[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

The flaws page

This is the page what IMHO is wrong with present distributions (if I
everything was right I wouldn't be doing Indy).

Title: Flaws in the development model for Linux desktop distributions

Flaws in the development model for the Linux desktop distributions

Linux has enjoyed a fantastic success in the last year, however its users have been hackers. Another step is now needed: bringing Linux to the desktop and to the common user. This will require not only easiness of use but a willingness to cope with the needs of those unusual (in the Unix world) users. Unfortunately the development model of present distributions both free and commercial is not suited to that new challenge. A newer one is needed.

Commercial distributions

In the last years commercial distributions have made enormous progress in the easiness of use field. However a commercial company goal is to make money and that sets two limits at their progress towards the desktop and personal computer:

Marketing driven features

A feature will be introduced according to its ability to generate revenue not on its usefulness. A pretty installer will get positive reviews in the press and thus will get highest priority. A feature who only be noticed through day to day use will get medium priority. Making the box robust to user mistakes is crucial in a system like Linux where many users have to act as system administrators from thee first day. However user does not know when distribution fixes a mistake he made who would have left him with an unusable machine had not the distribution defused it. Therefore those "invisible" features get very low priority.

Money in Linux comes from server not desktop

Nowadays CD burners are standard equipment, ADSL allows to download entire distributions and you can find Linux CDs even in the Ziff Davis press. That makes selling boxed distributions a far more difficult proposition than in 1995 when it was difficult to get Linux without buying a distro. Therefore distribution vendors rely on support for revenue. But private users and even corporate users when application is not critical can get very good support from the Net and supporting a large number of users who pay small sums per incident (typical of private users) requires a large infrastructure. It is for critical applications (ie servers) where companies will adopt a policy of "take no chances" and will pay for support. It is for such clients such clients when Linux companies will be able to charge large sums to a small number of customers. Since the desktop and private user don't pay the bills for Linux companies these will concentrate on the server. They could use the desktop to reach notoriety (eg Mandrake) but sooner or later they will reorient towards the server. Or die

Present non-commercial distributions

Non comercial distributions are not under pressure to make money and thus we could expect from them to be the spearhead of Linux in the desktop however it is not the case. Slackware, Stampede, Debian reputations are not precisely about easiness of installation and use. Part of the problem arises from the fact that they were founded by people who were "University Unix". Therefore they gave little thought to usability problems who are vital for those people who don't have a teacher to soften the learning curve and a sysadmin caring for them. Their founders were IT professionals or future professionals and reading dozens of HOWTOs was natural to them so they forgot that non-IT people have other priorities. Finally they ended making their distribution an end in itself and fulfilling its goals became more important than contributing to Linux expansion. "We don't care about usability since we do't try to make a distribution for beginners" is a sentence you hear frequently between activists of non-commercail distributions.

Developers living in another world

A common problem to commercial and non-comercial distributions is the people making them are different from their users and live in a different world. These are people with strong backgrounds in Unix programming and system administration while Linux users are far more diverse with many people being Windows or mainframe transfuges and having different needs and tastes from those of the Unix people. Environmental conditions are also at work for making distributions miss their target. Let's take an example from Mandrake, a distribution who is exceptionally creative and close to its users, however it falls very short for SMB (networking in Windows environments): you don't get the PopUp messages sent by printer servers to tell you your printing job is finished and the network browser they ship works in laboratory conditions but not in real world. Why people with as much flair as Mandrake people have done a so poor job? Because they have never used Linux as a client in a Windows centric company thus they don't know about printers who are twenty meters from your office or about network browsers who don't scale beyond tiny networks. And other distributions are far worse. That is why there is a need for a user-developed distribution: because devlopers living in a Linux stronghold cannot understand user's real needs.

Conclusion: The need for another model

As long as distributions can pay the bills only by concentrating on the server then significant progress on the desktop could only be made by a free distribution. The existing free distributions have an aristocratic spirit who makes them unsuitable so a new one is needed. It should be created in a spirit where developers pride doesn't arise from building a cathedral-like distribution but in doing their utmost for propagating Linux and helping people. It should not be a distribution where users passively wait for what developers build for them but one where they take an active part in development. Finally it should not be developed exclusively by Unix people (whose competency is still strongly needed), it should be developed by all the people who compose the Linux mosaic: Unix people, mainframe people and, shock, horror, people fresh from Windows.
That is what we try to do in Independence