Team Murder No Brain No Headache.


The Desktop That Isn’t And Shouldn’t

For the first time ever (I think) in the four year history of Team Murder I'm going to link to an article at OSNews that I agree with for the most part. Advocating Linux as some kind of technological panacea for all that ails both corporate/business and personal computing has always seemed like a pretty stupid idea. One, because the user isn't going to change and that necessarily leaves the obligation to adapt to the distribution instead.

I think it's become pretty obvious at this point that the strongest development done on either the kernel or the hojillion pieces that make up a Linux distribution may well be done under the auspices of a corporate benefactor but I have yet to see a commercial and end user focused distribution do very well. The ones developed commercially to begin with end up penniless and with a broken distribution where countless man hours have been wasted customizing desktop environment windows and wrapping sudo around fucking everything and distributions that have been forked or acquired seem to fare a little better on the server side of the table because they offer actual technical support when a production machine goes belly up whether or not that support is used. Big organizations love to have someone to yell at preferably attached to a toll free number. Anyway, right now a lot of companies with money to throw around are backing Linux development which is great because it subsidizes important and low level stuff that hobbyist hackers just don't have the resources to tackle as efficiently. That association seems to generally work: salaries are paid, code is hacked out full time, and everyone benefits whether they directly interact with that software or not -- chances are that they do without ever knowing it.

The problem with tailoring anything towards users is that users neither know the limitations of programming languages that are passed along to any given pile of software written in that language nor exactly what they want aside from vague suggestions like 'I want it to work with everything.' Taking these sorts of complaints as anything other than general feedback is the part of the reason that so many operating systems end up nightmarish and full of hacks. Part of that has something to do with the fact that 'desktop' is the most nebulous term every imaginable. What exactly does it mean when you say 'aimed at the desktop user'? Everything and nothing.

All of this is, of course, immaterial as people seem more than willing to launch projects with this imprecise audience as their intended target for effort. One of the concepts important to this that Martin Girard expresses very well in his post is the intentional accessibility and visibility of the guts of a Linux distribution to its user and how important this sort of accessibility is to keeping the geek crowd interested in an operating system and in developing for the operating system. While APIs are nice as a way to avoid unnecessary complexity when dealing with interface issues it is also preferable to see how the lower level components actually interact with each other. No sane person wants to develop exclusively on a platform that severely limits a programmer's access to a handful of methods that are determined, without any real developer input, behind closed doors. Some people are willing to balance between the two in order to develop commercial applications which more often than not requires the use of tools controlled by the same company.

The reason that I find myself agreeing with most of the what is said in the article is that people aren't necessarily talking about something that really exists. They're speaking about an idealized conceptualization (as incomplete and backasswards as that may be) of how an operating system should work for them which isn't implemented by any of the big players or even the more marginal entries. I think Linux development, more than any of the others that have an interest in the desktop market, will be more likely not to make stupid and hasty decisions based on the demands of users. Does that make Linux development inherently user hostile? Maybe but that also creates an environment more accommodating for people who develop software to use on the operating system and one that doesn't have the slew of exceptions to common sense rules of low level design. It makes sense to the people who build and use it which is far more than I can say for the commercial variants.

Filed under: General Comments Off
Comments (0) Trackbacks (0)

Sorry, the comment form is closed at this time.

Trackbacks are disabled.