Assessing the Role of Libre/Open Source Software for European Industry, The Hague, 19 November 2004

by Rowan Wilson on 30 November 2004 , last updated

Archived This page has been archived. Its content will not be updated. Further details of our archive policy.

Introduction

CALIBRE (Coordination Action for LIBRE software) is an EU-funded consortium of academic and industrial groups whose goal is to coordinate research into open source software and associated issues. The project is led by the University of Limerick. This, their first international conference, was titled Realising the Potential of Open Source Software for European Industry: the Next Generation of Software Methods and Services. This conference also saw the launch of CALIBRATION, an industry forum on open source software policy.

Introductions and Key Note Address

We were welcomed to the Rembrandt Hall in the Netherlands Congress Centre by Hans De Groene, Director of the Innovation Department in the Dutch Ministry of Economic Affairs. The Netherlands has much expertise in software development, particularly in the realm of embedded systems, he explained. Dutch companies such as TomTom (who produce GPS navigation devices and software) and Vialis (who produce transport-related hardware like parking ticket machines) are successful worldwide. They also make some of their software available under open source licences.

After Hans De Groene’s warm welcome, our next introduction came from Jesus Villasante, Head of the Information Society Unit at the European Commission. In a theme that was to be taken up many times throughout the day, Villasante stressed the strategic importance of open source to Europe. It was vital, he said, to convince European politicians of this strategic importance, in order that they might exploit it to the full.

Next to speak was Colm Butler, Director of Information Society Policy at the Department of the Taoiseach, in his third appearance at The Hague that week. We had seen Butler speak the previous day at the FLOSSPOLS conference across the hall, and so some of his themes were familiar. Butler advocated a careful policy with regard to intellectual propety management, which at the same time preserved the right to profit from innovation while minimising barriers to creativity. He also again stressed the difficulties over open source terminology that he and others had experienced, and called for an effort on the part of the open source community to help promote understanding through clarity.

Finally in this session of introductions Professor Brian Fitzgerald of the Univeristy of Limerick, CALIBRE’s coordinator, spoke to us. CALIBRE aims to research and investigate open source software, distributed development models and agile methods of generating software (such as extreme programming) in order to establish empirical methods of measuring their effectiveness and success. The academic approach to these issues would be multi-disciplinary, as was necessitated by such a complex phenomenon. In addition to this ambitious workplan CALIBRE will be promoting the results of this research to European industry, particularly the secondary software sector of industries that produce code as part of a larger product. How can development communities be built around this kind of specialist software? They will attempt to answer this question. They will also consider issues surrounding intellectual property management and other legal issues such as liability, and also assemble information on successful open source business models. Finally they will establish a European open source Policy Forum, called CALIBRATION, to help drive understanding of these many issues in European Industry.

Professor Fitzgerald then introduced the day’s key note speaker, Paul Everitt of Zope Europe. Everitt began by explaining how Zope had come to be an open source project. Digital Creations, a company that Everitt co-founded, created Zope as a proprietary application server in the late nineties. They struggled to sell licences, and made most of their money from consulting and support of their product. In late 1998 Digital Creations sought venture capital funding and received three quarters of a million dollars from investor Hadar Pedhazur. A speech by Eric Raymond at a Python conference, in combination with some well-timed advice from Pedhazur persuaded Everitt that they could benefit from releasing Zope as open source. After all - the media buzz around open source at the time would get them many column inches - and they were not making any money out of licensing anyway. Through the dot-com bubble Zope survived, winning industry awards and two more rounds of VC funding. Everitt now runs Zope Europe, a kind of guild of small firms specialising in Zope consulting and bespoke development. Everitt’s message is that open source products do not currently represent whole products in the way that many proprietary pieces of software do. Open source offerings almost universally lack the sales support teams, warranty provision and certification structures that are common in the proprietary sector. Syndicates of smaller firms clustered around an open source product can do this, Everitt asserts, as the Zope Europe organisation shows. In questions after the speech it was remarked that trust between these would be a major factor in the success of these whole product syndicates as they expanded. After all, the provision of the integrated service depended upon a smooth delegation of business to other firms who are, in the conventional view, potential competitors. Everitt replied that as things stood in Zope Europe, it was a friendly, socially cohesive group, and that this social cohesion was an essential part of the business model.

OSS in the European Secondary Software Sector

After a refeshing coffee break with excellent biscuits we resumed with a talk from Frank van der Linden of Philips Medical Systems. Van der Linden explained that Philips produce many medical hardware devices such as scanners and monitoring systems, and these all contain software components. Increasingly the differences between one manufacturer’s device and another’s equivalent was embodied in some aspects of the software controlling the device. Given this fact, it made sense for Philips to concentrate the majority of their resources on improving this software-based differentiating technology. However, according to van der Linden’s calculation, only approximately five percent of a device’s software was in fact differentiating - the remainder was either commercial off-the-shelf software or middleware developed independently within separate medical device firms but fulfilling precisely the same functions for each. It makes sense to take the software in this second category and open source it, producing a medical-oriented middleware project to which all interested parties can contribute and benefit from others’ work. This is not an anti-competitive love-in, however. It would, with any luck, have the desired effect of freeing resources within firms to concentrate on honing their differentiating technologies.

Next up was Heikki Sakkonen of Nokia Research Centre. Sakkonen described the Nokia Research Centre’s experiences with open source, and was refreshingly forthcoming about the rest of Nokia’s skeptical view of their enthusiasm for it. Nokia have had their fingers burnt once before through over-enthusiastic embracing of open source software. A finished prototype for a Linux-based set-top box had to be abandoned when it was realised that no-one had checked the intellectual property status of any of the media-playing codecs that they bundled with it. Despite this Nokia continue to get involved in open source, as with their release of the Affix bluetooth stack under the LGPL. They are also porting perl to their series 60 units and this is a potential driver to the development of open source software on smartphones. Sakkonen ended by identifying the risks that Nokia saw in their involvement with open source. As usual these centred around the legal issues of inadvertent infringement of others’ intellectual property rights and also liability for damages.

Finally in this session we heard from Daniel Weyl, of BMW’s Research and Technology Division. Weyl first explained to us that open source software is widely used within BMW for their internet site and some intranet functionality. Where it is not - thusfar - is in the microcontrollers in the vehicles themselves. A BMW 7-series has seventy controller units, and it takes between seven and ten hours to flash software onto all of them. If an upgrade is needed in any individual controller, all controllers must be re-flashed to make absolutely sure of compatibility. Weyl admitted that software was far from BMW’s primary expertise: “we are metal-benders”. Nevertheless his personal aim was to - at some point in the future - see BMW producing cars for which a documented API was distributed and for which technically-minded drivers could develop their own software. Media players and diagnostics programs were two examples of potential automotive applications given. Weyl was also frank about the resistance to this idea within BMW. Clearly there could be security and safety issues with this strategy, and there was some concern over whether these actions would marry well with the established BMW brand image.

OSS in SMEs

Anders Mattsson of Swedish embedded-systems software firm Combitech talked to us in this session. Mattsson spoke about his firm’s desire to move from proprietary UML tools to open source equivalents in order to cut outlay on licences, and to allow easy customisation. However the open source UML tools that are currently available (for example ArgoUML) lack some of the core functionality that Combitech need - full generation of code from a model, execution of a model, process modelling and XML model exchange. This leaves Combitech with three possible approaches if they wish to go open source to some degree: do the hard work of adding this functionality to the current open source tools, adopt a mixed set of open and proprietary tools to reduce costs moderately, and contribute to the open source tools by subjecting them to “industry-strength testing”. Mattsson concluded that in this area open source tools are immature.

Before lunch Andrea Deverell of CALIBRE told us about the post-prandial plan. Conference attendees were to be divided into small groups and asked to think about the role of open source in the secondary software sector in Europe from a variety of different perspectives. A colleague explained to me that this was a variation on Edward De Bono’s Six Thinking Hats exercise. I remained slightly confused. The results of our groups’ considerations would be presented by a group member to the conference, and a prize awarded to the best contribution. The combined results of the groups’ work would serve to sketch out the issues to be dealt with by the CALIBRATION forum, of which this was the formal launch. With some trepidation we went to lunch.

CALIBRATION Workshop Session

On our return we found that our chairs had been arranged into ovals dotted around the hall, each with a flipchart at its apex. Each group would consider the role of open source in the secondary software sector in Europe from two perspectives selected from a set of four (emotional, positive, negative and creative). The time allowed for each perspective was strictly limited and this combined with the half-heard shouted suggestions of other groups added considerably to the sense of urgency. In a piece of unintended irony the group that won had selected the only attendee from Microsoft to be their spokesman. Many useful observations were made by all the groups - too many to list here. There was a consensus that the development of the Chinese economy might pose a threat to European industry.

OSS in Multi-Nationals

Hastily rearranging our chairs into rows, we next heard from Sarah Ewen, of Sony Computer Entertainment Europe. Ewen explained that Sony’s Research and Development department in Japan consisted of approximately forty people all using open source desktops and development tools. This reflects the generally widespread use of Linux within Japan. The development kits shipped by Sony to third-party game developers run on Linux (although there are popular third-party ‘middleware’ development environments such as Criterion’s RenderWare are Windows-based). Sony have also shipped a Linux distro for the PS2, bundled with a keyboard peripheral, ethernet adaptor and hard drive. Sony also maintain a Linux community web site supporting private developers who want to program on the PS2 running Linux. The National Center for Supercomputing Applications in Illinois has built a sixty-four node cluster of Linux-based PS2s.

Next we heard from Simon Phipps, Chief Technology Evangelist at Sun Microsystems. Phipps’ message was that Sun was an open source-friendly company, and had been so since well before the phrase open source was coined. Questions began well before the end of his talk, in shouted form. Most concerned the Java licensing situation, and Phipps did not reply in precise terms.

The final session of the day was presented by Maureen O’Sullivan, Lecturer in Law at the University of the West Of England in Bristol. O’Sullivan began by speaking generally about open source licensing, characterising all open source licences as arrayed between the two extremes of BSD and GPL. All software licences depend upon copyright law, which is slightly problematic as there is no harmonization of copyright legislation worldwide, and so general licences like the GPL can be interpreted differently in different jurisdictions. One solution to this would be to create localised versions of the GPL for each legislative region. The downside of this would be that there can be no guarantee that all licences produced this way are exactly equivalent in meaning. The Creative Commons project is going down this route nevertheless. Other threats to FLOSS, O’Sullivan said, were the FUD-strategies of opponents of open source, well-financed-but-insubstantial vexatious lawsuits designed to spread worry, and software patent issues. What can be done to improve the situation? O’Sullivan recommends the framing of a piece of international law, possibly adopted by the UN, which attempts to pin down areas which are currently uncertain due to the problems mentioned above - a Free Software Act. The Act would essentially formalise the terms of FLOSS licences at a level above the national, thereby avoiding the issues associated with localisation. O’Sullivan proposed that, as well as codifying FLOSS licences, the act could provide specific exemptions, define who has standing to sue when a licence is breached, and try to promote rewriting of offending code over damages as a solution to violation.

Conclusions

The first international CALIBRE conference was an interesting event. Passions ran high at certain points, leading to some barracking of speakers by zealous open source enthusiasts. It was also interesting to hear questions from representatives of industry organisations like CompTIA, who robustly put the point that open source did not need any help from projects such as CALIBRE as it was succeeding in the market perfectly well on its own. Many speakers made the point that they personally were convinced of the business efficacy of open source, but that there were other bodies of opinion within their organisations which were not so sure. CALIBRE may be able to sway these opinions with its outputs, particularly its research into building development communities around specialised software.

Paul Everitt’s account of building a symbiotic group of companies to mimic the services of single large software firms was fascinating. However the social cohesion that allows this symbiosis may be a limiting factor as these groups grow beyond a certain size - it will be interesting to see how future examples of these groups develop.

We look forward to future CALIBRE events.