Louis Suárez-Potts - interview

From LXF Wiki

Flocking together

Linux Format was rather surprised to meet Louis Suárez-Potts in Montreal. What's an office app developer doing attending the Libre Graphics Meeting?

Earlier this year, Linux Format attended The Libre Graphics Meeting at École Polytechnique de Montréal in Canada, which touted itself as being, "all about participation; artists and developers, bring your laptops and show us what you can (and can't yet) do. Organise a BOF about your favourite project or feature." Unlike a lot of other conferences, which tend to be dominated by marketing and sales people, lead organizer Louis Desjardins arranged The Libre Graphics Meeting with the intention of it being for developers in particular: the air was abuzz with the heady idea of cross-pollination between open source projects ­ and not just those concerned primarily with graphics.

We caught up with Louis Suárez-Potts, Community Development Manager and Community Lead for the world's leading open source desktop application, OpenOffice.org (http://www.openoffice.org), to seek his opinion on the challenges facing many open source projects, particularly the implementation of the Open Document Format (ODF), and the advantages and disadvantages of corporate contributions to open source development.

Linux Format: What else fills your working day, as well as your broad and exhausting-sounding remit at OpenOffice.org?

Louis Suárez-Potts As well as being the primary liaison with the OpenOffice.org (OOo) primary contributor and sponsoring company ­ Sun Microsystems, CollabNet (www.collab.net), the software company hosting OpenOffice.org, employed me as the Senior Community Development Manager and open source strategy consultant from 2000 ­ 2007. Recently I've been examining the problematics of "infonomics," or the political economy of information. Some of my publications and papers can be found at http://homepage.mac.com/luispo/. One of my blogs touches on OpenOffice.org and open source; http://ooo-speak.blogspot.com, another centers on cultural criticism.

LXF: This second question may seem impertinent, but why did you want to come to the Libre Graphics Meeting? It seems like a strange choice of conference for someone so involved with the more er, Officey side of productivity applications...

LSP: Two main reasons. Firstly, this is a developer-focused conference. Most of the people here are involved with Open Source Graphics in some way. And of course there's the open document format. The charter of the ODF focuses mostly on documents, but there's also scope for graphics to be used in presentations and so forth. There has been a lot of interest and activity in how OpenOffice can transcend the boundaries of the traditional office suite, to start including things that are far more interesting from a creative point of view, and to solve the problem of how to deploy the ODF in other programs that aren't a part of the office suite. For example Scribus; if you can import your text in OOo format, why not also be able to save and share documents as an open source document, like ODF? The advantage is clear if for example you need to update a page regularly with data from a spreadsheet.

LXF: And the second reason?

LSP: It's not about competition ­ we're not all competing with one `bad guy'; in open source generally, we need to pull together. Whatever open source project you're involved in, you have your own mission, but the aim of all our activities is primarily with offering usability and choice to the end user. The very notion of collaboration should extend beyond the boundaries of any open source project. A common format is important, but so are things like similar technology, architecture, code etc ­ they're all important. No-one's going to actually buy applications like KOffice or OpenOffice.org, but they're going to choose one of these open source alternatives to the ubiquitous Microsoft Office on the strength of not only its own abilities, but also the way in which it interacts with the other programs on their desktop.

LXF: That is almost the entire purpose of open source, isn't it? It's one of the key differentiators between open source and proprietary software.

LSP: Supposedly! The Hobbesian ideal (that unrestrained competition is good for business) can be incompatible with the open source ethos; I'm not saying that developers should all collaborate on each other's projects ­ that would be unworkable ­ but what's needed is collaboration with other projects to capitalise on our strengths to address common issues. One good reason for OOo collaborating with Scribus: the latter application is very discriminating regarding fonts; they have to meet specifications that go far beyond what most other applications demand. The two teams have discussed the possibility of establishing a library of fonts ("library" in the old sense: a repository) that have been vetted according to Scribus' standards but can also work with other compliant applications, such as OOo. This would not only help Scribus and OOo, but the user too.

And, from my perspective, having ODF implemented by Scribus and other relevant applications seems desirable, though not as compelling as perfecting import functionality. One of the coolest things about The Libre Graphics Meeting is the superb broadsheet published by the conference team. It shows how good Scribus is (Louis Desjardins is a project member) and showcased the seriousness of the effort put into the event.

LXF: By extending an open source program's functionality, it will attract a broader base of potential users to migrate from the proprietary options?

LSP: I'm not for one moment suggesting that OOo ought to be the Borg of products, assimilating everything! Something as simple as a pop-up box in OOo prompting a user about the option of using Scribus if the way in which they were formatting their file suggested to the program that they would be better off using a DTP application would be a great example of the strides in strategy that open source could be making in interoperability and user-friendliness. Likewise with graphics ­ the same construct could point OOo users at Gimp if they are attempting to do anything more graphically complex than drawing a box in Ooo.

It's all about enabling users. Many ­ if not most ­ computer users never really go through the process of actually figuring out what the best program would be for completing the task in hand ­ they just end up using whichever product is already there, the one that came with their computer (or support package, in the case of larger business users) rather than evaluating the different options that are available. In the open source world, you lose nothing by advertising somebody else's product ­anything that enlarges the sphere of open source users is a good thing.

LXF: What sort of response do you get from other projects when you talk about this kind of interoperability?

LSP: Most of them are perfectly happy with this. Working together on compatibility is important. Yesterday me and some of the OOo team were talking with the KOffice guys and sharing tips about refining compatibility in general, and what the demands are surrounding ODF. And how to license the technology ­ an important consideration with the release of GPL 3.

LXF: Do you think that sponsor involvement introduces an overtone of obligation for attendees at events such as this?

LSP: It depends on the organisation. It can be difficult ethically ­ it's one thing at a conference to be treated to a sumptuous dinner by a commercial entity showing its good faith with the aims of the open source, but higher levels of involvement can introduce a level of commercial pressure that doesn't always sit well with the academic nature of discussion. Rather than a junket, perhaps the money could be better spent by sponsor companies on, for example, training project managers how to work in a more non- hierarchical fashion...

It's clear that a lot of companies feel that there is a conflict between their business model and the academic aims of open source, but the lines between sponsor and contributor are becoming more blurred than they used to be ­ Novell, it could be argued, falls into both categories: providing conference funding on one hand, and giving code to the community at the same time. There should be clear ethical boundaries, but who is going to take the decision at exactly where the line is drawn? Most corporations would typically fund some sort of event, but organisers need to be careful, otherwise someone's going to ask the awkward question, "How much does it cost to buy the interest of the developer?"

In my opinion, some companies don't really act as a sponsor in a true sense ­ they're prepared to contribute resources by supplying people to work on areas of programming that lie within their own development framework, to the detriment of unassociated areas, and don't really work with the non-developer community ­ that is, other businesses that are interested in the exact same open source areas on which the sponsor company is contributing to. These other parties will eventually get access to the code, because it all goes back into the source, but what they won't have an insight into the reasons why the code was put together in that way.

LXF: Is there an unassailable division between business and the community?

LSP: Not at all. In Ooo's case, back in 2002-3. the compatibility for the upcoming Mac OS X was handled by a widely distributed group of international contributors working towards a common goal ­ what better example of community could there be? It would be great to have companies like Novell working more fully with the community, and not just in this market-heavy fashion. Novell sponsors OOo Con, which is great, but rather than a barge cruise, using that money to help with QA, localisation and training etc would see a better return for both Novell and the community as a whole.

People think divisions exist because of the nature of different participants' contribution or sponsorship. For instance, IBM can't be more contributive to OpenOffice. org because Ooo competes with some of its existing products; or why should Adobe contribute to an open source project when it competes with a product that's an important part of its revenue stream?

LXF: Is it difficult to co-ordinate a project that has contributions from both volunteers and paid staff alike?

LSP: In the classic corporate structure, you have marketing, who say "This is what users want" then they synthesise that or clarify it for the engineers; then the engineers say "You've gotta be kidding!" Well, in open source, "You've gotta be kidding!" is phrased a little more harshly, as volunteer developers can just walk away because he or she has no interest in doing that ­ they want to do something that's more technologically interesting for them. This is the classic liberal approach (in the sense of liberty, rather than politics) and much of open source is these days conducted by people used to working in this fashion.

When a company is allocated a certain number of developers to work on a particular area, there can be tension as the developers aren't necessarily that keen to let the company allocate resources or dictate which areas of the project on which work has to be done. These areas aren't always that attractive to developers, because there is less opportunity to innovate. Many developers don't want to just produce open source equivalents of functionality already provided by programs from companies like Microsoft and Adobe; they are motivated by the opportunity to implement something that's cool. If marketing doesn't consider this to be worthwhile, or consumer group can't initially see the appeal for end-users, then commercial concerns can end up limiting the resources allocated to evolution of new features, and the developers end up having to work on these areas on their own rather than with the support of the sponsors. A lot of open source developers ­ both paid and volunteers ­ might be frustrated at this situation, but realise that this is the way things work at present, and just get on with it anyway.

LXF: Do you think that the PR aspect of sharing code with the open source community is an important motivator to do so?

LSP: I would be sceptical about how this would apply to the market as a whole, but it certainly has its advantages. Microsoft spends more money on recruiting developers than a company like Sun can. But you don't want Microserfs though: as a recruiter, you want developers that appreciate the technological and social freedom that open source provides ­ the PR aspect is important way of attracting innovative and interesting people; Sun's accomplishment is proven in that it's getting a lot of good developers now..

In contrast to what I said earlier about enterprise having too much influence in aspects of a project relating to their own profit model, it would be true to say that involvement of profit-driven companies can provide important direction for open source projects. Look at Debian...

LXF: Debian? Ha! Ha!

LSP: You may laugh! Debian is an experiment in democracy run amok. It's very proud of that: it's radical, but such a rampant democratic approach but can dissuade businesses' opinion of open source as it can impede progress. Development doesn't take place in a vacuum; in business, to actually launch products you need someone who's responsibility it is to say "this is where we're going," and "this is what we do next!" ­ that's why Mark Shuttleworth (supremo of Ubuntu, based on Debian) has been so important for Linux.

LXF: It's interesting to note that though ODF is not supported directly by Windows Office 2007, Microsoft finances the ODF add-in for Word project on SourceForge to create a plugin for Microsoft Office that will be freely available under a BSD license. How many companies understand that?

LSP: As open source becomes more widespread, companies are getting a feel for what it's about, and are headed in the right direction. When they recognise that they don't have to do all their research in-house, everyone wins ­ if they fund an open source project, it's like starting a subsidiary business, but with a lot less risk and is in many cases more cost-effective. LXF