Changes at Mandrakesoft

In recent weeks, Mandrakesoft has announced several wide-ranging changes affecting everything from the company's development model to incorporation of new technologies, and even its name. We have attempted to read between the lines of Mandrakesoft's press releases, interviews, FAQs, and IRC discussions, and this is what we think.

First, the good news: Mandrakesoft is doing well. The company has recently been awarded two multi-million euro contracts by the French government and it is likely that private enterprises in France have also started to contribute towards the company's positive cash flow. As a result, there has been a shift of focus by Mandrakesoft from developing a predominantly home user's product into more profitable enterprise-grade solutions and support. This is hardly surprising as -- and let's be honest about it -- that's where the real money is. If this model works so well for Red Hat on the other side of the Atlantic, there is no reason why it shouldn't work for Mandrakesoft, albeit on a somewhat smaller scale within its own sphere of influence.

This success is probably the main reason behind the latest round of changes in the development and release process of Mandrakelinux. Since the company was established in 1998, Mandrakelinux releases came out in regular 6-month intervals, but the high release frequency of two architectures has been putting strain on the developers, as witnessed by the delays in each betas and release candidates of all recent Mandrakelinux releases. From that point of view, the newly introduced annual release plan will make sense. Unfortunately, it will probably alienate some users many of whom have perceived Mandrakelinux as a solid, up-to-date distribution with frequent releases incorporating all the latest Linux technologies. Especially the current Mandrakeclub members will have a reason to complain since the €120/year membership fee originally entitled them to two Mandrakelinux releases per year. As a compromise, Mandrakesoft is now offering to fill the gap with an interim product - just for the club members. Even so, the skeptics will argue that this is likely to be a poorly-tested snapshot of the development tree, which has historically suffered from stability issues.

How the acquisition of Conectiva fits into Mandrakesoft's future plans is less clear. Although Conectiva employs many talented developers and has a history of several successfully implemented ideas (the port of Debian's apt to RPM-based distributions springs to mind), there seems to be little that the Brazilian company can offer Mandrakesoft. Also, as anybody who has worked for a multi-national software company can confirm, managing software development in a country halfway across the globe will almost certainly result in a substantial overhead in terms of traveling, communication, and bandwidth cost. Add to it the language barrier, and the benefits of acquiring the services of a few dozens of talented developers can be easily overshadowed by the increased expenditure. As such, it seems that Mandrakesoft's acquisition of Conectiva is largely a public relations stunt devised to convey a message saying that "Mandrakesoft is back" - healthier and more profitable than ever.

That said, some of Conectiva's ideas might end up being incorporated into Mandrakelinux in one form or another. The Mandrakesoft developers have hinted that they are examining some of Conectiva's kernel hacks and evaluating the possibility of incorporating elements of its package management into Mandrakelinux. But will Conectiva's apt replace Mandrakelinux's urpmi? There are reasons to believe that it might. Although both apt and urpmi are released under the GPL, urpmi is not used by any distribution outside Mandrakelinux, while apt is widely deployed by many RPM-based projects and it even became a very popular third-party package management tool for Fedora Core and SUSE LINUX. In fact, several distributions that were originally based on Mandrakelinux were quick to drop urpmi in favor of apt (e.g. PCLinuxOS or ALT Linux). There is little point for the unified company to continue developing two package management tools, so if one of them has to go, it will likely be urpmi.

Besides the major modifications in its development model, speculations are rife that the company will also change its name. Shortly after acquiring Conectiva, Mandrakesoft registered several top-level domain names for Mandriva, as well as a large number of regional domain names in many parts of the world. Of course, this is less surprising given the long-standing trademark dispute between the company and a US-based syndicate holding the rights to the comic-strip character "Mandrake the Magician". If the name is indeed retired, it will mean the end of one of the best-known and best-loved brands in the history of Linux distributions.

How to keep its existing user base in the atmosphere of frequent release and development model changes is an important challenge for Mandrakesoft right now. Lack of predictability is starting to become a major weakness of the distribution, especially when compared to some of its competitors that have clearly defined release processes and support periods. But if Mandrakesoft can get more business from large enterprises, losing a few home users to other distributions will be a small price to pay. In this respect, Mandrakesoft is wisely following in the footsteps of Red Hat and Novell/SUSE, especially if they can stick to the current plan and resist introducing any major new changes for some time to come.

---

(Log in to post comments)