Linux Standards Essay 1

Michael Bacarella wrote in with the first essay on Linux Standards. I've got several more of these left to post today, and I'll continue to post them as I have time.

The following was written by Slashdot reader Michael Bacarella

Certainly a topic of much bile and hatred.

This will be long, hopefully you'll have the patience to read through it all. :)

One might ask at first why we even need standards. I'm sure a lot of you know just why, but there are plenty out there who honestly can't see why this is such a big deal to us. From their perspective it looks like we're childish arrogant bastards who are frothing at the mouth over nothing.

Quicky: Linux is not centralized. That is an amazing strength as well as a weakness as it allows Linux to be used for almost anything. Super-computer to clustering to desktop PC to 386 resurrector to PalmPilot OS. While these all do not have official distributions, it's part of the principle.

The problem here is that when one develops under one distribution, there's often no guarantee it will run on other distributions or even different versions of the same distribution. Keep in mind that most Linux software runs fine between these distributions. Software authors with lots of Linux experience generally write cross-Linux as well as cross-Unix code.

Standardization for hardened Linux developers would merely make the job easier. If it's easier to write it tends to get done faster, if it gets done faster the programmer has more time to experiment and debug (or not fail school or raise children or other equally important things :)

The maddening standards debates didn't begin until (to my knowledge anyway) commercial software developers saw an interest in Linux. Most of the times these Linux ports tend to be aimed at a specific distribution as well as a specific platform. This happens because "commercial" software developers usually don't have the time or energy to learn Linux well enough to write a proper port.

This isn't an issue about standardizing Linux. It's an issue of making Linux easier to learn and develop for. When I first picked up one of those cryptic text editors and COMMAND LINE based compilers, I had no idea how to write anything. Where are the threads? What the hell are signals? WTF is fork? Why the hell are there all these zombie processes? (How come my system is no longer usable? :)

I wasn't up to speed for at least 10 months. Granted I wasn't giving it my full attention and total dedication but you still get the idea. New Linux apps are about a year behind the Windows world, if ported. As Linus himself even said, the problem isn't that it isn't there. The problem is that it isn't documented well enough. Commercial software development isn't undertaken by skilled hackers and kernel tinkerers.It's taken up by people who want to finish the port and get paid and go home. If we want them to work for us efficiently, we need to provide them gobs of documentation and tutorials so they don't have to think more than necesary.

No I'm not calling all commercial developers idiots, but you can see where I'm coming from here can you not? When you take it in this light, the only standardization Linux really needs is cross-distribution collaboration. It's in every distributors' best interests to make their distribution compatible with the rest. We don't need a standards organization we have to pay to do it for us, nor do we need super commercial giants involved.

We succeed on globalized teamwork and cooperation. Forcing us to all conform to a standard is the opposite of this idealism. This is how the Corporate world tries to work. We shouldn't settle for that. There are better ways, and there's no reason for us to become stupid (stupid as in doing what's technically bad for the system) all of a sudden.

A standard base wouldn't be such a bad idea, but as many an editorial on freshmeat stated: The LSB will succeed on how little it defines, not how much it defines. (I probably junked that up, but the idea is the same). A standards organization is going too far, and the idea bothers me to no end. Not because of 'Linux Standards' and it's rather "misguided" methods of going about it, but rather because it's a technically bad idea. (tm)

I'll use the OpenGL Architectural Review Board as an example. The ARB is a group that meets every so often to discuss what happens to OpenGL. This sounds innocent enough, but it's current members are Microsoft, SGI, Evans & Sutherland, HP, IBM, Intel, and Intergraph. The fate of OpenGL is decided by these people. How would you feel if Linux's future was decided this way as well?

Couldn't you just picture Microsoft and several other Microsoft-controlled Linux haters becoming Linux Standards members and running Linux to the ground? I sure could.

The future of Linux will be decided by it's users. That's the way it has always been, and short of me being thrown in prison, that is how it shall remain.

An Operating System of the people, by the people, and for the people.

Michael Bacarella
Systems Administrator / Development
New York Connect.NET