If it's just about "the reconsumption of the eclipse stack
for tool vendors" then what do we need an update site for? Tool vendors can deal
with humongous zip files. Tool vendors can deal with weird version dependencies.
I thought the main idea of an update site was to improve the experience for end
users.
So how about option
3:
- Have one update site, make the integration good, get
good reviews, *and* do synchronized updates ala IDEA and NetBeans. (I know it's
more than an IDE but having it be a super IDE gets it in the door for the other
things).
Once the update site is working you'll have about 3
months for integration polish, with community
feedback.
Even though I clearly understand
the goal of callisto, making all the things available for download from one
update site simply augment the belief that callisto is about integration. I
think this is because the term"simultaneous release" does not capture the
motivation behind callisto (which I believe is to ease the reconsumption of
the eclipse stack for tool vendors).
Yes and yes.
So there is two
possibilities:
- don't make one
update site and let people hunt down the required jars
- go ahead with one update site, fix the obvious
integration problem and get beaten by the reviews for poor integration so we
can do better the next time.
Even though it will hurt, I'm in favor of the second option as
otherwise we will never improve cross project integration and
communication.
Me too. (or "me two"
:-)