Hello all,

I've been trying to follow this thread, and I find myself agreeing (and
disagreeing) with people on both sides of it (might be better to say all
sides of it<g>).  A few random thoughts:

(1) Those concerned about the dynamic nature of Squeak might do well to
tune out of the update streams, or at least stay clear of the alpha
versions.  Either way, add lots of unit tests.  In that regard, I really
need to build (or better yet find) a filtered test runner that I like. 
Ignoring a few problems, the SUnitBrowser in Dolphin is really nice;
Squeak's improved test runner is close, but no cookie, IMHO.

(2) Somebody was kind enough to send me some code for this; I have it,
and _will_ try it - honest.  Been busy; really busy.  So what is the
topic?  Partly because Dolphin is a commercial product that appears as
releases (with some updates along the way), and partly to avoid some
problems listed in (1), I like  to grab a new release and move my code
into it, leaving my previous image in tact.  It helps a lot to have
tools that decide what is loaded, and how to get it loaded in the next
image.  Those with access to Dolphin might want to look at my Migrate
package.  It consists of a "current image analyzer" that writes a script
to build the "next image", and a loader that is called from the script,
which is simply filed into the new/clean image to build it..  As I said,
I have the initial pieces to do something like this via SM, but I would
not mind at all if somebody beat me to it as part of the ongoing
efforts.

(3) Release schedules.  This one might sting a little.  If it does,
please consider it is meant to be constructive.  Rather than fix release
dates and point out that "nobody cared enough to...", I think it would
be a far better use of human nature to post what/who is delaying the
release because their tests don't pass [*].  Being told you're a lazy
jerk will turn people off; but, no one will want their name on the
showstoppers' list, and they will work to get off of it.  At that's my
prediction.  One could further build in some notion of critical vs.
optional packages, with a max delay granted for inclusion of an optional
package.

Sincerely,

Bill


[*]  I am starting to sound like an XP advocate; I am not.  However, I
see value in several pillars of XP, and automated testing is high on the
list.  I tend to let development drive tests rather than the other way
around, except when (it seems to be rare), I indeed know, and can
reasonably test, the expected result.



Wilhelm K. Schwab, Ph.D.
University of Florida
Department of Anesthesiology
PO Box 100254
Gainesville, FL 32610-0254

Email: bills at anest4.anest.ufl.edu
Tel: (352) 846-1285
FAX: (352) 392-7029