Replies

You can talk a lot about best practices and libraries and everything else, but the punchline is that unless people know there’s something in house that’s already solved Problem X they aren’t going to go looking for that code to use, they’re going to write their own.

That's been my experience. Ada, the language put out by the DOD for bid and designed & built by multiple committees was supposed to address that. "A camel is a horse designed by a committee". And a camel is what they got. The earliest implementations, with perhaps the exception of Verdix Ada (after it matured a bit), were mostly pretty bad. A few iterations (and a lot of in-the-field experience) later, Ada has gotten pretty good.

Of course, the proliferation of different software licenses on code hasn't helped either. Basic example - you cannot incorporate GPL code into BSD licensed code (the GPL is a viral license that requires all of the code it is linked with become GPL), however you can incorporate LGPL code into BSD licensed code. Some Open Source projects, notably from the FSF are even stricter, they not only require GPLed submissions, but the copyright must also be assigned to the FSF.

Certain software projects like XEmacs (which was a fork of GNU Emacs in the early 1990s by the now defunct Lucid) require GPL licensed contributions. Others like Emacs require not only GPL licensed contributions, but they must also be copyright assigned to the FSF. When Lucid went bankrupt and orphaned XEmacs as Open Source in its bankruptcy, it was in the position of being GPL, but unable for the most part to share code back with the parent.

There is a time and place for lawyers, but in the field of software development their involvement has been little short of disastrous.

Been there, done that, tried to change the world and it didn't quite work out, although I'm quite proud of how XEmacs has turned out. It's still my favorite text editor. (grep for "altair" in XEmacs ChangeLog files, that's me).