> Perhaps you could clearly and briefly summarize the point
> that you're trying to make
Oh....the other thing. Business models.
Forget consumers for a minute -- just think IT depts and the like.
They need, in my opinion, at least lightly manned rel eng pipelines of
their very own. The product shouldn't be a boxed set or even a
subscription service to a package updator -- the product should be
somewhere upstream of that with some other services mixed in to make
it work smoothly. There should be a lot more trading, review, and
consideration of change sets --- you don't just buy what XYZZY corp
tells you is the "latest version of Linux" -- you buy a system that
gives you tangible proof that various invariants are preserved as you
change the software installed at your site.
Now, you can mix consumers back into this, but, in my opinion, they
need some proxies who are their IT depts (ISPs, for example).
Well, a cornerstone of a rel eng pipeline is the revision control
system, and a distributed revision control system can link customer
sites to their competing providers better than a centralized system.
Plus, it can let them mix and match from various suppliers.
Finally, rather than duplicating CVS, to make it really work well, we
have to "trick" programmers into producing clean change sets and the
best way to do that is with a different user interface (not a process
manual). By "trick", I mean, "make doing the right thing the easiest
course of action".
-t
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Aug 13 04:21:29 2002