Dog fooding SBlog

My SBlog participation so far has been a very nice distributed team project. Monticello and the development model around it really came through in terms of us being able to pick up one anothers code and add stuff with minimal spurious effort. I think we did find a bug in moving variables up/down the hierarchy, and I think we could go farther in the distributed model, abandoning even an "official" shared ftp space. I think the distributed model is really viable, and we should pursue it to the utmost. Software should model the core work process of its users, and the core of open source development is distributed, asynch. So we should strive to eliminate all artificial sources of synchronization/centralization requirements from the software.

I gained a lot from the IRC interaction - having at least occaisonal synchronous communications is really useful for speeding things up. I felt that better presence information would have been really useful - people leave their IRC on and go, and then the participant list "lies". This could be fixed on the client side, I guess.

I learned a bit about how to use XML-RPC from Squeak (a piece of cake, with the package on SM), and this should really be useful for other projects - for many projects, this should be the only kind of client/server communication needed (assuming we're writing both sides). And its language agnostic, which is always good...