With the advent of the Stablr concept, I've gotten a huge amount of interest from the general community. People have been throwing suggestions, ideas, and their talent my way.
The good part of this is, I've fleshed out so many important aspects about Stablr its not funny. The unfortunate part, is I didn't have the chance to write them down. Please, if you want to become involved - send me an email, leave a comment, etc. Contact me somewhere, and I'll make sure to write it down.
Now, for some details.
I got some very good insight into how the system should be laid out, and had several good conversations about frameworks, and why I shouldn't use one. We also discussed MySQL data access engines, and which one would be fastest, as well as caching, and more. It was an incredibly productive weekend, and Matthew Turland's wife was correct. This was all fleshed out over at a bar. (Homeless Shoes was it?)
Anyway. I purchased http://stablr.net, and have written up a 5-page document on implementation details. I've also started working on the base code.
Again, anyone want to become involved :D ?
That link to the 5 page document is 404. Also, what will be the final license of the codebase - will it be open source?
You might be interested in this post. Some nice architecture in it...
http://blog.fav.or.it/2008/05/26/fixing-twitter/
My comment would be not to replicate twitters mistakes by building this system as a CMS.
Also why re invent the wheel in terms of message delivery. I would look at open source MTAs and hooking into their ability to scale as a delivery system.
Or even look at twitters open source ruby stuff.
The 5-page document link is not working..