On Monday, July 8, 2002, at 03:37 PM, Peter Schuller wrote:

> All that may be true; my point was the there are existing improvements
> out there that don't get accepted even if they are clearly superior to
> what's in the image. Surely Microsoft didn't have Windows 2000 laying
> around 8 years ago, or they would have released it (given the pethetic
> state of Windows back then compared to competetors). The situation I'm
> talking about is that their *IS* code out there - just not in the image.
> I'm not criticizing the amount of development for Squeak (I have no
> right to do so), I'm just point out that there seems to be certain
> "inhibitors" in the release process that results in less high-quality
> Squeak enhancements being in the image than is desirable.

All evidence to the contrary.  Over many years with this project, much, 
probably most, of the high-quality code put "out there" that makes sense 
to go into the image has been incorporated, so far as I have seen.  Yes, 
sometimes it may take an iteration or two, and sometimes high-quality 
code is not put out because of honest disagreements over whether the 
change is "clearly superior to what's in the image," or compatibility 
issues, testing issues and so forth.   Reasonable policy issues can also 
give rise to refusing to immediately incorporate content, such as 
"avoiding image bloat," which *IS* a serious problem.

In short, I think Peter's criticisms are widely overstated, and I have 
yet to see any proposed alternative that would be better than the status 
quo ante.  This is, in fact, a tired argument, based upon little, if 
any, empirical evidence that a problem exists, and no real alternative 
is presented with a likelihood to "fix" the "problem" to a state better 
than any we have today.