> that a lot of us are building the same components over and
> over again with almost zero portability.
That's exactly what I mean.
I think that a standard a la W3C doesn't make sense. There are too few common design
patterns to just standardize the schema, or handshake, or plugin API.
On the other hand the current situation is really unproductive. It is a lot of work
to do the basics right. If you are aiming to meet a narrow, targetted need it should
be possible without inventing a cosmology.
For example, I would really like to make the reputation management components of
WorldOS available to projects based on other infrastructures, but it is impossible.
That way I can get away from general infrastructure, which I don't have the resources
to win at, and focus on a narrow, clear niche. I have a consulting client in the
exact same position. He has a way cool thing and needs to be able to hook it up to
multiple P2P platforms.
I think that starting with requirements gathering in the way Dan Moniz has been doing
is a workable approach. There doesn't need to be One True Language, just something
better than the current tower of babel.