[SF][VM] SF for VMs: Phoenix out of ashes?

Peter Crowther Peter.Crowther at melandra.com
Thu Feb 8 11:29:58 UTC 2001


> From: Raab, Andreas [mailto:Andreas.Raab at disney.com]
[...]
> The reason why most people use VMs provided by
> certain maintainers is simply that those people have proven 
> over a long time that they are willing and able to do the
> constant maintenance effort.
[...]
> But these people you are putting your trust into are usually of a very
> particular kind. They know that people put their trust into 
> them and they
> want to provide the best service they can. So they put a lot 
> of effort into
> making sure everything runs smoothly and everything works across the
> variants of that OS they are working on. And they hate it, 
> they absolutely
> hate it, if there is some modification that doesn't fit into 
> their working
> style, that isn't quite compatible. Or at least I do.
[...]

Some of these people get pretty snowed under, unfortunately.  Ian, in
particular, no longer seems to have the time to maintain the UNIX VMs --- I
was surprised and heartened to see his posting about a new VM recently, but
it's the first we've heard from him for a long time.  Ian's someone else who
gets very tetchy about anyone else touching his code.  I know; I used to
work with him!  But it's difficult to resist the temptation when the code
appears to be gathering cobwebs!

Andreas, in an ideal world where the VM maintainers have unlimited time to
devote to the task, I would agree with you.  But time is limited for
everyone on this list, and it can get frustrating when the Mac VM or the
Windows VM or the UNIX VM supports some new feature but the maintainers of
the other VMs cannot put the effort in to integrate the features in a
sufficiently timely fashion for the crowd of enthusiastic test-pilots on
this list.

I'm quite willing to believe that SourceForge is the wrong solution to this
problem, but I think it's worth a degree of discussion as to how we could
create a win-win situation that satisfies the test-pilots and the
experienced VM maintainers.  And I think it's worth taking a good look at
Tim's build structures as well as Ian's and your own.  The end decision on
which to use or which features to use will have to be taken by fiat, because
structures designed by committee are too unwieldy; but I feel that the ideas
coming in from experienced developers who aren't among the existing VM Mafia
shouldn't be dismissed out of hand simply because they're outside the VM
Mafia.

I'll go and hide in a relatively flameproof bunker now...

		- Peter





More information about the Squeak-dev mailing list