Here are some not-well-thought-out opinions. They are really a reply to Liz's post, but they are long enough that I am breaking convention and throwing them to the top level:
I still like the blogging center of excellence idea, but I also think it may be slightly premature. Such a (fundable) center, it seems to me, would be a natural spin-off of any sizable collaborative research venture.
I think the combination of your points would make for an interesting proposal. I'll restate these with my own spin (and a focus on your third point, I suppose):
1. Is there a clear way to measure the effects of blogging within a research field?
There is an inherent difficulty in establishing such metrics, as we are looking at a swiftly moving target. Nonetheless, we should be able to establish, in more than an anecdotal way, that blogging leads to significant progress within scientific fields of study.
I think we can borrow pretty heavily here from a tradition in measuring R&D capacity, especially at the national level, and extend these measures to the blogosphere.
2. What (formal or systemic) factors lead to the highest degree of (e.g.): (a) community formation, (b) production of new ideas, (c) whatever metrics we establish in #1>
That is, are some blogs better than others. Once we know what is good, we can start to talk about some--perhaps abstract--means to achieve better collaboration.
The literature on "communities of practice" would provide some basis for establishing a model here, though there is a lot of chaff to get through there (IMHO).
3. How can the diffusion of such technologies of collaboration be encouraged?
Liz, you've suggested Turkey dinners, but I don't know that this is a scalable solution :). The question is: once we know that these things are good, and how to make them better, how do we sell them to a scientific community. I'm thinking about this in terms of my graduate students and fellow faculty. How do I sell them on the idea that taking a few minutes (who am I kidding?) of their day out to read and write in blogs is a productive pursuit?
One of the ways I do this is by establishing empirically (in #1 and #2) that there are productivity gains--I prefer "creativity gains"--to be had by getting on the blog wagon.
Another, as you suggest, is to look for the reason folks pick up blogging (uses and gratifications) in law, journalism, etc., and how this might be translated into scientific endevors.
My guess is that we may be able to borrow something from the push on electronic portfolios in the education world, as well.
In any case, I think it would be beneficial to keep the idea of what concrete results we can produce at the forefront. Faculty that made visits last year also suggested ITR was open to social scientific work--more open than some other obvious contenders in NSF. But I think it's a very good idea to frame our research in a strongly empirical--and dare I say quantitative--way.
Here is one final suggestion. Perhaps not now, but at some point we might want to invite one of the program officers to check out the blog, and perhaps offer comments. The former, rather than the latter, seems more plausible. As Seb has already mentioned, this approach to public grantwriting is not only unusual, but something of a proof-of-concept in terms of our proposal itself: i.e., collaboration via blogs leads to significant research.