streetwikibot and streetwikis setup standart

[Michael Allan]

> of course, I was thinking of giving arguments to the bot for the namespaces
> to backup instead of setting it in the config file :
> 
> ./streetwikibot.pl 10,102
> 
> would backup the Template and Property namespaces.

Arguments sound like a good idea:

 * No setup required (if no config file needed)

 * You can call it repeatedly (e.g. from another script), with
   different arguments for each call

> >  * Can we configure it to backup only the latest revision of each
> >    page, and not the entire history?
> 
> no problem :)  I can add a limit in the config file.

There's an option on the Special:Export page for that:
http://okidoke.referata.com/wiki/Special:Export

I was surprised that export isn't included here:
http://search.cpan.org/~dcollins/MediaWiki-Bot-2.3.1/lib/MediaWiki/Bot.pm
(I wish I looked like that guy.  That's how a programmer should
 look. ;^)

> > At some point, I guess the wikis will be too big to backup in a single
> > HTTP fetch.  Then we'll need some kind of caching to avoid
> > re-exporting pages that haven't changed.  But we don't need that, yet.
> 
> This shouldn't be difficult to implement, I can probably get the last
> modified date of the pages, store it in a file and compare before fetching
> the pages.

Backup one page at a time, each in a separate file, like
NAMESPACE/PAGE?  Then I guess you set the modtime of the file to the
revision time.  (But we don't need this, yet.  I wouldn't code it.)

> lets go for the version 1.1 !

Command-line arguments like "--namespace=NAMESPACE" and
"--revision=current|all", would make it nearly perfect. :^)
Here's examples of option handling:
http://zelea.com/project/votorola/b/build
http://zelea.com/system/host/obsidian/usr/local/bin/desk-launch

We're now thinking of using Referata for pollwikis, too.  So there
should be lots of work for bots like this.

Mike