CLIM mail archive

[Prev][Next][Index][Thread]

labels for panes



   Date: Wed, 23 Sep 1992 14:43-0400
   From: John C. Mallery <jcma@reagan.ai.mit.edu>

   Personally, I would like to see a release strategy with smaller, upward
   compatible releases with higher frequency than big, possibly incompatible
   releases over longer periods.  This is good for two reasons:

	   1) Steady progess is observed by users, inspiring confidence
	   that things will get (even) better;

	   2) Users are confronted with a smaller number of things to learn 
	   with each release, and consequently, will make use of them.

   And if they don't want to update their code, that is there choice; they can
   wait for a few more minor releases.

There is a substantial fixed cost (i.e. almost independent of the
"size" of the release) in doing a software product release.  QA,
documentation updates, manufacturing, keeping track of which customers
have which versions.  A smallish operation is therefore under some
pressure to amortize the cost of the release process over a
substantial chunk of development work, if it wants to get any real
work done...



0,,

References:

Main Index | Thread Index