The Great Leap Backwards for Open-Source

The Schema Czar watched the 'Open Source' movement from its onset with interest and sympathy. Now, I need to support my family.  Working for "free software" never quite made it for me personally.  But when I started, the term "hacking" was as yet unblemished - whether by computer criminals or by poor programming practices.  It was cool looking through Unix source code and finding #ifdef's for the PDP-8 and Honeywell computers.  It was neat that when you went through manpages, you would find things from several universities as well as commercial entities. As the new millennium entered, the tools and collaboration mechanisms for open-source matured along with machine and network capacity.  Linux had proven itself, and communities around Linux, Java, and other base technologies were asserting themselves as a force for delivering software.  I was cautiously optimistic - and cautiously cautious, because I still needed to earn a living! What would happen? Now I think I see what has happened.  Open-source software stagnates.  Critical tools bit-rot and die.   For example: Rather than fix CVS, (which was itself a refusal to fix RCS,) we got Subversion.  When Subversion failed to follow through on its excellent start, we got external tools like svnmerge implementing things that should have been in the base implementation.  Now we have still more version systems. This isn't some ancillary tool - software development organizations are LIVING on a version control system that is not being enhanced and barely being maintained. Perl is now TWENTY years old.  PHP is what Perl should probably have become, but it didn't, and whoo-hoo, now we have Python.  All three are gappy, idiosyncratic and sclerotic. If I see yet another "improved" markup language like Wiki or ReST, it will provoke me to violence.  Why don't any of these open-source luminaries fix someone else's code before starting from scratch? If this is what Java has to look forward to, we're in trouble.  The resiliency heretofore of Linux and MySQL are the only bright lights, and how much of this is due to for-profit sponsors? Open-source has a huge streak of "not-invented here," "I want to be the famous creator of a great new thing," and "I'm too good to maintain my own code" attitudes.  These are destroying its value proposition for business.  Not only do you have to customize it before it's usable to begin with, you're taking on a full plate of maintenance work once the "open-source community" has lost its infatuation with a technology.  That is, 15 - 21 months after an open-source technology gets popularized, it's on its way out. Open-source is poised to leap backwards in one of two ways:  either it will keep going backwards to the same practice of reinventing a wheel that could be repaired, or it will embrace a a positive approach to backwards-compatibility.