LWN.net Weekly Edition for March 23, 2006

The Global Desktop Project

Free software has always seemed like a good match for the developing world. It makes top-quality software available to all, without forcing choices between buying expensive licenses (using hard-to-get foreign currencies) or dealing with the the consequences of wide-scale copying of proprietary code. Free software is one bit of technology which is just as available in the poorer parts of the world as it is in the richer countries.

What has been observed, however, is that, while use of free software in the developing world is taking off, participation in the development process is not growing at the same rate. This is true even in countries where there is no shortage of people with the technical skills needed to hack on free software. To a great extent, much of the developing world is a consumer of free software, but takes a relatively small role in its production. This costs the development community, which has no end of projects which could benefit from more developer attention. But it especially hurts the people who don't participate. A consumer of free software remains dependent on imported code without developing the ability to improve that code or influence its further evolution.

The United Nations University (UNU) recently sent out a press release on this issue:

The UNU, working with governments and industry, has set out on an effort to improve participation rates in the developing world. Part of this effort is the Global Desktop Project, an initiative to increase the number of free software hackers by encouraging improvements to the Linux desktop experience. The leader of this project turns out to be a familiar name: Scott McNeil. Among other things, Scott has served as the head of SUSE's US operation, the "open source strategist" for VA Linux Systems, and the executive director of the Free Standards Group.

Scott tells us that the desktop focus was chosen because the desktop tends to be one of the most interesting areas for aspiring hackers. There is also a great deal of desktop-oriented work - such as internationalization - which is best done by locals. Within this focus, there are four separate initiatives being pursued. These include a "train the trainers" seminar series designed to help spread the free software methodology, the establishment of a set of open source labs, the creation of a series of FDL-licensed courses, and a student mentoring effort.

The first labs are expected to open in China later this year; they will be financed from initial funding received from Intel and a couple of Chinese government agencies. The labs will hire students - mostly at the graduate level - to work on free software projects. Lab staff will also work on mentoring to help new developers work with the community. Says Scott:

Working with the community is often one of the biggest stumbling blocks for developers coming from outside of Europe and North America. These developers have all the technical skills they need, but there can be a strong impedance mismatch between the culture they grew up with and the often, um, impolite nature of discussion in the development communities. Being flamed on a mailing list is unpleasant for most of us (though there do seem to be people who live for that experience), but it can be shocking to somebody from a culture where people do not talk to each other that way. It can also lead to workplace difficulties. Even more gentle, well-meaning criticism can be problematic for some developers.

So developers from those areas tend to avoid the community - and not contribute back improvements they may have made. And that hurts everybody involved. That is why, as Scott says, "the Global Desktop Project is as much a socialization exercise as it is an engineering project." It is an effort to integrate these developers into the growing worldwide development community - a result which should be beneficial to everybody involved.

This project - which is expected to last three to five years - is just getting started, so it will be a while before results will be visible. There will, doubtless, be many cultural and funding hurdles to overcome during that time. But, if all goes well, the Global Desktop Project has the potential to increase the rate at which the developing world joins the free software development community. And, as a bonus, we might just get a better desktop out of the deal.

Comments (8 posted)

Xara releases the source

Xara Xtreme is yet another drawing and image composition tool for the Windows platform. A few months ago, Xara announced that it would branch out and make this application available for Mac OS and Linux platforms as well. Even better, it would be released under the GNU GPL. The result, it was said, would be a top-quality drawing tool for the free software community.

The first part of that promise has now been fulfilled: the source code is now available for the project now known as Xara LX. This version of Xara, it turns out, is a GTK+ application by way of wxWidgets. It comes with plenty of warnings: many of the features are not yet ported, and the whole thing can be somewhat unstable. But, the tool is now out there for people to play with.

Your editor has a hard time resisting an invitation like that. The unstable Xara LX build ran nicely, and it was easy to put together a simple drawing with features like transparency and blending. It was also not particularly hard to make the whole thing crash. But, suitable warnings had been given; this tool is not being provided for production use at this time. For an example of what can be done by users who know what they are doing, see this screenshots page.

Once it stabilizes, the Linux community should have another nice drawing package in its toolbox. Linux may not yet be poised to displace proprietary packages from the systems used by professional artists, but things are clearly headed in the right direction. With tools like the Gimp, Inkscape, Krita, and, now, Xara LX, we are getting closer to the day when there is no need to use those other, proprietary platforms even for the most demanding graphical tasks.

Comments (5 posted)

Money trouble at OpenBSD

The OpenBSD project sets the standard for security in free operating systems. More than with any other project, the OpenBSD hackers work at tracking down potential security problems before they affect users. This work has earned OpenBSD a well-deserved reputation for being hard to break.

A recent posting to the openbsd-misc mailing list has raised a non-technical issue: it seems that OpenBSD's finances are not as solid as its software. The project has been running at a $20,000 (US) annual deficit for the last couple of years, with no relief in sight. The problem, it is said, is that OpenBSD users have stopped buying CDs; instead, they content themselves with grabbing a copy from a network server for free. The sales of CDs and related items are a major source of money for the project; if CD sales do not live up to expectations, income will fall short.

LWN asked the OpenBSD project if there was any sort of public information on the group's budget and how it is spent. Unfortunately, it seems that there is no such information. From looking at what information is available, it appears that the biggest single expense is the occasional "hackathons" - coding-intensive developer meetings - run by the project. Beyond that, there's the usual costs for Internet service, equipment, and so on. It appears that very little of OpenBSD's budget goes toward paying salaries to developers.

To support its activities, OpenBSD would like to bring in about $100,000 per year. Donations recently have been a very small fraction of that, however. What the OpenBSD folks are saying now is: something has to change, or the project will be unable to continue at its current level of activity in the future.

Every free software project must support its work somehow. For small projects, that support may consist of no more than the occasional donation of development time by an interested hacker or two. Larger projects require more, however, in the way of infrastructure and developer time. So most projects, once they achieve a certain level of success, have to find a revenue stream from somewhere.

That, often, is when the core developers try to form a business around the project. It may just be a matter of lining up some consulting work to pay for the continued development and maintenance of the code, or there may be a more advanced business plan involved. Sometimes projects are able to obtain sponsors which have some interest in the project's success; witness Google's support of the Mozilla Foundation, for example. Sometimes developers will be hired by a company to work on free code; many Linux kernel hackers make their living in that way.

It is a rare project, however, which is able to get very far with sales of CDs and donations. There is little motivation for anybody with a broadband network connection to order CDs; a simple download gets more current software more quickly. This is why Linux distributors have been moving away from CD sales as a business model for years - and those which haven't are wishing they had. The OpenBSD project is simply discovering the same thing others have found out: the value of a CD is quite low. Anybody who is in the business of selling CDs full of free software is in a commodity business, and one which is in competition with its own customers at that.

There is no other business of any consequence built around OpenBSD, however. There are few products which incorporate OpenBSD, and few high-profile network-based services which use it. While OpenBSD does not lack for users, it seems there are relatively few who see a business interest in supporting its development. It must be said that the abrasive nature of OpenBSD's leadership cannot be helping in this regard.

The same posting hints at one approach for generating some cash:

The project, in other words, is appealing to "entities" which obtain some value from OpenSSH to kick into the OpenBSD coffers. It is hard to imagine that, for example, Linux distributors - all of which distribute OpenSSH - are not among the "entities" being targeted here. This is just a bit ironic, given how the OpenBSD founder has chosen to trash Linux recently.

More disconcerting, however, is the implicit threat: support OpenBSD, or OpenSSH may go down the tubes. The answer the project is likely to hear may not be the one they are looking for; the world may ask, instead, why are OpenBSD and OpenSSH funded from the same pool of money? Might it not be better to separate the two - by forking OpenSSH, if necessary? Certainly some way could be found to keep OpenSSH going if OpenBSD were to come to an end.

The end of OpenBSD would be an unfortunate event, however. The project's uncompromising focus on security has raised the bar for all systems and made all of us - even those who have never run OpenBSD - more secure. We all benefit from having a group out there doing the work that the OpenBSD people have taken on. But it is up to the OpenBSD folks to put some of the same attention into securing their financial future, and that means finding a way to obtain money from those who benefit most from OpenBSD's existence. Given the size of the OpenBSD user base and the modest nature of its financial needs, it seems like this problem should have a solution.

Comments (31 posted)

Page editor: Jonathan Corbet

Inside this week's LWN.net Weekly Edition

• Security: One year of RHEL 4 security; A vast number of scary new vulnerabilities.
• Kernel: What's coming in 2.6.17; The last-minute unshare() discussion; Starvation in the scheduler.
• Distributions: New releases: Fedora Core 5, CentOS 4.3, Mandriva One, Pie Box Enterprise Linux 3 AS7; DPL elections; Dapper in June; FUDCon Boston
• Development: The Firefox Bon Echo Alpha 1 web browser, Roxkbox Release Schedule, Zenoss project, new versions of Rivendell, LAT, Sussen, mnoGoSearch, DANCE, PythonCAD, Lightning, GNOME, GARNOME, kicad, Wine, Freecycle, MusE, Azureus, GPA, Camomile, Urwid, Valgrind.
• Press: Open-source business, the Maximum Net, coverage of EclipseCon, Novell BrainShare and PS3 conf, Duval to sue, Dutch Court upholds Creative Commons, interviews with Eben Moglen and Jack Messman, GNOME 2.14 review, free Unicode typefaces, FOSS community and the disabled.
• Announcements: MySQL Joins Eclipse, Novell Launches Open Enterprise, Linuxaudio.org progress report, open source tax credit proposal, Jolt Awards, numerous calls for papers, Akademy 2006, Embedded Linux Conf, FUDCon, Gelato ICE, ETech keynote audio.

Next page: Security>>