One of the things that depresses me about the software industry is the transformation of version numbers from being a useful indicator of a software package’s maturity to marketing tool. In the commercial software world, all bets are off with regard to version numbers. Windows NT debuted at version 3.1. Netscape has been skipping version numbers like crazy in recent years. Plenty of other companies muck around with their version numbers as well, basing them as much on the version numbers of their competitors and how they want their software to be percieved as on the actual number of releases of the software itself.
In the open source world, though, there’s still some integrity with regard to version umbers (even if there’s not consistency in conventions between projects). For example, when the 2.5.x series of development Linux kernels is released, will the release version be 2.6 or 3.0. Based on this KernelTrap report, much wrangling is going into this decision. If anything, the Linux kernel version numbering scheme is a bit too conservative. Given the degree to which Linux has changed since it was introduced, it could probably easy be a few rev numbers higher than it is right now. Other open source projects mirror this humility, Mozilla debuted properly at version 1.0 (even though its lesser sibling is called Netscape 7), and for a long time Perl was issuing fairly major revisions and only incrementing the second sub-version when they did so (that was actually confusing).
This entry is really nothing more than a lament, the world of commercial software certainly isn’t going to change back to responsible version number usage anytime soon. If anything, the move is toward dropping numbers altogether and using things like XP, MX, etc.