From: "Economou, Matthew" To: "'malda@slashdot.org'" Subject: Possible concerns with Linux standardization (my editorial) Date: Thu, 20 Aug 1998 11:27:35 -0500 Rob, I'm not certain the following editorial delivered to you (I've also modified it slightly). You'd requested comments about Linux standards, and here's my best shot. Hope it works! --matt (aka Xenophon Fenderson) >From the Desk of "Xenophon Fenderson, the Carbon(d)ated" , http://www.cs.rose-hulman.edu/~econommx/: It'd probably be a good idea to read this with "IMHO, " prepended to each sentence. :) Many Linux hackers feel that standardizing what gets installed will reduce the degrees of freedom the heretofore open system has provided them. Stallman's concept of /libre/ is very important to many in the Linux crowd, and often times this translates into a 'damn-the-torpedoes-full-speed-ahead' attitude regarding "rules" and "regulations" and "the-way-to-you-must-do- things" (versus the hackish "the-way-things-should-be-done-to-be-done-right,- obviously"). (It is my (possibly flawed) understanding that this attitude led to the various Linux-specific releases of various components of the GNU toolchain, resulting in the Linux-libc and the host of Linux-specific patches to binutils, gdb, gcc, and so forth.) I also think "Linux standardization" (as the effort to specify what all gets installed and where everything goes is called) scares a lot of hackers (if my own initial reaction to the LSA is any indication). I think that many of us fear being left out (even if we've only contributed to the GNU/Linux body count). Hackers tend end toward a quasi-anti-commercial attitude, and Business and Government taking the ball and running with it (yay!) make many of us (or maybe just me) freak out (oh my God, they're stealing MY system!). This isn't to say that Linux standardization is a bad thing. It allows developers to expend less effort on supporting binary releases of their code. Both J. Random Corporation AND J. Random Hacker benefit...while I personally prefer to compile everything myself, binary releases are often more convenient to install (e.g. I *could* compile X11R6 myself, but I really would rather not wait that long). It would relieve commercial developers of having to worry about where to dump log files, or what library bugs to work with/around, or any of a hundred other trivial details that make up many of the differences among Linux distributions. I don't know where I stand on the issue of Linux standardization bodies. I personally belive Bruce Perens' effort to be an exercise in ego, and I think the LSA is ignoring the community that (literally) made Linux the success it is today. WRT to the LSA's membership fees, the organizers of the LSA better realize, and do it soon, that many of the Linux developers have worked for free, /gratis/, and that to charge money for membership in the standardization body is philosophically incompatible with Linux specifically and GNU in general. In my mind, it would be like Linus trying to charge people for the right to develop kernel code. (As an aside, I also think that any Linux standard must be consensual, because of the nature of the community. Heck, maybe Linux standards are a kind of social programming. I can hear it now... people will be petitioning Goddess to release the source code to our branes! Memetics, anyone?) I haven't had a chance to look into the most recent effort (the collaboration between Debian and Red Hat), but I hope it will have a more open attitude towards both the Open Source community AND any commercial/governmental entities that care to contribute. And besides, any well-written program should be able to cope with default configuration paths changing, right? ;-P