, multi-channel sound, 80-column text, Viewdata/Teletext, cassette and diskette storage, local area networking, interfacing to printers, joysticks and other input/output devices, as well as to things like robots and user-developed devices. Although it is easy to pick out one or two of these requirements, move forwards a year or two, increase the budget two- or three-fold, or any combination of these things, and to nominate various other computers, there really were few existing systems that could deliver all of the above, at least at an affordable price at the time.

Some microcomputers of the early 1980s
Computer RAM Text Graphics Year Price
Apple II Plus Up to 64K 40 x 25 (upper case only) 280 x 192 (6 colours), 40 x 48 (16 colours) 1979 £1500 or more
Commodore PET 4032/8032 32K 40/80 x 25 Graphics characters (2 colours) 1980 £800 (4032), £1030 (8032) (including monochrome monitor)
Commodore VIC-20 5K 22 x 23 176 x 184 (8 colours) 1980 (1981 outside Japan) £199
IBM PC (Model 5150) 16K up to 256K 40/80 x 25 640 x 200 (2 colours), 320 x 200 (4 colours) 1981 £1736 (including monochrome monitor, presumably with 16K or 64K)
BBC Micro (Model B) 32K 80/40/20 x 32/24, Teletext 640 x 256 (2 colours), 320 x 256 (2/4 colours), 160 x 256 (4/8 colours) 1981 £399 (originally £335)
Research Machines LINK 480Z 64K (expandable to 256K) 40 x 24 (optional 80 x 24) 160 x 72, 80 x 72 (2 colours); expandable to 640 x 192 (2 colours), 320 x 192 (4 colours), 190 x 96 (8 colours or 16 shades) 1981 £818
ZX Spectrum 16K or 48K 32 x 24 256 x 192 (16 colours applied using attributes) 1982 £125 (16K), £175 (48K)
Commodore 64 64K 40 x 25 320 x 200 (16 colours applied using attributes) 1982 £399

Perhaps the closest competitor, already being used in a fairly limited fashion in educational establishments in the UK, was the Commodore PET. However, it is clear that despite the adaptability of that system, its display capabilities were becoming increasingly uncompetitive, and Commodore had chosen to focus on the chipsets that would power the VIC-20 and Commodore 64 instead. (The designer of the PET went on to make the very capable, and understandably more expensive, Victor 9000/Sirius 1.) That Apple products were notoriously expensive and, indeed, the target of Commodore’s aggressive advertising did not seem to prevent them from capturing the US education market from the PET, but they always remained severely uncompetitive in the UK as commentators of the time indeed noted.

Later, the ZX Spectrum and Commodore 64 were released. Technology was progressing rapidly, and in hindsight one might have advocated waiting around until more capable and cheaper products came to market. However, it can be argued that in fulfilling virtually all aspects of the ambitious specification and pricing, it would not be until the release of the Amstrad CPC series in 1984 that a suitable alternative product might have become available. Even then, these Amstrad computers actually benefited from the experience accumulated in the UK computing industry from the introduction of the BBC Micro: they were, if anything, an iteration within the same generation of microcomputers and would even have used the same 6502 CPU as the BBC Micro had it not been for time-to-market pressures and the readily-available expertise with the Zilog Z80 CPU amongst those in the development team. And yet, specific aspects of the specification would still be unfulfilled: the BBC Micro had hardware support for Teletext displays, although it would have been possible to emulate these with a bitmapped display and suitable software.

Arise Sir Clive

Much has been made of the disappointment of Sir Clive Sinclair that his computers were not adopted by the BBC as products to be endorsed and targeted at schools. Sinclair made his name developing products that were competitive on price, often seeking cost-reduction measures to reach attractive pricing levels, but such measures also served to make his products less desirable. If one reads reviews of microcomputers from the early 1980s, many reviewers explicitly mention the quality of the keyboard provided by the computers being reviewed: a “typewriter” keyboard with keys that “travel” appear to be much preferred over the “calculator” keyboards provided by computers like the ZX Spectrum, Oric 1 or Newbury NewBrain, and they appear to be vastly preferred over the “membrane” keyboards employed by the ZX80, ZX81 and Atari 400.

For target audiences in education, business, and in the home, it would have been inconceivable to promote a product with anything less than a “proper” keyboard. Ultimately, the world had to wait until the ZX Spectrum +2 released in 1986 for a Spectrum with such a keyboard, and that occurred only after the product line had been acquired by Amstrad. (One might also consider the ZX Spectrum+ in 1984, but its keyboard was more of a hybrid of the calculator keyboard that had been used before and the “full-travel” keyboards provided by its competitors.)

Some people claim that they owe nothing to the BBC Micro and everything to the ZX Spectrum (or, indeed, the computer they happened to own) for their careers in computing. Certainly, the BBC Micro was an expensive purchase for many people, although contrary to popular assertion it was not any more expensive than the Commodore 64 upon that computer’s introduction in the UK, and for those of us who wanted BBC compatibility at home on a more reasonable budget, the Acorn Electron was really the only other choice. But it would be as childish as the playground tribalism that had everyone insist that their computer was “the best” to insist that the BBC Micro had no influence on computer literacy in general, or on the expectations of what computer systems should provide. Many people who owned a ZX Spectrum freely admit that the BBC Micro coloured their experiences, some even subsequently seeking to buy one or one of its successors and to go on to build a successful software development career.

The Costly IBM PC

Some commentators seem to consider the BBC Micro as having been an unnecessary diversion from the widespread adoption of the IBM PC throughout British society. As was the case everywhere else, the de-facto “industry standard” of the PC architecture and DOS captured much of the business market and gradually invaded the education sector from the top down, although significantly better products existed both before and after its introduction. It is tempting with hindsight to believe that by placing an early bet on the success of the then-new PC architecture, business and education could have somehow benefited from standardising on the PC and DOS applications. And there has always been the persistent misguided belief amongst some people that schools should be training their pupils/students for a narrow version of “the world of work”, as opposed to educating them to be able to deal with all aspects of their lives once their school days are over.

What many people forget or fail to realise is that the early 1980s witnessed rapid technological improvement in microcomputing, that there were many different systems and platforms, some already successful and established (such as CP/M), and others arriving to disrupt ideas of what computing should be like (the Xerox Alto and Star having paved the way for the Apple Lisa and Macintosh, the Atari ST, and so on). It was not clear that the IBM PC would be successful at all: IBM had largely avoided embracing personal computing, and although the system was favourably reviewed and seen as having the potential for success, thanks to IBM’s extensive sales organisation, other giants of the mainframe and minicomputing era such as DEC and HP were pursuing their own personal computing strategies. Moreover, existing personal computers were becoming entrenched in certain markets, and early adopters were building a familiarity with those existing machines that was reflected in publications and materials available at the time.

Despite the technical advantages of the IBM PC over much of the competition at the beginning of the 1980s, it was also substantially more expensive than the mass-market products arriving in significant numbers, aimed at homes, schools and small businesses. With many people remaining intrigued but unconvinced by the need for a personal computer, it would have been impossible for a school to justify spending almost £2000 (probably around £8000 today) on something without proven educational value. Software would also need to be purchased, and the procurement of expensive and potentially non-localised products would have created even more controversy.

Ultimately, the Computer Literacy Project stimulated the production of a wide range of locally-produced products at relatively inexpensive prices, and while there may have been a few years of children learning BBC BASIC instead of one of the variants of BASIC for the IBM PC (before BASIC became a deprecated aspect of DOS-based computing), it is hard to argue that those children missed out on any valuable experience using DOS commands or specific DOS-based products, especially since DOS became a largely forgotten environment itself as technological progress introduced new paradigms and products, making “hard-wired”, product-specific experience obsolete.

The Good and the Bad

Not everything about the BBC Micro and its introduction can be considered unconditionally good. Choices needed to be made to deliver a product that could fulfil the desired specification within certain technological constraints. Some people like to criticise BBC BASIC as being “non-standard”, for example, which neglects the diversity of BASIC dialects that existed at the dawn of the 1980s. Typically, for such people “standard” equates to “Microsoft”, but back then Microsoft BASIC was a number of different things. Commodore famously paid a one-off licence fee to use Microsoft BASIC in its products, but the version for the Commodore 64 was regarded as lacking user-friendly support for graphics primitives and other interesting hardware features. Meanwhile, the MSX range of microcomputers featured Microsoft Extended BASIC which did provide convenient access to hardware features, although the MSX range of computers were not the success at the low end of the market that Microsoft had probably desired to complement its increasing influence at the higher end through the IBM PC. And it is informative in this regard to see just how many variants of Microsoft BASIC were produced, thanks to Microsoft’s widespread licensing of its software.

Nevertheless, the availability of one company’s products do not make a standard, particularly if interoperability between those products is limited. Neither BBC BASIC nor Microsoft BASIC can be regarded as anything other than de-facto standards in their own territories, and it is nonsensical to regard one as non-standard when the other has largely the same characteristics as a proprietary product in widespread use, even if it was licensed to others, as indeed both Microsoft BASIC and BBC BASIC were. Genuine attempts to standardise BASIC did indeed exist, notably BASICODE, which was used in the distribution of programs via public radio broadcasts. One suspects that people making casual remarks about standard and non-standard things remain unaware of such initiatives. Meanwhile, Acorn did deliver implementations of other standards-driven programming languages such as COMAL, Pascal, Logo, Lisp and Forth, largely adhering to any standards subject to the limitations of the hardware.

However, what undermined the BBC Micro and Acorn’s related initiatives over time was the control that they as a single vendor had over the platform and its technologies. At the time, a “winner takes all” mentality prevailed: Commodore under Jack Tramiel had declared a “price war” on other vendors and had caused difficulties for new and established manufacturers alike, with Atari eventually being sold to Tramiel (who had resigned from Commodore) by Warner Communications, but many companies disappeared or were absorbed by others before half of the decade had passed. Indeed, Acorn, who had released the Electron to compete with Sinclair Research at the lower end of the market, and who had been developing product lines to compete in the business sector, experienced financial difficulties and was ultimately taken over by Olivetti; Sinclair, meanwhile, experienced similar difficulties and was acquired by Amstrad. In such a climate, ideas of collaboration seemed far from everybody’s minds.

Since then, the protagonists of the era have been able to reflect on such matters, Acorn co-founder Hermann Hauser admitting that it may have been better to license Acorn’s Econet local area networking technology to interested competitors like Commodore. Although the sentiments might have something to do with revenues and influence – it was at Acorn that the ARM processor was developed, sowing the seeds of a successful licensing business today – the rest of us may well ask what might have happened had the market’s participants of the era cooperated on things like standards and interoperability, helping their customers to protect their investments in technology, and building a bigger “common” market for third-party products. What if they had competed on bringing technological improvements to market without demanding that people abandon their existing purchases (and cause confusion amongst their users) just because those people happened to already be using products from a different vendor? It is interesting to see the range of available BBC BASIC implementations and to consider a world where such building blocks could have been adopted by different manufacturers, providing a common, heterogeneous platform built on cooperation and standards, not the imposition of a single hardware or software platform.

But That Was Then

Back then, as Richard Stallman points out, proprietary software was the norm. It would have been even more interesting had the operating systems and the available software for microcomputers been Free Software, but that may have been asking too much at the time. And although computer designs were often shared and published, a tendency to prevent copying of commercial computer designs prevailed, with Acorn and Sinclair both employing proprietary integrated circuits mostly to reduce complexity and increase performance, but partly to obfuscate their hardware designs, too. Thus, it may have been too much to expect something like the BBC Micro to have been open hardware to any degree “back in the day”, although circuit diagrams were published in publicly-available technical documentation.

But we have different expectations now. We expect software to be freely available for inspection, modification and redistribution, knowing that this empowers the end-users and reassures them that the software does what they want it to do, and that they remain in control of their computing environment. Increasingly, we also expect hardware to exhibit the same characteristics, perhaps only accepting that some components are particularly difficult to manufacture and that there are physical and economic restrictions on how readily we may practise the modification and redistribution of a particular device. Crucially, we demand control over the software and hardware we use, and we reject attempts to prevent us from exercising that control.

The big lesson to be learned from the early 1980s, to be considered now in the mid-2010s, is not how to avoid upsetting a popular (but ultimately doomed) participant in the computing industry, as some commentators might have everybody believe. It is to avoid developing proprietary solutions that favour specific organisations and that, despite the general benefits of increased access to technology, ultimately disempower the end-user. And in this era of readily available Free Software and open hardware platforms, the lesson to be learned is to strengthen such existing platforms and to work with them, letting those products and solutions participate and interoperate with the newly-introduced initiative in question.

The BBC Micro was a product of its time and its development was very necessary to fill an educational need. Contrary to the laziest of reports, the Micro Bit plays a different role as an accessory rather than as a complete home computer, at least if we may interpret the apparent intentions of its creators. But as a product of this era, our expectations for the Micro Bit are greater: we expect transparency and interoperability, the ability to make our own (just as one can with the Arduino, as long as one does not seek to call it an Arduino without asking permission from the trademark owner), and the ability to control exactly how it works. Whether there is a need to develop a completely new hardware solution remains an unanswered question, but we may assert that should it be necessary, such a solution should be made available as open hardware to the maximum possible extent. And of course, the software controlling it should be Free Software.

As we edge gradually closer to September and the big deployment, it will be interesting to assess how the device and the associated initiative measures up to our expectations. Let us hope that the right lessons from the days of the BBC Micro have indeed been learned!

Når sannheten kommer frem

Monday, April 28th, 2014

For noen måneder siden skrev jeg om avgjørelsen foretatt av ledelsen til Universitetet i Oslo å innføre Microsoft Exchange som gruppevareløsning. I teksten viste jeg til kommentarer som handlet om Roundcube og hvorfor den løsningen ikke kunne brukes i fremtiden hos UiO. Nå etter at universitetets IT-avdeling publiserte en nyhetssak som omtaler SquirrelMail, får jeg muligheten til å grave opp og dele noen detaljer fra mine samtaler med talspersonen tilknyttet prosjektet som vurderte Exchange og andre løsninger.

Når man ser rundt på nettet, virker det ikke uvanlig at organisasjoner innførte Roundcube parallelt med SquirrelMail på grunn av bekymringer om “universell utforming” (UU). Men i samtaler med prosjektet ble jeg fortalt at diverse mangler i UU-støtte var en aktuell grunn for at Roundcube ikke kunne bli en del av en fremtidig løsning for webmail hos UiO. Nå kommer sannheten frem:

“Ved innføringen av Roundcube som UiOs primære webmail-tjeneste fikk SquirrelMail lov til å leve videre i parallell fordi Roundcube hadde noen mangler knyttet til universell utforming. Da disse var forbedret hadde ledelsen besluttet at e-post og kalender skulle samles i et nytt system.”

Det finnes to ting som fanger vår oppmerksomhet her:

  1. At Roundcube hadde mangler “ved innføringen”, mens Roundcube hadde vært i bruk i noen år før den tvilsomme prosessen ble satt i gang for å vurdere e-post og kalenderløsninger.
  2. At forbedringene kom tilfeldigvis for sent for å påvirke ledelsens beslutning å innføre Exchange.

I fjor sommer, uten å dele prosjektgruppens påstander om mangler i Roundcube direkte i offentlighet, hørte jeg med andre om det fantes kjente mangler og forbedringspotensial i Roundcube i dette området. Var UU virkelig noe som hindret utbredelsen av Roundcube? Jeg satte meg inn i UU-teknologier og prøvde noen av dem med Roundcube for å se om situasjonen kunne forbedres med egen innsats. Det kan hende at det var store mangler i Roundcube tilbake i 2011 da prosjektgruppen begynte sitt arbeid – jeg velger ikke å hevde noe slikt – men etter at slike mangler ikke kom frem i 2012 i prosjektets sluttrapport (der Roundcube faktisk ble anbefalt som en del av den åpne kandidaten), må vi konkludere at slike bekymringer var for lengst borte og at universitetets egen webmail-tjeneste, selv om den er tilpasset organisasjonens egen visuelle profil (som kan ha noe å si om vedlikehold), var og fortsatt er tilgjengelig for alle brukere.

Hvis vi våger å tro at utdraget ovenfor forteller sannheten må vi nå konkludere at ledelsens beslutning fant sted lenge før selve prosessen som skulle underbygge denne beslutningen ble avsluttet. Og her må vi anse ordene “ledelsen besluttet” i et annet lys enn det som ellers er vanlig – der ledelsen først drar nytte av kompetanse i organisasjonen, og så tar en informert beslutning – med å anta at, som tidligere skrevet, noen “måtte ha” noe de likte, tok beslutningen de ville ta uansett, og så fikk andre til å finne på unnskyldninger og et grunnlag som virker fornuftig nok overfor utenforstående.

Det er én sak at en tilstrekkelig og fungerende IT-løsning som også er fri programvare svartmales mens tilsynelatende upopulære proprietære løsninger som Fronter når ikke opp slik at problemer rapportert i 2004 ble omtalt igjen i 2012 uten at produsenten gjør noe vesentlig mer enn å love å “jobbe med tilgjengelighet” fremover. Det er en annen sak at bærekraftige investeringer i fri programvare virker så fremmed til beslutningstakere hos UiO at folk heller vil snu arbeidsdagen til den vanlige ansatte opp-ned, slik utskiftningen av e-postinfrastrukturen har gjort for noen, enn å undersøke muligheten til å foreta relativt små utviklingsoppgaver for å oppgradere eksisterende systemer, og slippe at folk må “[stå] på dag og natt” i løpet av “de siste månedene” (og her viser vi selvsagt ikke til folk i ledelsen).

At ansatte i IT-avdelingen fikk munnkurv omkring prosessen og måtte forlange at deler av beslutningsgrunnlaget ikke når frem til offentligheten. At “menige” IT-ansvarlige må løpe hit og dit for å sørge for at ingen blir altfor misfornøyd med utfallet til en avgjørelse hverken de eller andre vanlige ansatte hadde vesentlig innflytelse på. At andre må tilpasse seg preferansene til folk som egentlig burde hatt ingenting å si om hvordan andre utfører sitt daglig arbeid. Alt dette sier noe om ledelseskulturen og demokratiske tilstander internt i Universitetet i Oslo.

Men sannheten kommer langsomt frem.

Jeg må ha… derfor må vi ha…

Monday, February 17th, 2014

For et par år siden satte Universitetet i Oslo i gang en prosess for å vurdere gruppevareløsninger (systemer som skal motta, lagre og sende organisasjonens e-postmeldinger og samtidig tillate lagring og deling av kalenderinformasjon som avtaler og møter). Prosessen vurderte et ukjent antall løsninger, droppet alle bortsett fra fire hovedkandidater, og lot være å vurdere en av kandidatene fordi den var det eksisterende kalendersystemet i bruk hos universitetet. Prosessens utfall var en oppsummering av de tre gjenstående løsningene med fordeler og ulemper beskrevet i en 23-siders rapport.

Rapportens konklusjon var at én av løsningene ikke nådde opp til forventninger omkring brukervennlighet eller åpenhet, én hadde tilstrekkelig brukervennlighet og baseres på fri programvare og åpne standarder, og én ble betraktet som å være mest brukervennlig selv om den baseres på et proprietært produkt og proprietære teknologier. Det ble skrevet at universitetet kunne “eventuelt” drifte begge de to sistnevnte løsningene men at den åpne bør foretrekkes med mindre noen spesielle strategiske hensyn styrer valget mot den proprietære, og om det skulle være tilfellet, så ville implementeringen innebære en betydelig arbeidsmengde for institusjonens IT-organisasjon.

Man kan vel komme med kritikk til måten selve prosessen ble utført, men etter å ha lest rapporten ville man trodd at i en offentlig organisasjon som ofte sliter med å dekke alle behov med tilstrekkelige midler, ville man valgt løsningen som fortsetter “tradisjonen” for åpne løsninger og ikke belastet organisasjonen med “ressurskrevende innføring” og andre ulemper. Prosjektets konklusjonen, derimot, var at universitetet skulle innføre Microsoft Exchange – den proprietære løsningen – og dermed skifte ut vesentlige deler av institusjonens infrastruktur for e-post, overføre lagrede meldinger til den proprietære Exchange løsningen, og flytte brukere til nye programvarer og systemer.

Man begynner å få en følelse at på et eller annet sted i ledelseshierarkiet noen har sagt “jeg må ha”, og ettersom ingen har nektet dem noe de “måtte ha” tidligere i livet, så har de fått det de “måtte ha” i denne situasjonen også. Man får også en følelse at rapporten i sitt endelige utkast ble formulert slik at beslutningstakerne raskt kunne avfeie ulempene og risikoene og så få “grønt lys” for et valg de ville ta helt på begynnelsen av prosessen: når det står i “svart på hvitt” at innføringen av den forhåndsutvalgte løsningen er gjennomførbar, føler de at de har alt de trenger for å bare sette i gang, koster det som det bare må.

For Nyhets Skyld

Men tilbake til prosessen. Det første som forbauset meg med prosessen var at man skulle vurdere implementering av helt nye systemer i det hele tatt. Prosessen fikk sin opprinnelse i et tilsynelatende behov for en “integrert” e-post- og kalenderløsning: noe som anses som ganske uviktig hos mange, men man kan ikke nekte for at noen ville opplevd en slik integrasjon av tjenester som nyttig. Men oppgaven å integrere tjenester trenger ikke å innebære en innføring av kun ett stort system som skal dekke alle funksjonelle områder og som skal erstatte alle eksisterende systemer som hadde noe å gjøre med tjenestene som integreres: en slik forenklet oppfatning av hvordan teknologiske systemer fungerer ville føre til at man insisterer at hele Internettet drives fra kun én stormaskin eller på kun én teknologisk plattform; alle som har satt seg inn i hvordan nettet fungerer vet at det ikke er slik (selv om noen organisasjoner ville foretrukket at alle kunne overvåkes på ett sted).

Det kan hende at beslutningstakerne faktisk tror på utilstrekkelige og overforenklede modeller av hvordan teknologi anvendes, eller at de velger bare å avfeie virkeligheten med en “få det gjort” mentalitet som lett kobles sammen med vrangforestillingen at ett produkt og én leverandør er som regel “løsningen”. Men infrastruktur i de fleste organisasjonene vil alltid bestå av forskjellige systemer, ofte med forskjellige opphav, som ofte fungerer sammen for å levere det som virker som kun én løsning eller tjeneste til utenforstående. Akkurat dette har faktisk vært situasjonen inne i universitetets infrastruktur for e-post frem til nå. Hvis den nåværende infrastrukturen mangler en forbindelse mellom e-postmeldinger og kalenderdata, hvorfor kan man ikke legge til en ny komponent eller tjeneste for å realisere den manglende integrasjonen som folk savner?

Og så blir det interessant å se nærmere på løsninger som likner på den eksisterende infrastrukturen universitetet bruker, og på produkter og prosjekter som leverer den savnede delen av infrastrukturen. Hvis det finnes liknende løsninger og tilsvarende infrastrukturbeskrivelser, spesielt om de finnes som fri programvare akkurat som programvarene universitetet allerede bruker, og hvis avstanden mellom det som kjøres nå og en fremtidig “fullstendig” løsning består bare av noen tilleggskomponenter og litt arbeid, ville det ikke vært interessant å se litt nærmere på slike ting først?

Gruppevare: Et Perspektiv

Jeg har vært interessert i gruppevare og kalenderløsninger i ganske lenge. For noen år siden utviklet jeg en nettleser-basert applikasjon for å behandle personlige kommunikasjoner og avtaler, og denne brukte eksisterende teknologier og standarder for å utveksle, fremvise og redigere informasjonen som ble lagret og behandlet. Selv om jeg etterhvert ble mindre overbevist på akkurat den måten jeg hadde valgt å implementere applikasjonen, hadde jeg likevel fått et innblikk i teknologiene som brukes og standardene som finnes for å utveksle gruppevaredata. Tross alt, uansett hvordan meldinger og kalenderinformasjon lagres og håndteres må man fortsatt forholde seg til andre programmer og systemer. Og etter at jeg ble mer opptatt av wikisystemer og fikk muligheten til å implementere en kalendertjeneste for en av de mest utbredte wikiløsningene, ble jeg oppmerksom på nytt på standardiseringsretninger, praktiske forhold i implementasjon av gruppevaresystemer, og ikke minst hva slags eksisterende løsninger som fantes.

Et gruppevaresystem jeg hadde hørt om for mange år siden var Kolab, som består av noen godt etablerte komponenter og programmer (som også er fri programvare), ettersom produktet og prosjektet bak ble grunnlagt for å styrke tilbudet på serversiden slik at klientprogrammer som Evolution og KMail kunne kommunisere med fullstendige gruppevaretjenester som også ville være fri programvare. Et slikt behov ble identifisert da personer og organisasjoner tilknyttet Kolab leverte en e-postløsning bygget på KMail (med kryptering som viktig element) til en del av den tyske stat. Hvorfor bruke fri programvare kun på klientsiden og dermed måtte tåle skiftende forhold og varierende støtte for åpne standarder og interoperabilitet på serversiden?

Nesten Framme

Når man ser på Kolab er det slående hvor mye løsningen har til felles med universitetets nåværende infrastruktur: begge bruker LDAP som autentiseringsgrunnlag, begge bruker felles antispam og antivirus teknologier, og Roundcube står ganske sentralt som “webmail” løsning. Selv om noen funksjoner leveres av forskjellige programvarer, når man sammenlikner Kolab med infrastrukturen til UiO, kan man likevel påstå at avstanden mellom de ulike komponentene som brukes på hver sin side ikke er altfor stor for at man enten kunne bytte ut en komponent til fordel for komponenten Kolab helst vil bruke, eller kunne tilpasse Kolab slik at den kunne bruke den eksisterende komponenten istedenfor sin egen foretrukne komponent. Begge måter å håndtere slike forskjeller kunne benytte dokumentasjonen og ekspertisen som finnes på en mengde steder på nettet og i andre former som bøker og – ikke overraskende – organisasjonens egne fagpersoner.

Man trenger dog egentlig ikke å “bytte” til Kolab i det hele tatt: man kunne ta i bruk delene som dekker manglene i den nåværende infrastrukturen, eller man kunne ta i bruk andre komponenter som kan utføre slike jobber. Poenget er at det finnes alternativer som ligger ganske nær infrastrukturen som brukes i dag, og når man må velge mellom en storslått gjenimplementering av den infrastrukturen eller en mindre oppgradering som sannsynligvis kommer til å levere omtrent det samme resultatet, burde man begrunne og dokumentere avgjørelsen om å ikke se nærmere på slike nærliggende løsninger istedenfor å bare droppe dem fra vurderingen i stillhet og håpe at ingen ville merket at de forsvant uten å bli omtalt i det hele tatt.

I et dokument som oppsummerer prosjektets arbeid står det nettopp at SOGo – den foretrukne åpne løsningen – opptrer som et ledd mellom systemer som allerede er i drift. Man kunne godt tenkt at en vurdering av Kolab ville informert vurderingen av SOGo, og omvendt, slik at forståelsen for en eventuell oppgraderingsløsning kunne blitt grundigere og bedre, og kanskje førte til at en blanding av de mest interessante elementene tas i bruk slik at organisasjonen får ut det beste fra begge (og enda flere) løsninger.

Gjennom Nåløyet

Som skrevet ovenfor er jeg interessert i gruppevare og bestemte meg for å oppdage hvorfor en løsning slik som Kolab (og liknende løsninger, selvsagt) kunne vært fraværende fra sluttrapporten. Selv om det ikke ville vært riktig å publisere samtalene mellom prosjektets talsperson og meg, kan jeg likevel oppsummere det jeg lærte gjennom meldingsutvekslingen: en enkel oversikt over hvilke løsninger som ble vurdert (bortsett fra de som omtales i rapporten) finnes trolig ikke, og prosjektdeltakerne kunne ikke med det første huske hvorfor Kolab ikke ble tatt med til sluttrapporten. Dermed, når rapporten avfeier alt som ikke blir beskrevet i sin tekst, som om det fantes en grundig prosess for å sive ut alt som ikke nådde opp til prosjektets behov, kan dette anses nesten som ren bløff: det finnes ingen bevis for at evalueringen av tilgjengelige løsninger var på noen som helst måte omfattende eller balansert.

Man kan vel hevde at det skader ingen å publisere en rapport på nettet som velger ut verdige systemer for en organisasjons gruppevarebehov – konklusjonene vil naturligvis dreie seg ganske mye om hva slags organisasjon det er som omtales, og man kan alltid lese gjennom teksten og bedømme kompetansen til forfatterne – men det er også tilfellet at andre i samme situasjon ofte vil bruke materialer som likner på det de selv må produsere som utgangspunkt for sitt eget evalueringsarbeid. Når kjente systemer utelates vil lesere kanskje konkludere at slike systemer måtte ha hatt grunnleggende mangler eller vært fullstendig uaktuelle: hvorfor ellers ville de ikke blitt tatt med? Det er flott at organisasjoner foretar slikt arbeid på en åpen måte, men man bør ikke undervurdere propagandaverdien som sitter i en rapport som dropper noen systemer uten begrunnelse slik at systemene som gjenstår kan tolkes som de aller beste eller de eneste relevante, og at andre systemer bare ikke når opp og dermed har ingen plass i slike vurderinger hos andre institusjoner.

Etter litt frem og tilbake ble jeg endelig fortalt at Kolab ikke hadde vist tilstrekkelige livstegn i utvikler- og brukersamfunnet rundt prosjektet, og det var grunnen til at programvaren ikke ble vurdert videre; eksempler trukket frem handlet om prosjektets dokumentasjon som ikke var godt nok oppdatert. Selv om prosjektet har forbedringspotensial i noen områder – jeg har selv foreslått at noe gjøres med wikitjenesten til prosjektet i perioden etter jeg avsluttet mine samtaler med universitetets prosjektdeltakere og begynte å sette meg inn i situasjonen angående fri programvare og gruppevare – må det sies at tilbakemeldingene virket litt forhastede og tilfeldige: e-postlistene til Kolab-prosjektet viser relativt mye aktivitet, og det finnes ikke noen tegn som tyder på at noen fra universitetet tok kontakt med prosjektets utviklersamfunn for å fordype seg i mulige forbedringer i dokumentasjon, fremtidige planer, og muligheter for å gjenbruke kompetanse og materialer for komponenter og systemer som Kolab har til felles med andre løsninger, blant annet de som brukes hos universitetet.

Enda merkeligere var kommentarer fra prosjektgruppen som handlet om Roundcube. Rapporten omtalte Roundcube som en tilstrekkelig løsning som ikke bare brukes frem til i dag hos universitetet, men også ville vært brukt som erstatning for webmail-løsningen i SOGo. Plutselig, ifølge prosjektets talsperson, var Roundcube ikke godt nok i ett navngitt område, men rapporten brukte ikke noen som helst plass om slike påståtte mangler: rart når man tenker at det ville helt sikkert vært en veldig god anledning til å beskrive slike mangler og så forenkle beslutningsgrunnlaget vesentlig. Det han hende at slike ting ble funnet ut i etterkant av rapportens publisering, men man får et inntrykk – begrunnet eller ikke – at slike ting også kan finnes opp i etterkant for å begrunne en avgjørelse som ikke nødvendigvis trenger å bli forankret i faktagrunnlaget.

(Jeg ble bedt om å ikke dele prosjektgruppens oppfatninger om Roundcube med andre i offentligheten, og selv om jeg foreslo at de tok opp de påståtte manglene direkte med Roundcube-prosjektet, er jeg ikke overbevist at de hadde noen som helst hensikt å gjøre det. Beleilig at man kan kritisere noe eller noen uten at de har muligheten til å forklare eller forsvare seg!)

Og til slutt, brukte rapporten en god del plass for å beskrive proprietære Exchange-teknologier og hvordan man kunne få andre systemer til å bruke dem, samtidig som åpne og frie løsninger måtte tilpasses Outlook og det produktets avhengighet på proprietær kommunikasjon for å fungere med alle funksjoner slått på. Relativt lite plass og prioritering ble tildelt alternative klienter. Til tross for bekymringer om Thunderbird – den foretrukne e-postklienten frem til nå – og hvordan den skal utvides med kalenderfunksjon, har jeg aldri sett Kontact eller KMail nevnt en eneste gang, ikke en gang i Linux-sammenheng til tross for at Kontact har vært tilgjengelig i universitetets påbudte Linux-distribusjon – Red Hat Enterprise Linux – i årevis og fungerer helt greit med e-postsystemene som nå skal vrakes. Det kan hende at folk anser Kontact som gammeldags – en holdning jeg oppdaget på IRC for noen uker siden uten at den kunne begrunnes videre eller med mer substans – men den er en moden e-postklient med kalenderfunksjon som har fungert for mange over lengre tid. Pussig at ingen i vurderingsarbeidet vil nevne denne programvaren.

Må Bare Ha

Det begynner med en organisasjon som har godt fungerende systemer som kunne bygges på slik at nye behov tilfredsstilles. Imidlertid, ble det kjørt en prosess som ser ut til å ha begynt med forutsetningen at ingenting kan gjøres med disse godt fungerende systemene, og at de må skiftes ut med “ressurskrevende innføring” til fordel for et proprietært system som skal skape en dypere avhengighet på en beryktet leverandør (og monopolist). Og for å styrke denne tvilsomme forutsetningen ble kjente løsninger utelatt fra vurderingene som ble gjort, tilsynelatende slik at en enklere og “passende” avgjørelse kunne tas.

Inntrykket som blir igjen tyder på at prosessen ikke nødvendigvis ble brukt til å informere avgjørelser, men at avgjørelser informerte eller dirigerte prosessen. Hvorfor dette kan ha skjedd er kanskje en historie for en annen gang: en historie som handler om ulike perspektiver i forhold til etikk, demokrati, og investeringer i kunnskap og kompetanse. Og som man kunne forvente ut ifra det som er skrevet ovenfor, er det ikke nødvendigvis en historie som setter organisasjonens “dirigenter” i et så veldig godt lys.

Når man fokuserer på merkevarer og ikke standarder

Friday, January 24th, 2014

Det var interessant å lese et leserbrev i Uniforum med tittelen “Digitale samarbeidsverktøy i undervisning“: noe som trykker alle de riktige knappene i dagens akademiske landskap med et økende fokus på nettbasert opplæring, bredere tilgang til kurs, og mange andre ting som kanskje motiveres mer av “prestisje” og “å kapre kunder” enn å øke generell tilgjengelighet til en institusjons kunnskap og ekspertise. Brevforfatterne beskriver utbredt bruk av videokonferanseløsninger og skryter på en proprietær løsning som de tilsynelatende liker ganske godt, og henviser til universitetets anbefalinger for tekniske løsninger. Etter at man graver litt til å finne disse anbefalingene ser man ganske fort at de dreier seg om tre proprietære (eller delvis proprietære) systemer: Adobe Connect, Skype, og spesialiserte videokonferanseutstyr som finnes i noen møterom.

Det er i all sannsynlighet en konsekvens av forbrukersamfunnet vi lever i: at man tenker produkter og merkevarer og ikke standarder, og at etter man oppdager et spennende produkt vil man gjerne øke oppmerksomheten på produktet blant alle som har liknende behov. Men når man blir produkt- og merkevarefokusert kan man fort miste blikket over det egentlige problemet og den egentlige løsningen. At så mange fortsetter å insistere på Ibux når de kunne kjøpe generiske medikamenter for en brøkdel av merkevarens pris er bare et eksempel på at folk ikke lenger er vant til å vurdere de virkelige forholdene og vil heller ty til merkevarer for en enkel og rask løsning de ikke trenger å tenke så mye på.

Man bør kanskje ikke legge så veldig mye skyld på vanlige databrukere når de begår slike feil, spesielt når store deler av det offentlige her i landet fokuserer på proprietære produkter som, hvis de utnytter genuine standarder i det hele tatt, pleier å blande dem med proprietære teknologier for å styre kunden mot en forpliktelse overfor noen få leverandører i en nesten ubegrenset tid fremover. Men det er litt skuffende at “grønne” representanter ikke vurderer bærekraftige teknologiske løsninger – de som gjøres tilgjengelig som fri programvare og som bruker dokumenterte og åpne standarder – når det ville forventes at slike representanter foreslår bærekraftige løsninger i andre områder.

The Organisational Panic Button and the Magic Single Vendor Delusion

Wednesday, November 27th, 2013

I have had reason to consider the way organisations make technology choices in recent months, particularly where the public sector is concerned, and although my conclusions may not come as a surprise to some people, I think they sum up fairly well how bad decisions get made even if the intentions behind them are supposedly good ones. Approaching such matters from a technological point of view, being informed about things like interoperability, systems diversity, the way people adopt and use technology, and the details of how various technologies work, it can be easy to forget that decisions around acquisitions and strategies are often taken by people who have no appreciation of such things and no time or inclination to consider them either: as far as decision makers are concerned, such things are mere details that obscure the dramatic solution that shows them off as dynamic leaders getting things done.

Assuming the Position

So, assume for a moment that you are a decision-maker with decisions to make about technology, that you have in your organisation some problems that may or may not have technology as their root cause, and that because you claim to listen to what people in your organisation have to say about their workplace, you feel that clear and decisive measures are required to solve some of those problems. First of all, it is important to make sure that when people complain about something, they are not mixing that thing up with something else that really makes their life awkward, but let us assume that you and your advisers are aware of that issue and are good at getting to the heart of the real problem, whatever that may be. Next, people may ask for all sorts of things that they want but do not actually need – “an iPad in every meeting room, elevator and lavatory cubicle!” – and even if you also like the sound of such wild ideas, you also need to be able to restrain yourself and to acknowledge that it would simply be imprudent to indulge every whim of the workforce (or your own). After all, neither they nor you are royalty!

With distractions out of the way, you can now focus on the real problems. But remember: as an executive with no time for detail, the nuances of a supposedly technological problem – things like why people really struggle with some task in their workplace and what technical issues might be contributing to this discomfort – these things are distractions, too. As someone who has to decide a lot of things, you want short and simple summaries and to give short and simple remedies, delegating to other people to flesh out the details and to make things happen. People might try and get you to understand the detail, but you can always delegate the task of entertaining such explanations and representations to other people, naturally telling them not to waste too much time on executing the plan.

Architectural ornamentation in central Oslo

Architectural ornamentation in central Oslo

On the Wrong Foot

So, let us just consider what we now know (or at least suspect) about the behaviour of someone in an executive position who has an organisation-wide problem to solve. They need to demonstrate leadership, vision and intent, certainly: it is worth remembering that such positions are inherently political, and if there is anything we should all know about politics by now, it is that it is often far more attractive to make one’s mark, define one’s legacy, fulfil one’s vision, reserve one’s place in the history books than it is to just keep things running efficiently and smoothly and to keep people generally satisfied with their lot in life; this principle alone explains why the city of Oslo is so infatuated with prestige projects and wants to host the Winter Olympics in a few years’ time (presumably things like functioning public transport, education, healthcare, even an electoral process that does not almost deliberately disenfranchise entire groups of voters, will all be faultless by then). It is far more exciting being a politician if you can continually announce exciting things, leaving the non-visionary stuff to your staff.

Executives also like to keep things as uncluttered as possible, even if the very nature of a problem is complicated, and at their level in the organisation they want the explanations and the directives to be as simple as possible. Again, this probably explains the “rip it up and start over” mentality that one sees in government, especially after changes in government even if consecutive governments have ideological similarities: it is far better to be seen to be different and bold than to be associated with your discredited predecessors.

But what do these traits lead to? Well, let us return to an organisational problem with complicated technical underpinnings. Naturally, decision-makers at the highest levels will not want to be bored with the complications – at the classic “10000 foot” view, nothing should be allowed to encroach on the elegant clarity of the decision – and even the consideration of those complications may be discouraged amongst those tasked to implement the solution. Such complications may be regarded as a legacy of an untidy and unruly past that was not properly governed or supervised (and are thus mere symptoms of an underlying malaise that must be dealt with), and the need to consider them may draw time and resources away from an “urgently needed” solution that deals with the issue no matter what it takes.

How many times have we been told “not to spend too much time” on something? And yet, that thing may need to be treated thoroughly so that it does not recur over and over again. And as too many people have come to realise or experience, blame very often travels through delegation: people given a task to see through are often deprived of resources to do it adequately, but this will not shield them from recriminations and reprisals afterwards.

It should not demand too much imagination to realise that certain important things will be sacrificed or ignored within such a decision-making framework. Executives will seek simplistic solutions that almost favour an ignorance of the actual problem at hand. Meanwhile, the minions or underlings doing the work may seek to stay as close as possible to the exact word of the directive handed down to them from on high, abandoning any objective assessment of the problem domain, so as to be able to say if or when things go wrong that they were only following the instructions given to them, and that as everything falls to pieces it was the very nature of the vision that led to its demise rather than the work they did, or that they took the initiative to do something “unsanctioned” themselves.

The Magic Single Vendor Temptation

We can already see that an appreciation of the finer points of a problem will be an early casualty in the flawed framework described above, but when pressure also exists to “just do something” and when possible tendencies to “make one’s mark” lie just below the surface, decision-makers also do things like ignore the best advice available to them, choosing instead to just go over the heads of the people they employ to have opinions about matters of technology. Such antics are not uncommon: there must be thousands or even millions of people with the experience of seeing consultants breeze into their workplace and impart opinions about the work being done that are supposedly more accurate, insightful and valuable than the actual experiences of the people paid to do that very work. But sometimes hubris can get the better of the decision-maker to the extent that their own experiences are somehow more valid than those supposed experts on the payroll who cannot seem to make up their minds about something as mundane as which technology to use.

And so, the executive may be tempted to take a page from their own playbook: maybe they used a product in one of their previous organisations that had something to do with the problem area; maybe they know someone in their peer group who has an opinion on the topic; maybe they can also show that they “know about these things” by choosing such a product. And with so many areas of life now effectively remedied by going and buying a product that instantly eradicates any deficiency, need, shortcoming or desire, why would this not work for some organisational problem? “What do you mean ‘network provisioning problems’? I can get the Internet on my phone! Just tell everybody to do that!”

When the tendency to avoid complexity meets the apparent simplicity of consumerism (and of solutions encountered in their final form in the executive’s previous endeavours), the temptation to solve a problem at a single stroke or a single click of the “buy” button becomes great indeed. So what if everyone affected by the decision has different needs? The product will surely meet all those needs: the vendor will make sure of that. And if the vendor cannot deliver, then perhaps those people should reconsider their needs. “I’ve seen this product work perfectly elsewhere. Why do you people have to be so awkward?” After all, the vendor can work magic: the salespeople practically told us so!

Nothing wrong here: a public transport "real time" system failure; all the trains are arriving "now"

Nothing wrong here: a public transport "real time" system failure; all the trains are arriving "now"

The Threat to Diversity

In those courses in my computer science degree that dealt with the implementation of solutions at the organisational level, as opposed to the actual implementation of software, attempts were made to impress upon us students the need to consider the requirements of any given problem domain because any solution that neglects the realities of the problem domain will struggle with acceptance and flirt with failure. Thus, the impatient executive approach involving the single vendor and their magic product that “does it all” and “solves the problem” flirts openly and readily with failure.

Technological diversity within an organisation frequently exists for good reason, not to irritate decision-makers and their helpers, and the larger the organisation the larger the potential diversity to be found. Extrapolating from narrow experiences – insisting that a solution must be good enough for everyone because “it is good enough for my people” – risks neglecting the needs of large sections of an organisation and denying the benefits of diversity within the organisation. In turn, this risks the health of those parts of an organisation whose needs have now been ignored.

But diversity goes beyond what people happen to be using to do their job right now. By maintaining the basis for diversity within an organisation, it remains possible to retain the freedom for people to choose the most appropriate systems and platforms for their work. Conversely, undermining diversity by imposing a single vendor solution on everyone, especially when such solutions also neglect open standards and interoperability, threatens the ability for people to make choices central to their own work, and thus threatens the vitality of that work itself.

Stories abound of people in technical disciplines who “also had to have a Windows computer” to do administrative chores like fill out their expenses, hours, travel claims, and all the peripheral tasks in a workplace, even though they used a functioning workstation or other computer that would have been adequate to perform the same chores within a framework that might actually have upheld interoperability and choice. Who pays for all these extra computers, and who benefits from such redundancy? And when some bright spark in the administration suggests throwing away the “special” workstation, putting administrative chores above the real work, what damage does this do to the working environment, to productivity, and to the capabilities of the organisation?

Moreover, the threat to diversity is more serious than many people presumably understand. Any single vendor solution imposed across an organisation also threatens the independence of the institution when that solution also informs and dictates the terms under which other solutions are acquired and introduced. Any decision-maker who regards their “one product for everybody” solution as adequate in one area may find themselves supporting a “one vendor for everything” policy that infects every aspect of the organisation’s existence, especially if they are deluded enough to think that they getting a “good deal” by buying all their things from that one vendor and thus unquestioningly going along with it all for “economic reasons”. At that point, one has to wonder whether the organisation itself is in control of its own acquisitions, systems or strategies any longer.

Somebody Else’s Problem

People may find it hard to get worked up about the tools and systems their employer uses. Surely, they think, what people have chosen to run a part of the organisation is a matter only for those who work with that specific thing from one day to the next. When other people complain about such matters, it is easy to marginalise them and to accuse them of making trouble for the sake of doing so. But such reactions are short-sighted: when other people’s tools are being torn out and replaced by something less than desirable, bystanders may not feel any urgency to react or even think about showing any sympathy at all, but when tendencies exist to tackle other parts of an organisation with simplistic rationalisation exercises, who knows whose tools might be the next ones to be tampered with?

And from what we know from unfriendly solutions that shun interoperability and that prefer other solutions from the same vendor (or that vendor’s special partners), when one person’s tool or system gets the single vendor treatment, it is not necessarily only that person who will be affected: suddenly, other people who need to exchange information with that person may find themselves having to “upgrade” to a different set of tools that are now required for them just to be able to continue that exchange. One person’s loss of control may mean that many people lose control of their working environment, too. The domino effect that follows may result in an organisation transformed for the worse based only on the uninformed gut instincts of someone with the power to demand that something be done the apparently easy way.

Inconvenience: a crane operating over one pavement while sitting on the other, with a sign reading "please use the pavement on the other side"

Inconvenience: a crane operating over one pavement while sitting on the other, with a sign reading "please use the pavement on the other side"

Getting the Message Across

For those of us who want to see Free Software and open standards in organisations, the dangers of the top-down single vendor strategy are obvious, but other people may find it difficult to relate to the issues. There are, however, analogies that can be illustrative, and as I perused a publication related to my former employer I came across an interesting complaint that happens to nicely complement an analogy I had been considering for a while. The complaint in question is about some supplier management software that insists that bank account numbers can only have 18 digits at most, but this fails to consider the situation where payments to Russian and Chinese accounts might need account numbers with more than 18 digits, and the complainant vents his frustration at “the new super-elite of decision makers” who have decided that they know better than the people actually doing the work.

If that “super-elite” were to call all the shots, their solution would surely involve making everyone get an account with an account number that could only ever have 18 digits. “Not supported by your bank? Change bank! Not supported in your country? Change your banking country!” They might not stop there, either: why not just insist on everyone having an account at just one organisation-mandated bank? “Who cares if you don’t want a customer relationship with another bank? You want to get paid, don’t you?”

At one former employer of mine, setting up a special account at a particular bank was actually how things were done, but ignoring peculiarities related to the nature of certain kinds of institutions, making everyone needlessly conform through some dubiously justified, executive-imposed initiative whether it be requiring them to have an account with the organisation’s bank, or requiring them to use only certain vendor-sanctioned software (and as a consequence requiring them to buy certain vendor-sanctioned products so that they may have a chance of using them at work or to interact with their workplace from home) is an imposition too far. Rationalisation is a powerful argument for shaking things up, but it is often used by those who do not care how much it manages to transfer the inconvenience in an organisation to the individual and to other parties.

Bearing the Costs

We have seen how the organisational cost of short-sighted, buy-and-forget decision-making can end up being borne by those whose interests have been ignored or marginalised “for the good of the organisation”, and we can see how this can very easily impose costs across the whole organisation, too. But another aspect of this way of deciding things can also be costly: in the hurry to demonstrate the banishment of an organisational problem with a flourish, incremental solutions that might have dealt with the problem more effectively can become as marginalised as the influence of the people tasked with the job of seeing any eventual solution through. When people are loudly demanding improvements and solutions, an equally dramatic response probably does not involve reviewing the existing infrastructure, identifying areas that can provide significant improvement without significant inconvenience or significant additional costs, and committing to improve the existing solutions quietly and effectively.

Thus, when faced with disillusionment – that people may have decided for themselves that whatever it was that they did not like is now beyond redemption – decision-makers are apt to pander to such disillusionment by replacing any existing thing with something completely new. Especially if it reinforces their own blinkered view of an organisational problem or “confirms” what they “already know”, decision-makers may gladly embrace such dramatic acts as a demonstration of the resolve expected of a decisive leader as they stand to look good by visibly banishing the source of disillusionment. But when such pandering neglects relatively inexpensive, incremental improvements and instead incurs significant costs and disruptions for the organisation, one can justifiably question the motivations behind such dramatic acts and the level of competence brought to bear on resolving the original source of discomfort.

The electrical waste collection

The electrical waste collection

Mission Accomplished?

Thinking that putting down money with a single vendor will solve everybody’s problems, purging diversity from an organisation and stipulating the uniformity encouraged by that vendor, is an overly simplistic and even deluded approach to organisational change. Change in any organisation can be very expensive and must therefore be managed carefully. Change for the sake of change is therefore incredibly irresponsible. And change imposed to gratify the perception of change or progress, made on a superficial basis and incurring unnecessary and avoidable burdens within an organisation whilst risking that organisation’s independence and viability, is nothing other than indefensible.

Be wary of the “single vendor fixes it all” delusion, especially if all the signs point to a decision made at the highest levels of your organisation: it is the sign of the organisational panic button being pressed while someone declares “Mission Accomplished!” Because at the same time they will be thinking “We will have progress whatever the cost!” And you, not them, will be the one bearing the cost.

Students: Beware of the Academic Cloud!

Sunday, July 21st, 2013

Things were certainly different when I started my university degree many years ago. For a start, institutions of higher education provided for many people their first encounter with the Internet, although for those of us in the United Kingdom, the very first encounter with wide area networking may well have involved X.25 and PAD terminals, and instead of getting a “proper” Internet e-mail address, it may have been the case that an address only worked within a particular institution. (My fellow students quickly became aware of an Internet mail gateway in our department and the possibility, at least, of sending Internet mail, however.)

These days, students beginning their university studies have probably already been using the Internet for most of their lives, will have had at least one e-mail address as well as accounts for other online services, may be publishing blog entries and Web pages, and maybe even have their own Web applications accessible on the Internet. For them, arriving at a university is not about learning about new kinds of services and new ways of communicating and collaborating: it is about incorporating yet more ways of working online into their existing habits and practices.

So what should students expect from their university in terms of services? Well, if things have not changed that much over the years, they probably need a means of communicating with the administration, their lecturers and their fellow students, along with some kind of environment to help them do their work and provide things like file storage space and the tools they won’t necessarily be able to provide themselves. Of course, students are more likely to have their own laptop computer (or even a tablet) these days, and it is entirely possible that they could use that for all their work, subject to the availability of tools for their particular course, and since they will already be communicating with others on the Internet, communicating with people in the university is not really anything different from what they are already doing. But still, there are good reasons for providing infrastructure for students to use, even if those students do end up working from their laptops, uploading assignments when they are done, and forwarding their mail to their personal accounts.

The Student Starter Kit

First and foremost, a university e-mail account is something that can act as an official communications channel. One could certainly get away with using some other account, perhaps provided by a free online service like Google or Yahoo, but if something went wrong somewhere – the account gets taken over by an imposter and then gets shut down, for example – that channel of communication gets closed and important information may be lost.

The matter of how students carry out their work is also important. In computer science, where my experiences come from and where computer usage is central to the course, it is necessary to have access to suitable tools to undertake assignments. As everyone who has used technology knows, the matter of “setting stuff up” is one that often demands plenty of time and distracts from the task at hand, and when running a course that requires the participants to install programs before they can make use of the learning materials, considerable amounts of time are wasted on installing programs and troubleshooting. Thus, providing a ready-to-use environment allows students to concentrate on their work and to more easily relate to the learning materials.

There is the matter of the nature of teaching environments and the tools chosen. Teaching environments also allow students to become familiar with desirable practices when finding solutions to the problems in their assignments. In software engineering, for example, the use of version control software encourages a more controlled and rational way of refining a program under development. Although the process itself may not be recognised and rewarded when an assignment is assessed, it allows students to see how things should be done and to take full advantage of the opportunity to learn provided by the institution.

Naturally, it should be regarded as highly undesirable to train students to use specific solutions provided by selected vendors, as opposed to educating them to become familiar with the overarching concepts and techniques of a particular field. Schools and universities are not vocational training institutions, and they should seek to provide their students with transferable skills and knowledge that can be applied generally, instead of taking the easy way out and training those students to perform repetitive tasks in “popular” software that gives them no awareness of why they are doing those things or any indication that the rest of the world might be doing them in other ways.

Construction of the IT department's new building, University of Oslo

Minority Rule

So even if students arrive at their place of learning somewhat equipped to learn, communicate and do their work, there may still be a need for a structured environment to be provided for them. At that place of learning they will join those employed there who already have a structured environment in place to be able to do their work, whether that is research, teaching or administration. It makes a certain amount of sense for the students to join those other groups in the infrastructure already provided. Indeed, considering the numbers of people involved, with the students outnumbering all other groups put together by a substantial margin, one might think that the needs of the students would come first. Sadly, things do not always work that way.

First of all, students are only ever “passing through”. While some university employees may be retained for similar lengths of time – especially researchers and others on temporary contracts (a known problem with social and ethical dimensions of its own) – others may end up spending most of their working life there. As a result, the infrastructure is likely to favour such people over time as their demands are made known year after year, with any discomfort demanding to be remedied by a comfortable environment that helps those people do their work. Not that there is anything wrong with providing employees with a decent working environment: employers should probably do even more to uphold their commitments in this regard.

But when the demands and priorities of a relatively small group of people take precedence over what the majority – the students – need, one can argue that such demands and priorities have begun to subvert the very nature of the institution. Imposing restrictions on students or withholding facilities from them just to make life easier for the institution itself is surely a classic example of “the tail wagging the dog”. After all, without students and teaching an institution of higher education can no longer be considered a university.

Outsourcing Responsibility

With students showing up every year and with an obligation to provide services to them, one might imagine that an institution might take the opportunity to innovate and to evaluate ways in which it might stand out amongst the competition, truly looking after the group of people that in today’s increasingly commercialised education sector are considered the institution’s “customers”. When I was studying for my degree, the university’s mathematics department was in the process of developing computer-aided learning software for mathematics, which was regarded as a useful way of helping students improve their understanding of the course material through the application of knowledge recently acquired. However, such attempts to improve teaching quality are only likely to get substantial funding either through teaching-related programmes or by claiming some research benefit in the field of teaching or in another field. Consequently, developing software to benefit teaching is likely to be an activity located near the back of the queue for attention in a university, especially amongst universities whose leadership regard research commercialisation as their top priority.

So it becomes tempting for universities to minimise costs around student provision. Students are not meant to be sophisticated users whose demands must be met, mostly because they are not supposed to be around for long enough to be comfortable and for those providing services to eventually have to give in to student demands. Moreover, university employees are protected by workplace regulation (in theory, at least) whereas students are most likely protected by much weaker regulation. To take one example, whereas a university employee could probably demand appropriate accessibility measures for a disability they may have, students may have to work harder to get their disabilities recognised and their resulting needs addressed.

The Costs of Doing Business

So, with universities looking to minimise costs and to maximise revenue-generating opportunities, doing things like running infrastructure in a way that looks after the needs of the student and researcher populations seems like a distraction. University executives look to their counterparts in industry and see that outsourcing might offer a solution: why bother having people on the payroll when there are cloud computing services run by friendly corporations?

Let us take the most widely-deployed service, e-mail, as an example. Certainly, many students and employees might not be too concerned with logging into a cloud-based service to access their university e-mail – many may already be using such services for personal e-mail, and many may already be forwarding their university e-mail to their personal account – and although they might be irritated by the need to use one service when they have perhaps selected another for their personal use, a quick login, some adjustments to the mail forwarding settings, and logging out (never to return) might be the simple solution. The result: the institution supposedly saves money by making external organisations responsible for essential services, and the users get on with using those services in the ways they find most bearable, even if they barely take advantage of the specially designated services at all.

However, a few things may complicate this simplified scenario somewhat: reliability, interoperability, lock-in, and privacy. Reliability is perhaps the easiest to consider: if “Office 365” suddenly becomes “Office 360” for a few days, cloud-based services cannot be considered suitable for essential services, and if the “remedy” is to purchase infrastructure to bail out the cloud service provider, one has to question the decision to choose such an external provider in the first place. As for interoperability, if a user prefers Gmail, say, over some other hosted e-mail solution where that other solution doesn’t exchange messages properly with Gmail, that user will be in the awkward position of having to choose between a compromised experience in their preferred solution or an experience they regard as inconvenient or inferior. With services more exotic and less standardised than e-mail, the risk is that a user’s preferred services or software will not work with the chosen cloud-based service at all. Either way, users are forced to adopt services they dislike or otherwise have objections to using.

Miscellaneous waste

Product Placement

With users firmly parked on a specific vendor’s cloud-based platform, the temptation will naturally grow amongst some members of an organisation to “take advantage” of the other services on that platform whether they support interoperability or not. Users will be forced to log into the solution they would rather avoid or ignore in order to participate in processes and activities initiated by those who have actively embraced that platform. This is rather similar to the experience of getting a Microsoft Office document in an e-mail by someone requesting that one reads it and provides feedback, even though recipients may not have access to a compatible version of Microsoft Office or even run the software in question at all. In an organisational context, legitimate complaints about poor workflow, inappropriate tool use, and even plain unavailability of software (imagine being a software developer using a GNU/Linux or Unix workstation!) are often “steamrollered” by management and the worker is told to “deal with it”. Suddenly, everyone has to adapt to the tool choices of a few managerial product evangelists instead of participating in a standards-based infrastructure where the most important thing is just getting the work done.

We should not be surprised that vendors might be very enthusiastic to see organisations adopt both their traditional products as well as cloud-based platforms. Not only are the users exposed to such vendors’ products, often to the exclusion of any competing or alternative products, but by having to sign up for those vendors’ services, organisations are effectively recruiting customers for the vendor. Indeed, given the potential commercial benefits of recruiting new customers – in the academic context, that would be a new group of students every year – it is conceivable that vendors might offer discounts on products, waive the purchase prices, or even pay organisations in the form of services rendered to get access to new customers and increased revenue. Down the line, this pays off for the vendor: its organisational customers are truly locked in, cannot easily switch to other solutions, and end up paying through the nose to help the vendor recruit new customers.

How Much Are You Worth?

All of the above concerns are valid, but the most important one of all for many people is that of privacy. Now, most people have a complicated relationship with privacy: most people probably believe that they deserve to have a form of privacy, but at the same time many people are quite happy to be indiscreet if they think no-one else is watching or no-one else cares about what they are doing.

So, they are quite happy to share information about themselves (or content that they have created or acquired themselves) with a helpful provider of services on the Internet. After all, if that provider offers services that permit convenient ways of doing things that might be awkward to do otherwise, and especially if no money is changing hands, surely the meagre “payment” of tedious documents, mundane exchanges of messages, unremarkable images and videos, and so on, all with no apparently significant value or benefit to the provider, gets the customer a product worth far more in return. Everybody wins!

Well, there is always the matter of the small print – the terms of use, frequently verbose and convoluted – together with how other people perceive the status of all that content you’ve been sharing. As your content reaches others, some might perceive it as fair game for use in places you never could have imagined. Naturally, unintended use of images is no new phenomenon: I once saw one of my photographs being used in a school project (how I knew about it was that the student concerned had credited me, although they really should have asked me first, and an Internet search brought up their page in the results), whereas another photograph of mine was featured in a museum exhibition (where I was asked for permission, although the photograph was a lot less remarkable than the one found by the student).

One might argue that public sharing of images and other content is not really the same as sharing stuff over a closed channel like a social network, and so the possibility of unintended or undesirable use is diminished. But take another look at the terms of use: unlike just uploading a file to a Web site that you control, where nobody deems to claim any rights to what you are sharing, social networking and content sharing service providers frequently try and claim rights to your work.

Privacy on Parade

When everyone is seeking to use your content for their own goals, whether to promote their own businesses or to provide nice imagery to make their political advocacy more palatable, or indeed to support any number of potential and dubious endeavours that you may not agree with, it is understandable that you might want to be a bit more cautious about who wants a piece of your content and what they intend to do with it once they have it. Consequently, you might decide that you only want to deal with the companies and services you feel you can trust.

What does this have to do with students and the cloud? Well, unlike the services that a student may already be using when they arrive at university to start their studies, any services chosen by the institution will be imposed on the student, and they will be required to accept the terms of use of such services regardless of whether they agree with them or not. Now, although it might be said that the academic work of a student might be somewhat mundane and much the same as any other student’s work (even if this need not be the case), and that the nature of such work is firmly bound to the institution and it is therefore the institution’s place to decide how such work is used (even though this could be disputed), other aspects of that student’s activities and communications might be regarded as beyond the interests of the institution: who the student communicates with, what personal views they may express in such communications, what academic or professional views they may have.

One might claim that such trivia is of no interest to anyone, and certainly not to commercial entities who just want to sell advertising or gather demographic data or whatever supposedly harmless thing they might do with the mere usage of their services just to keep paying the bills and covering their overheads, but one should still be wary that information stored on some remote server in some distant country might somehow make its way to someone who might take a closer and not so benign interest in it. Indeed, the matter of the data residing in some environment beyond their control is enough for adopters of cloud computing to offer specially sanctioned exemptions and opt-outs. Maybe it is not so desirable that some foreign student writing about some controversial topic in their own country has their work floating around in the cloud, or as far as a university’s legal department is concerned, maybe it does not look so good if such information manages to wander into the wrong hands only for someone to ask the awkward question of why the information left the university’s own systems in the first place.

A leaflet for a tourist attraction in the Cambridge area

Excuses, Excuses

Cloud-based service providers are likely to respond to fears articulated about privacy violations and intrusions by insisting that such fears are disproportionate: that no-one is really looking at the data stored on their servers, that the data is encrypted somewhere/somehow, that if anything does look at the data it is merely an “algorithm” and not a person. Often these protests of innocence contradict each other, so that at any point in time there is at least one lie being told. But what if it is “only an algorithm” looking at your data? The algorithm will not be keeping its conclusions to itself.

How would you know what is really being done with your data? Not only is the code being run on a remote server, but with the most popular cloud services the important code is completely proprietary – service providers may claim to support Free Software and even contribute to it, but they do so only for part of their infrastructure – and you have no way of verifying any of their claims. Disturbingly, some companies want to enforce such practices within your home, too, so that when Microsoft claims that the camera on their most recent games console has to be active all the time but only for supposedly benign reasons and that the data is only studied by algorithms, the company will deny you the right to verify this for yourself. For all you know the image data could be uploaded somewhere, maybe only on command, and you would not only be none the wiser but you would also be denied the right to become wiser about the matter. And even if the images were not shared with mysterious servers, there are still unpleasant applications of “the algorithm”: it could, for example, count people’s faces and decide whether you were breaking the licensing conditions on a movie or other content by holding a “performance” that goes against the arbitrary licensing that accompanies a particular work.

Back in the world of the cloud, companies like Microsoft typically respond to criticism by pointing the finger at others. Through “shell” or “front” organisations the alleged faults of Microsoft’s competitors are brought to the attention of regulators, and in the case of the notorious FairSearch organisation, to take one example, the accusing finger is pointed at Google. We should all try and be aware of the misdeeds of corporations, that unscrupulous behaviour may have occurred, and we should insist that such behaviour be punished. But we should also not be distracted by the tactics of corporations that insist that all the problems reside elsewhere. “But Google!” is not a reason to stop scrutinising the record of a company shouting it out loud, nor is it an excuse for us to disregard any dubious behaviour previously committed by the company shouting it the loudest. (It is absurd that a company with a long history of being subject to scrutiny for anticompetitive practices – a recognised monopoly – should shout claims of monopoly so loudly, and it is even more absurd for anyone to take such claims at face value.)

We should be concerned about Google’s treatment of user privacy, but that should not diminish our concern about Microsoft’s treatment of user privacy. As it turns out, both companies – and several others – have some work to do to regain our trust.

I Do Not Agree

So why should students specifically be worried about all this? Does this not also apply to other groups, like anyone who is made to use software and services in their job? Certainly, this does affect more than just students, but students will probably be the first in line to be forced to accept these solutions or just not be able to take the courses they want at the institutions they want to attend. Even in countries with relatively large higher education sectors like the United Kingdom, it can be the case that certain courses are only available at a narrow selection of institutions, and if you consider a small country like Norway, it is entirely possible that some courses are only available at one institution. For students forced to choose a particular institution and to accept that institution’s own technological choices, the issue of their online privacy becomes urgent because such institutional changes are happening right now and the only way to work around them is to plan ahead and to make it immediately clear to those institutions that the abandonment of the online privacy rights (and other rights) of their “customers” is not acceptable.

Of course, none of this is much comfort to those working in private businesses whose technological choices are imposed on employees as a consequence of taking a job at such organisations. The only silver lining to this particular cloud is that the job market may offer more choices to job seekers – that they can try and identify responsible employers and that such alternatives exist in the first place – compared to students whose educational path may be constrained by course availability. Nevertheless, there exists a risk that both students and others may be confronted with having to accept undesirable conditions just to be able to get a study place or a job. It may be too disruptive to their lives not to “just live with it” and accept choices made on their behalf without their input.

But this brings up an interesting dilemma. Should a person be bound by terms of use and contracts where that person has been effectively coerced into accepting them? If their boss tells them that they have to have a Microsoft or Google account to view and edit some online document, and when they go to sign up they are presented with the usual terms that nobody can reasonably be expected to read, and given that they cannot reasonably refuse because their boss would then be annoyed at that person’s attitude (and may even be angry and threaten them with disciplinary action), can we not consider that when this person clicks on the “I agree” button it is only their employer who really agrees, and that this person not only does not necessarily agree but cannot be expected to agree, either?

Excuses from the Other Side

Recent events have probably made people wary of where their data goes and what happens with it once it has left their own computers, but merely being concerned and actually doing something are two different things. Individuals may feel helpless: all their friends use social networks and big name webmail services; withdrawing from the former means potential isolation, and withdrawing from the latter involves researching alternatives and trying to decide whether those alternatives can be trusted more than one of the big names. Certainly, those of us who develop and promote Free Software should be trying to provide trustworthy alternatives and giving less technologically-aware people the lifeline that they need to escape potentially exploitative services and yet maintain an active, social online experience. Not everyone is willing to sacrifice their privacy for shiny new online toys that supposedly need to rifle through your personal data to provide that shiny new online experience, nor is everyone likely to accept such constraints on their privacy when made aware of them. We should not merely assume that people do not care, would just accept such things, and thus do not need to be bothered with knowledge about such matters, either.

As we have already seen, individuals can make their own choices, but people in organisations are frequently denied such choices. This is where the excuses become more irrational and yet bring with them much more serious consequences. When an organisation chooses a solution from a vendor known to share sensitive information with other, potentially less friendly, parties, they might try and explain such reports away by claiming that such behaviour would never affect “business applications”, that such applications are completely separate from “consumer applications” (where surveillance is apparently acceptable, but no-one would openly admit to thinking this, of course), and that such a vendor would never jeopardise their relationship with customers because “as a customer we are important to them”.

But how do you know any of this? You cannot see what their online services are actually doing, who can access them secretly, whether people – unfriendly regimes, opportunistic law enforcement agencies, dishonest employees, privileged commercial partners of the vendor itself – actually do access your data, because how all that stuff is managed is secret and off-limits. You cannot easily inspect any software that such a vendor provides to you because it will be an inscrutable binary file, maybe even partially encrypted, and every attempt will have been made to forbid you from inspecting it both through licence agreements and legislation made at the request of exactly these vendors.

And how do you know that they value your business, that you are important to them? Is any business going to admit that no, they do not value your business, that you are just another trophy, that they share your private data with other entities? With documentation to the contrary exposing the lies necessary to preserve their reputation, how do you know you can believe anything they tell you at all?

Clouds over the IT building, University of Oslo

The Betrayal

It is all very well for an individual to make poor decisions based on wilful ignorance, but when an organisation makes poor decisions and then imposes them on other people for those people to suffer, this ignorance becomes negligence at the very least. In a university or other higher education institution, apparently at the bottom of the list of people to consult about anything, the bottoms on seats, the places to be filled, are the students: the first in line for whatever experiment or strategic misadventure is coming down the pipe of organisational reform, rationalisation, harmonisation, and all the other buzzwords that look great on the big screen in the boardroom.

Let us be clear: there is nothing inherently wrong with storing content on some network-accessible service, provided that the conditions under which that content is stored and accessed uphold the levels of control and privacy that we as the owners of that data demand, and where those we have elected to provide such services to us deserve our trust precisely by behaving in a trustworthy fashion. We may indeed select a service provider or vendor who respects us, rather than one whose terms and conditions are unfathomable and who treats its users and their data as commodities to be traded for profits and favours. It is this latter class of service providers and vendors – ones who have virtually pioneered the corrupted notion of the consumer cloud, with every user action recorded, tracked and analysed – that this article focuses on.

Students should very much beware of being sent into the cloud: they have little influence and make for a convenient group of experimental subjects, with few powerful allies to defend their rights. That does not mean that everyone else can feel secure, shielded by employee representatives, trade unions, industry organisations, politicians, and so on. Every group pushed into the cloud builds the pressure on every subsequent group until your own particular group is pressured, belittled and finally coerced into resignation. Maybe you might want to look into what your organisation is planning to do, to insist on privacy-preserving infrastructure, and to advocate Free Software as the only effective way of building that infrastructure.

And beware of the excuses – for the favourite vendor’s past behaviour, for the convenience of the cloud, for the improbability that any undesirable stuff could happen – because after the excuses, the belittlement of the opposition, comes the betrayal: the betrayal of sustainable and decentralised solutions, the betrayal of the development of local and institutional expertise, the betrayal of choice and real competition, and the betrayal of your right to privacy online.

Norwegian Voting and the Illusion of “Open Source”

Sunday, June 30th, 2013

I was interested to read about the new Norwegian electronic voting administration system, EVA, and a degree of controversy about whether the system can scale up to handle the number of votes expected in Oslo during the upcoming parliamentary elections. What I find more controversial is the claim that the system has been made available as “open source” software. In fact, a quick look at the licence is enough to confirm that the source code is really only available as “shared source”: something which people may be able to download and consult, but which withholds the freedoms that should be associated with open source software (or Free Software, as we should really call it).

So, here is why the usage of the term “open source” is dishonest in this particular case:

  • The rights offered to you (as opposed to the Norwegian authorities) cover only “testing, reviewing or evaluating the code”. (The authorities have geographical and usage restrictions placed on them.)
  • The licence restricts use to “non-commercial purposes”.
  • Anything else you might want to do requires you to get “written approval” from the vendors of the different parts of the system in question.
  • The software is encumbered by patents, but there is no patent grant.

Lost in Translation

In fairness, the page covering the source code does say the following about the use of the term “open source” which I have translated from the Norwegian:

When we use the term “open source”, we do not mean that the entire solution can be regarded as “free software”, meaning that you can download the software, inspect it, change it, and use it however you like. Anyone can download and inspect the source code, but it may only be used to carry out Norwegian elections. The solution should be available for research, however, and you are allowed to develop it further for such academic purposes.

Why translate from Norwegian when the English is underneath in the same page? Here’s why:

When we are using the term “open source code”, we do not say that the source code as a whole is what’s known as “open source”, that is code that can be freely downloaded, examined, changed and used. Anyone can download our source code, but it is only permittable to use it for elections in Norway. However, research on the solution is allowed and you are hence allowed to develop the system for an academic purpose.

In other words, while there are people involved who are clearly aware of what Free Software is, they apparently reserve the right to misuse the term “open source” to mean something other than what it was intended to mean. By consulting the Norwegian text first, I even give them the benefit of the doubt, but the contradiction of saying that something both is and is not something else should have awakened a degree of guilt that this could be regarded as deception.

Open Season

Now, one may argue that the Open Source Initiative should have been more aggressive in upholding its own brand and making sure that “open source” really does act as a guarantee of openness, but the “open” prefix is perhaps one of the most abused terms in the field of computing, and thus one might conclude that the OSI were fighting a losing battle from the outset. Of course, people might claim that the term “free software” is ambiguous or vague, which is why I prefer to write it as Free Software: the use of capital letters should at least get English readers unfamiliar with the term to wonder whether there is more involved than the mere juxtaposition of two widely-used words.

In Norwegian, “fri programvare” communicates something closer to what is meant by Free Software: the “fri” does not generally mean “for free” or “gratis”, but rather communicates a notion of freedom. I imagine that those who translated the text quoted above need to improve their terminology dictionary. Nevertheless, this does not excuse anyone who takes advantage of the potential ambiguity in the common sense perception of the term to promote something that has only superficial similarities to Free Software.

The Good, the Bad, and the Evry

I suppose we should welcome increased transparency in such important systems, and we should encourage more of the same in other areas. Nevertheless, with a substantial amount of activity in the field of electronic voting, particularly in the academic realm and in response to high-profile scandals (with some researchers even apparently experiencing persecution for their work), one has to wonder why the Norwegian government is not willing to work on genuinely open, Free Software electronic voting systems instead of partnering with commercial interests who, by advertising patent coverage, potentially threaten research in this area and thus obstruct societal progress, accountability and democracy.

One might suggest that the involvement of EDB ErgoGroup, now known as Evry, provides some answers to questions about how and where taxpayer money gets spent. With substantial state involvement in the company, either through the state-owned postal monopoly or the partly-state-owned incumbent telephone operator, one cannot help thinking that this is yet another attempt to funnel money to the usual beneficiaries of public contracts and to pick winners in an international market for electronic voting systems. Maybe, with voting problems, banking system problems and the resulting customer dissatisfaction, public departments and ministries think that this vendor needs all the help it can get, despite being appointed to a dominant position in the nation’s infrastructure and technology industry to the point of it all being rather anti-competitive. When the financial sector starts talking of monopolies, maybe the responsible adults need to intervene to bring a degree of proper functioning to the marketplace.

But regardless of who became involved and their own rewards for doing so, the Norwegian ministry responsible for this deception should be ashamed of itself! How about genuinely participating in and advancing the research and development of voting systems that everyone on the planet can freely use instead of issuing dishonest press releases and patronising accounts of how Norway doesn’t really need to learn from anyone? Because, by releasing patent-encumbered “shared source” and calling it “open source”, some decision makers and their communications staff clearly need lessons in at least one area: telling the truth.

An Aside

I also find it distasteful that the documentation hosted on the government’s electoral site has adverts for proprietary software embedded in it, but I doubt that those working in the Microsoft monoculture even notice their presence. These people may be using taxpayer money to go shopping for proprietary products, but I do not see why they should then be advertising those products to us as well using our money.