Re: [Social-discuss] On the architecture of a GNU Social node

[Ted Smith]

On Sat, 2010-04-24 at 12:09 -0400, Sean Corbett wrote:
> > I just posted this on libreplanet - it outlines in a more specific way
> > how I think GNU Social nodes should be built.
> >
> > <http://groups.fsf.org/wiki/User:Teddks/Social>
> >    
> I definitely agree with the high level design presented in this 
> document. This is basically what Ian and I envisioned when we've sat 
> around, shooting the shit and thinking about how we'd implement GNU 
> Social. I especially like the idea of UI transports; the use of core 
> transports and UI transports solve many of the disagreements on 
> implementation that constitute much of the mailing list's activity over 
> the past few months. Thanks for writing this document out; it's good to 
> see a clear, solid idea of how the node should be designed. A lack of a 
> document like this (and a starting-off point from which to get some code 
> written) was why we had the meeting last Wednesday in the first place.
> 
You're quite welcome. :-)

> Speaking of the meeting, I'm sorry there wasn't nearly enough "coverage" 
> as there probably should have been. I'm also sorry if we gave the 
> impression of trying to form some elitist group of developers near 
> Boston who would make decisions for the whole community. All I really 
> wanted was some clarity as to where to start with the project, because I 
> saw the list starting to stagnate and I felt like I should do 
> *something.* I agree that in the future, meetings should take place 
> through IRC, especially given the amount of hackers that aren't even in 
> the U.S.
> 

I might have been too harsh in my calling-out of the closed nature of
that meeting; it hit a nerve I've been feeling quite acutely since
reading *The Tyranny of Structurelessness*
(<http://flag.blackened.net/revolt/hist_texts/structurelessness.html>),
which describes in detail the danger of this sort of thing. The great
reaction from those involved has mitigated any fears I had, though. :-)

signature.asc
Description: This is a digitally signed message part