Re: EMC version numbers

[shackle <shackle-at-cme.nist.gov>]

I tried to do a major number minor number thing with the RCS library releases.
Although this is certainly what many people expect from experience with
commercial software. I am not sure that it has worked out very well for
software using the open source release early/ release often strategy.

What happened to me was that almost all revisions were minor revisions. Since
each release adds at least a couple new features the bug # is constantly reset
to zero. Also if we release often there would never really be a major release,
but instead many minor revisions would add up to  a larger difference than you
might have in a single major release. Even though I arbitrarily occasionally
called a release "major", the version numbers were ussually not as helpful as
the dates they were released. 3.72 (the last release before an arbitrary
major.) is actually closer to 4.00 than 4.00 is to 4.32.

I tried freezing one release and calling that the stable release. But since I
ended up for the most part only bothering to fix bugs in the current release.
The current releases ussually had about the same level of bugginess as the
stable release and since it had more features very few used the "stable"
release. I am sure there is a point where that method becomes useful but I am
not sure if EMC has reached the
critical mass  where benefits out weigh costs  to maintain seperate stable and
development versions as they do for the linux kernel.

-- Will






  

-------------------------------------------------
William Penn Shackleford III			shackle-at-nist.gov
National Institute of Standards & Technology	Tel:	(301) 975-4286
100 Bureau Drive Stop 8230			FAX:	(301) 990-9688
Gaithersburg MD  20899	USA
http://www.isd.mel.nist.gov/personnel/shackleford/
Office Location: Bldg. 220 Rm A253