From: Eivind Eklund Date: 2004-10-15T03:11:00+09:00 Subject: Re: Eating sandwiches in the park On Fri, 15 Oct 2004 02:51:07 +0900, Brian Wisti wrote: > I like the sandwich analogy. I may have taken the idea into a new and > more nauseating direction, but the analogy works. Code which abuses the > flexibility of the language will not spread very far, simply because > people won't use it. That has been my experience so far over the last > few years, and I don't expect that pattern to change any time soon. If > you see code which has had such abuse inflicted on it, you need to > either fix it or avoid it. You're certainly not forced to eat it. I'll just have to say that this does not match my experience. I regularly end up being forced to use and fix software that has serious problems because somebody else has made the decisions. Examples: - Perl, where the interpreter design and language is just icky, but I've had to deal with it because there are applications/modules my employer needs to work - MySQL, where a migration is too expensive - The NFS code for FreeBSD, where we had a spaghetti of C macros making the code almost impossible to deal with, but cleaning it up would result in divergence from NetBSD/OpenBSD so we had less bugfixes. - Donald Becker's ethernet drivers in Linux - the code was fairly horrible, but they were the only source of some functionality. - A million lines of C++ code in various styles and for various projects at a former employer, where I did troubleshooting and ended up debugging many of the "funny" problems that the people that had written the code couldn't debug themselves. If you don't have to deal with other people's bad choices on a daily basis, you live a sheltered life. When it comes to dealing with Ruby code: Read my .signature (though we do some selection, we're also committed to providing packages over time, so giving the authors enough rope to hang themselves is bad for us.) Eivind. -- Hazzle free packages for Ruby? RPA is available from http://www.rubyarchive.org/