Monthly Archives: January 2011

Lower the barriers to entry

Getting new users actively involved in your open source project is one of the most important aspects of community development. A healthy open source project welcomes new contributors of all kinds and makes it easy for them to contribute. Prospective contributors feel welcome and are guided towards their first contribution, whatever their skills are. My OSS Watch colleague Steve Lee pointed out the website of LibreOffice; they managed to do this very well.

Care for your users

It’s very unlikely that any new contributor to the project will not have used the software before starting to contribute. Getting people to contribute therefore starts with helping your users. LibreOffice has a lot of channels to assist users in getting started and getting the most out of the software, ranging from mailing lists to IRC channels, FAQ and documentation. Having a dedicated and clearly marked ‘Get Help’ section on your website will help new users to get going and feel welcome.

It’s not all about code

As we’ve discussed before, contributing to an open source project is about much more than just writing code. Make sure you highlight all the different ways in which your users can get involved in your project. LibreOffice provides a nice example of how to do this. All of their documentation carry a Creative Commons licence, so don’t be afraid to copy from their pages, as long as you make sure you acknowledge them and stick a compatible CC licence on your documentation as well.

Lower the barriers to entry

So let’s say you would like to get more developers on your project (and you should!). It can be quite difficult for developers to know what they can do to help out. Having that clearly marked out is therefore important. A first code contribution is usually an easy hack like removing dead code or adding a unit test. Some people even argue that you should leave some of those easy hacks in, just to tempt potential contributors to fix it! If you can, list the easy hacks on your wiki, like the LibreOffice project has done.

But my project is really small!

Especially when your project is small, it’s important to involve your users and get more contributors. You may not have all the channels (like mailling lists and forums and IRC and FAQs) available from the start, but you can improve this gradually. For example, if your project is hosted on GoogleCode or SourceForge, it is really easy to set up a wiki so you have the framework in place. That is an easy win and will encourage users to contribute. And don’t feel afraid to ask people to contribute. If they are providing a good solution to a common problem on a mailing list, ask them if they’d put it on the wiki.

So check out the LibreOffice website for inspiration, and while you’re there, make sure you download their software. They launched their first release today!

Open Source: we are buying, but are we engaging?

We recently completed our 2010 national survey of open source in the UK academic sector (the full report is currently being finalised). This post examines the extent to which suppliers to, and staff of, UK universities and colleges contribute to open source as a matter of course. This kind of engagement is important, since it realises the maximum benefits of open source software.

Our 2010 survey has shown that there has been another significant increase in the number of organisations with a policy to consider open source solutions during procurement. This is due to a combination of factors, such as the government’s open source action plan, our own work here at OSS Watch and the sector’s success with open source Virtual Learning Environments (Moodle has once again taken a significant share from closed source Blackboard/WebCT).

Alongside this increase in open source friendly policies, we are seeing a smaller, but still significant, increase in the amount of open source in use within these institutions. However, the balance is still firmly with closed source solutions. The reasons for rejecting open source remain fairly consistent with previous years’ results. The top three reasons for rejecting open source are lack of support, interoperability/migration problems and lack of staff expertise.

This is a pattern that is also visible in non-academic sectors. Put simply, people are becoming increasingly comfortable with open source as a viable way to develop software, although there is still room for improvement where suppliers are concerned. As a result, new companies are springing up to fill the gap and  venture captialists are happy to support them as shown by the increasing level of investment in open source related companies.

All this is very comforting. It means that as we move forwards, universities and colleges will be able to satisfy more of their needs using open source supplied and supported by appropriate third parties. However, in some cases, particularly for Higher Education, where IT departments can be large, a reliance on external providers may be a limiting factor when considering the benefits of open source beyond lower licence fees.

One of the key advantages of open source is that enhancements can be made to the software and then submitted upstream. A good open source company will manage this process for you, but in these circumstances the customer remains insulated from the core product in much the same way that they are with closed source software.

This is not necessarily a problem and can certainly bring the benefit of reduced costs that result from a more efficient development process. However, it often makes sense for staff to become actively engaged in the projects, particularly where the product has been modified for local use.

UK universities and colleges already innovate around teaching and research methods. Organisations like the JISC exist to manage investment of public money for such innovation. Open source provides a means for these innovations to find their way directly into the software in use and thus increase the return on investment in those solutions by making them available to all.

Whilst we are seeing an increase in open source friendly policies and actual use, we are not seeing a similarly strong growth in universities’ policies relating to staff engagement with open source software development. Where local innovation is taking place, it makes sense to actively engage with the open source community. For example, Stuart Lee, director of Oxford University Computing Services, told audiences at the recent FutureCampus event in Kuala Lumpur: “We have to develop our own systems off the shelf for things like tutorial recording.” Being able to integrate those systems with other software in use at the institution is critical to their success; this is what Oxford is doing with its Sakai Virtual Learning Environment and it is certainly reaping the rewards.

Ensuring that local innovations are closely embedded with the upstream project helps ensure that upgrades of systems are reasonably painless and it enables others to use, maintain and develop the innovative new features. It is therefore important that when procuring open source software solutions you also plan to properly resource collaboration work.

The march of WebM

On Tuesday Google announced that it would be withdrawing support for the video standard h.264 from its Chrome browser in two months’ time. Back in June 2009 we covered the contortions that Google had gone through to use the open source video codec FFmpeg in Chrome to decode embedded videos without risking the wrath of the the owners of that standard, the MPEG LA h.264 patent pool. Essentially Google used FFmpeg but did not acquire a licence from the patent pool for FFmpeg itself, but instead for its Chrome browser. This was a cunning move but at the time it annoyed some around the open web standards community who felt – with some justification – that Google’s move was something of an ‘I’m alright Jack’ statement to the rest of them. Representatives of the Mozilla project who produce Firefox scalded Google on public lists, as reported in that previous post.

In August 2009, a couple of months later, Google bought the media compression company On2 for about £100m. This was widely interpreted as an initial move to bypass the stranglehold that h.264 had on web video and perhaps produce a genuinely open standard with all necessary patents licensed for use by all. After all, On2 was the company that had contributed its (somewhat outdated) video codec VP3 to the web standards community, a codec which had come to be the basis of open video encoding standard Theora. During the arguments over Chrome and FFmpeg, many had argued that Theora should be the basis for web video and its embedding in the new HTML 5 standard using the new <video> tag. At the time Google had pointed out that Theora/VP3 was just not as efficient as h.264, and that its use would mean hosts and users paying for more bandwidth for the same quality video as h.264 provided.

In May last year, Google announced a new video container standard – WebM – which incorporated three related open media standards. Digital media files typically contain multiple streams of data encoding video, audio and subtitle information. To manage these multiple streams, media files are packaged in what are often called ‘container’ formats – essentially standards that describe how streams can be packaged together and decoded by playback hardware. Early container formats such as MPEG and AVI did not support complex options such as multiple audio and video streams, or incorporated subtitles. As it became clear that these older formats were unable to mimic the experience end users expected from their DVD players, more complex containers – such as the open standard Matroska – were developed. Matroska – which enabled multiple audio, video and subtitle tracks in a single file – became popular quickly, in part due to its ability to allow Japanese Anime video files to be exchanged over the internet in a form that satisfied both native Japanese speakers and others who required subtitles or dubbed audio. Google’s WebM was a new container format based heavily on the Matroska template. Within each WebM file there is typically at least an audio stream encoded by the open Ogg Vorbis standard and a video stream encoded by the VP8 standard (a much more efficient progression of the VP3/Theora codec) which Google had acquired with On2 the previous August. In the same announcement, Google, Mozilla and Opera revealed new builds of their browsers which fully support the WebM standard.

So since last year there has been – thanks to Google – an entirely open group of video and audio standards that anyone can use in their web content and encode and decode in their software. The question remained, though, would it get used? Microsoft and Apple were both licensors in the h.264 patent pool, and stood to make money if they could keep h.264 the de facto web video standard. The day after the WebM announcement Microsoft revealed that Internet Explorer would also support the new open standard. That left – and leaves – Apple. Apple’s browser Safari does not support WebM and Apple has announced no plans for it to do so. This is not such a big problem in the desktop, where Chrome is available for Mac OSX with WebM support, but in the mobile space, it potentially is. Mobile Safari, the only web browser currently permitted by Apple to render HTML on the iPhone, iPad and iPod, accounts for 40-50% of mobile web traffic. While WebM can be a huge success on the desktop with or without Apple’s support, in the mobile space Apple’s cooperation is more important. So would they play ball? It seems unlikely, at least for the moment. The day after the Google WebM announcement, an Apple engineer briefed news site Apple Insider saying that the VP8 codec element of WebM was ‘a mess’ and not competitive with h.264.

By dumping h.264 support from Chrome, Google is making a strong statement that it is prepared to use its power to drive WebM adoption. After all, Google owns Youtube, the most accessed video hosting site on the internet.  If Google shuts down h.264 support on Youtube, it could kill that standard – at least in web video – overnight. Whether it is prepared to take this action remains to be seen. Certainly it seems likely that the MPEG LA h.264 licensors might have something to say about such an action –  and its anti-competitive effect – in court.

For the open source and open standards community, WebM is unequivocally good news, as is Google’s willingness to strong-arm its adoption. Open source video decoding has long been a dark art for legal rather than technical reasons. The existence of a genuinely royalty-free video format set removes the threat of legal action from open source video projects which adopt it. For educational establishments, the current lack of involvement of Apple is troubling. iTunesU has been extremely successful in the UK, with UK institutions producing a surprisingly high proportion of the most downloaded material worldwide. With the current emphasis on open educational resources in UK policy, a genuinely open video standard would seem a natural way for open educational resources to be distributed – opening as it does the possibility of entirely legitimate royalty-free open source video editing software and a consequent widening of the audience who are capable of remixing the material for their own purposes. However while iTunesU remains the primary portal for UK educational video content, and Apple’s native software (iTunes, iOS video decoding, Quicktime) does not support WebM, it seems unlikely that we will be able to take full advantage of this advance.