Yes, after reading more of the posts I realized that ebase is an existing and heavily used piece of work. Getting these guys on board would be exciting.
I agree with your assessment that 4.x provides a bunch of the framework but has a long ways to go. I do think the notion of packages of portlets can go a long ways towards lowering the curve necessary for folks to deploy something useful in any given "vertical app" space. I have some questions about the scalability of this approach, even with heavy portlet caching, but let's face it, if a non-profit has 1,000,000 users then they can afford a customized, more
efficient approach (if each user is worth $0.25 to them, that's $250,000 available for customization).
It's the quick-and-dirty, let's get something up cool and quickly that
will serve my 10,000 member organization's needs package that needs to be simple, and deployable quickly with a minimal learning curve. These are the organizations that have little money available.
I think one message that could be brought to the NOSI folks is that OpenACS 4.x will, by end of 2001, do a much better job of supporting more-or-less out of the box use (the portlets required for ACES include some very nice calendaring and similar things) while NOT SACRIFICING THE FLEXIBLE TOOLKIT APPROACH THAT'S MOST USEFUL FOR HIGH-END CUSTOMIZATION.
This blend will make the toolkit useful in a wide variety of contexts.
We could certainly use some of the folks, though some (Michelle, for instance) seem more interested in rolling their own world for the sake of doing so, others have their pet solution based on non-objective criteria (Zope is elegant but only gets you 10% of the way there - hmmm, I wonder what it might look like after it gets 60% of the way there?).
Also, when done, OpenACS 4 + bits and pieces of ACES/.LRN will have moved the kit considerably further than 60% of the way there, IMO.
It might also be worth stressing that portions the OpenACS 4 world will be moving in the same general direction that NOSI is discussing, with or without them. It only makes sense to pool resources and, well, we do have a significant headstart in terms of base platform stuff.
Meanwhile, I've been playing with 4.x somewhat and am already generating a list of loose ends and *bleeping aD!*-isms that make me realize we need to have a more rigorous definition of "product" than that used by aD (in the past only, not the present, I hope).
Since you guys are already rolling this site out and the openacs.org site out using OpenACS 4.x, I hope you're making your own list of demerits to enter into the SDM!
BTW - is the NOSI site OpenACS 4 PG or Oracle? It looks good!
