Rapid deployment of Votorola

[Michael Allan]

Philippe Schelté wrote:
> Sorry to not comment this, but I probably miss some elemnts to be sure
> that I understand well the process. I'm sure I will begin to
> contribute in term of ideas when I will test or code the first
> prototypes.

We probably can't deploy it as I suggested.  I tried it on paper, and
it came out a train wreck:

  http://t.zelea.com/wiki/User:Mike-ZeleaCom/Architecture

It's funny in a way, because now I think the solution is to do (in a
sense) what Bitbucket and github are doing: take a radically
distributed design and *not* distribute it.  Not distribute it *yet*,
because first we need:

  * Control over MediaWiki configuration and extensions
  * Tight communication lines for rapid turn-around
  * Critical mass of developers, admins, and users

So we should probably deploy a single central pollwiki/streetwiki.  If
it gets too crowded in there at some point, *then* we start
demonstrating our distributed chops.  Meantime, we focus on
prototyping the common facilities on which it all depends - free-range
voting and residential registration - and on developing the separate
voting tools.

I bounced the idea off Thomas.  (He carries this stuff around in his
head.)  It's hard for anyone else to comment without a proper overview
doc.  The design is scattered here:

  http://t.zelea.com/wiki/User:ThomasvonderElbe_GmxDe/Vote_mirroring
  http://t.zelea.com/wiki/User:Mike-ZeleaCom/p/edor
  http://metagovernment.org/wiki/Streetwiki

Plus these hacks to speed things up:

  http://t.zelea.com/wiki/User:Mike-ZeleaCom/Opening_an_architecture#Rapid_prototyping

I'll try pulling it all together, soon (2nd attempt).  Right now, I'm
cobbling together a faster server.  (It's something else we need,
nobody will disagree!)

-- 
Michael Allan

Toronto, +1 647-436-4521
http://zelea.com/