OSS Watch is pleased to announce the first issue of our newsletter, available as a PDF. It will be published monthly, and include a selection of feature items, news, and upcoming events.
Author Archive for Rowan Wilson
Rejoice, licence geeks! A new international journal has been launched covering developments in law relating to free and open source software. To quote the journal’s web site:
The International Free and Open Source Software Law Review (IFOSS L. Rev.) is a collaborative legal publication aiming to increase knowledge and understanding among lawyers about Free and Open Source Software issues. Topics covered include copyright, licence implementation, licence interpretation, software patents, open standards, case law and statutory changes.
In the first issue there are discussions of the Jacobsen vs Katzer case in the US and its repercussions for both in the US and the UK, the peer-to-patent programme and a proposal to help organisations procuring free and open source software understand and manage their risks. It’s good if slightly technical stuff, and I highly recommend it.
HTML 5 - the next version of the language used to create web pages - is due to be finalised somewhere between 2010 and 2012, depending on whether you believe the W3C or the WHATWG, the two groups who are contributing to the language specification. One of the expected benefits of the new version is that it will allow page authors to reference and embed videos in their pages without requiring end users to have a specific browser plugin installed. At the moment, Adobe’s Flash and Apple’s Quicktime are the most popular methods for getting video into web pages, but this irks many web purists who feel that the requirement to install these plugins undermines the goal of a truly ‘open’ standards-based web. After all, while web video playback depends upon vendor-specific software, the vendors in question have enormous power to control which browsers and operating systems they will support and therefore which will be widely used. There is of course a powerful pragmatic counter-argument to this, which says that the success of the web is driven by adoption not contemplation and any solution that has been generally adopted has to be considered the correct one. If you are a business wanting to outshine your competitor’s web presence, you are likely to adopt whatever is the most attractive, eye-catching and available technology at the time without worrying to much about high-minded ideals of openness. Into this argument, HTML5 - and specifically its <video> tag - strides to resolve the question once and for all. That at least was the idea.
Proprietary plugins are not the only things that threaten the ‘open’ web. Patents covering the use and creation of certain types of media file are also a problem. It would be fairly pointless to sweep away proprietary video playback plugins if you then replace them with in-browser functionality that requires a patent licence to actually include. Therefore as well as mandating that compliant browsers should be able to playback video URLs indicated by the <video> tag, the HTML5 specification originally directed that every browser should - at the very least - be able to decode the ogg theora multimedia encoding format. Theora - named after a character from Max Headroom an 80s sci-fi TV drama - is an open standard for the encoding of video which is of a similar quality to the patented H.264 MPEG4 encoding standard used in Quicktime and on many Blu-Ray High Definition video discs. Theora derives from work done by technology firm On2 who have granted a worldwide royalty-free licence to anyone who wishes to incorporate theora encoding and decoding functionality into their products. The software for actually doing the encoding and decoding is released under the open source BSD licence. It all sounds great I’m sure you’ll agree - patent-free video (and audio - ogg has a project called vorbis that handles that) playing right in your browser and no-one has to install a plugin or pay a patent licence.
Unfortunately Apple and Nokia - companies that would be creating HTML5 browsers in the future - lobbied successfully to have the theora/vorbis mandate removed. Nokia puzzlingly claimed(pdf) that theora was proprietary, and less puzzlingly pointed out that the format had no accompanying Digital Rights Management implementation that might help providers make money from their content. Apple argued that although the On2 patents in theora were licensed royalty-free, there was no guarantee that there were not other video-related patents out there that might be infringed upon by theora (particularly submarine patents - meaning patents that have yet to be fully published for administrative reasons but which it is still possible to inadvertently infringe). Apple also claimed that there were few-if-any hardware implementations of theora decoders, meaning that low-powered devices like mobile phones could not incorporate a specialised decoding chip to take the pressure of video work off their overworked central processors and sagging batteries. Open web enthusiasts pointed out that no video format is free of submarine patent concerns, and mandating theora in HTML5 would almost certainly mean that there soon would be theora decoding hardware available. Some uncharitable commentators also pointed out that Apple held one of the essential patents for MPEG4 decoding, and so stood to make money from additional licence sales if MPEG4 became a common format for the contents of the HTML5 <video> tag.
So the issue was already a vexed one when last week Google announced on the WHATWG mailing list that their support for the <video> tag in version 3 of their browser Google Chrome was going to use LGPLv2.1-licensed free software video library ffmpeg to decode both theora and MPEG4 H.264. Why was this controversial? Well as I mentioned above, H.264 is subject to a whole cupboard-load of patents, and anyone who looks like they have a bit of money would be well advised to get a comprehensive licence from the relevant group or be prepared for a writ on the corporate doormat one morning. Clearly anyone with a pocket-book the size of Google’s would have taken this essential step. However even doing this might well not allow you to distribute ffmpeg safely. The LGPLv2.1 - like the GPLv2 - contains a clause which aims to prevent the software it covers becoming subject to additional responsibilities and fees. This clause (section 11) reads:
If, as a consequence of a court judgment or allegation of patent infringement or
for any other reason (not limited to patent issues), conditions are imposed on
you (whether by court order, agreement or otherwise)that contradict the
conditions of this License, they do not excuse you from the conditions of this
License. If you cannot distribute so as to satisfy simultaneously your obligations
under this License and any other pertinent obligations, then as a consequence you
may not distribute the Library at all. For example, if a patent license would
not permit royalty-free redistribution of the Library by all those who receive copies
directly or indirectly through you, then the only way you could satisfy both it
and this License would be to refrain entirely from distribution of the Library.
This clause is generally understood to mean that paying for patent licences to distribute LGPLv2.1-licensed code is impossible; once you pay for the licence you lose your right to distribute, so it’s never worth doing (and this in turn acts as a deterrent to patent owners looking to get per-copy royalties on LGPLv2.1- and GPLv2-licensed software - even if you win the case against the distributor, you kill the product and will get no royalties going forward). Nevertheless, Google’s lawyers seemed to have found a way through waters which had previously been thought unnavigable. As the argument on the thread continued, it emerged that Google’s approach had been to seek a patent licence for Chrome itself, not the ffmpeg library that was actually doing the video decoding. The LGPLv2.1 is designed to allow code it covers to co-exist in an uneasy peace with code covered by any other kind of licence, including patent licences. In this was free software libraries could be used with proprietary software in a way that would be impossible if they were covered by the full GPLv2. Google argued that as long as they never sought patent rights to use ffmpeg, no-one could accuse them of not passing on those rights to others in a way that ffmpeg’s licence mandated. All their permissions to use patents rested with Chrome itself - however it chose to do the decoding.
Now this is a sly move, and not one that has been attempted in as high-profile a piece of software as Chrome up to now. Will Google get away with it? Well that question really reduces to the question: Who might try to stop them? In this case Google Chrome is relying on licences from the MPEG-LA for the patents used to decode MPEG4 (and possibly theora if they want to rule out any future challenges from MPEG-LA on that score) and a software licence from the ffmpeg project. The likelihood is that Google have paid well for complete coverage from the MPEG-LA, and so a challenge from that quarter seems unlikely. What about the ffmpeg project? Well they clearly explain that they have no specific knowledge of any patents infringed by distributing ffmpeg. Of course, everyone is aware that the project very likely infringes on many patents in the H.264 patent pool at the very least, but time and experience has taught programmers operating in patent-rich problem domains to never read the relevant patents. It only increases the likelihood that you will infringe upon them and increase the damages if it can be shown that you knowingly infringed. What this means for ffmpeg is that it would be very difficult for them to argue that Google is not in compliance with the LGPLv2.1 for ffmpeg because they have licensed some patents ostensibly for use in Chrome. To do so they would have to explain exactly which of the patents licensed were in reality embodied in ffmpeg, and would therefore trigger Google’s responsibilities inder the LGPLv2.1. This could be seen as an unwise move for ffmpeg. With neither the MPEG-LA or ffmpeg likely to question Google’s actions, it seems likely they will get no substantial challenge to their actions.
From time to time OSS Watch will be publishing guest blog posts here to highlight interesting opinions from around the FOSS community. This post is by Gerry Gavigan, chairman of the Open Source Consortium.
Elections for the European Parliament are scheduled for 2 June 2009, and this seems like an ideal opportunity to broach the subject of ICT policies and draw the attention of candidates to the issues surrounding choices by government and its administration concerning software and systems. Hopefully, this will at least raise the profile and gain better consideration of the underpinning issues and lead to wider adoption of Free and Open Source Software.
One initiative attempting to do this is the Free Software Pact (FSP) providing candidates with a method to inform the voting public that “they favour the development and use of Free Software, and will protect it from possible threatening EU legislation.”
As Chairman of the Open Source Consortium I was asked to help promote the FSP. Fundamentally I agree with the underlying objectives of the FSP but I am not sure that the approach to implementation is well thought through.
I am convinced that level playing fields coupled with open and transparent processes topped off with effective governance (easier said than done, of course) are all that Free Software needs for new installations and also to trickle through the arteries and capillaries of existing ICT deployments, eventually eliminating the externalities of proprietary software (largely summarisable as the deadweight costs arising from distortions in the market). I should add that this opinion is vendor neutral. I don’t care who it is or how Free Software is supplied.
If we are to achieve this nirvana we need to approach those we are seeking to influence in terms that are meaningful to them. This is not always easy but that is not an excuse not to try. And it may require several goes. Additionally if one is seeking to influence politicians or administrations is is easier to go with the grain of things than start again somewhere else.
With that in mind and prompted by the FSP, it seems right to put my drafting where my opinions were. Accordingly the OSC has created a draft form letter that hope as many people as possible will send it to the candidates in their constituency.
Dear…
Support Implementation of draft European Interoperability Framework V2 “EIF V2″
As you are a candidate in the forthcoming elections for the European Parliament I am writing to seek your support for a simple measure that will help citizens, businesses and European society.
Information and Communications Technology (ICT) has become a critical part of European infrastructure; a modern economy cannot function without it.
Optimal choices in ICT need a European policy on interoperability in ICT.
Interoperability means it does not matter how you mix and match the constituent parts of ICT, they will just work with each other without problems or issues. Such choice also means that users of your ICT systems (e.g., users of European or national government on-line services) do not have to make ICT choices based on your decisions, e.g., having to use a specific web browser (that itself may require a specific operating system).
Many European and national ICT systems have been implemented using software which does not enable easy interoperability, and instead creates an effect where it becomes easier, and some cases necessary, to choose more of the same supplier’s products, and harder to choose competing products.
Moreover, once governments or local administrations have decided to use software that does not enable easy interoperability, that choice imposes a requirement for the citizen or customer to choose the same software.
This network effect prevents choice, competition and limits opportunities to promote innovation outside vendor control, in the whole European arena for ICT.
To have agreed as policy the (draft) EIF V2 ( http://ec.europa.eu/idabc/en/document/7728) will halt and eventually reverse this situation.
This draft has been criticised by vendors currently benefiting from the current uneven playing field, but its adoption would not prevent any vendor, current or future, from supplying ICT to European or national governments.
The EIF is not considered to by all to be perfect, but getting it firmly “inside the tent” would enable it to be improved. I urge you to adopt or adapt this letter and write.
You can find your existing MEPs and candidates here.
Gerry Gavigan
As we mentioned back in February, Microsoft has sued Dutch GPS device and software manufacturer Tomtom over - among other things - Tomtom’s use of GNU Linux; specifically a Microsoft-developed file-system component within GNU/Linux, FAT. In our previous post we speculated that the Open Invention Network might play a role in this drama and indeed on the 23rd March, a few days after they countersued Microsoft over Microsoft’s alleged implementation of Tomtom patents in their own products, Tomtom became a member of the OIN family. A week later, Tomtom and Microsoft announced that they were settling their disputes. Unfortunately the detail of the settlement is not public, beyond the information in the Microsoft announcement. Of greatest interest to the free and open source software community is precisely how the deal interacts with Tomtom’s responsibilities under their licence to distribute GNU/Linux, the GNU General Public License (GPL) v2. Section 7 of the GPL has this to say:
7. If, as a consequence of a court judgment or allegation of patent infringement or for any other reason (not limited to patent issues),conditions are imposed on you (whether by court order, agreement or otherwise) that contradict the conditions of this License, they do not excuse you from the conditions of this License. If you cannot distribute so as to satisfy simultaneously your obligations under this License and any other pertinent obligations, then as a consequence you may not distribute the Program at all. For example, if a patent license would not permit royalty-free redistribution of the Program by all those who receive copies directly or indirectly through you, the only way you could satisfy both it and this License would be to refrain entirely from distribution of the Program.
What this meant for Tomtom was that if they settled with Microsoft by - say - agreeing to pay them 1p for each copy of GNU/Linux they distributed, then they would in fact have invalidated their licence to distribute GNU/Linux at all. The GPL v2 insists that either the code it covers goes out with no additional encumbrances or it doe not go out at all. Now Microsoft is very aware of this provision. Their cross-licensing deal with Novell - detailed in an OSS Watch blog post here - was structured precisely so as to avoid this consequence. Many in the free and open source community feared that Tomtom would strike a similar deal with Microsoft, and fans of the lastest version of the GPL, v3, announced Microsoft’s actions in this case should act as a ‘wake-up call’ for developers to begin using the licence, as it includes changes designed to make such deals apply to the entire community if they exist at all. In the event, it would seem from the information we have that Tomtom has avoided emulating Novell, and has instead agreed to remove the FAT components from the version GNU/Linux they distribute on their devices within two years. This strongly indicates that Tomtom have accepted no licence from Microsoft relating to GNU/Linux itself, and Microsoft have agreed to just give them breathing space to remove allegedly infringing components, rather than insisting they take a licence. So - a happy ending for all concerned? Not really. Microsoft must be pleased, as they have managed to give the appearance of winning a battle in which they were accusing GNU/Linux of infringing on their patents while not actually having to suffer examination of the validity of their patents in court. The free and open source software community is left wondering whether the FAT patents could be used against other distributors of GNU/Linux. Jim Zemlin, Executive Director of the Linux Foundation, seemed to be calling on developers to cut FAT implementations out of their software altogether. A couple of days ago, OIN announced the following:
…the patents used in the recent TomTom patent action have been posted by OIN for review and submission of prior art by the Linux community. Submissions may be made by visiting http://www.post-issue.org, clicking on the appropriate patent and selecting “Submit Prior Art”.
The Peer to Patent web site that OIN cite in that link is an initiative designed to allow the technology community at large to pool their collective expertise and thereby perhaps invalidate patents that ought not to have been granted. One way to invalidate a patent is to show that it was not innovative at the time the patent was applied for, that there was in the jargon ‘prior art’. The Peer to Patent site acts as a clearing-house for evidence of prior art on specific patents. In fact one of the FAT patents in question, 5,597,517, has already been re-examined, provisionally invalidated, amended and then declared valid again after the Public Patent Foundation or ‘PubPat’ (tagline: ‘representing the public’s interest in the patent system’) asked for the US Patent Office to take another look back in 2004. It will be interesting to see how far this pillorying of the Microsoft’s FAT patents will go towards their invalidation, given that this is their second re-examination.So another Microsoft patent case goes by and we are still no closer to knowing how justified Microsoft’s patent threats against Linux really are. Tomtom clearly found the effort involved in engineering out the FAT patents a worthwhile expense, but only tells us that it was likely to work out cheaper than the lawyer’s fees, and that Tomtom understandably had no interest in acting as a vanguard for the community on the issue. Perhaps the most reassuring thing for the community is that - so far - Microsoft has only claimed ownership of technologies which can be engineered out of GNU/Linux without breaking it.
