<URL: http://bugs.freeciv.org/Ticket/Display.html?id=40592 >

> [sch...@debian.org - Wed Dec 10 22:50:45 2008]:
> On Wed, Dec 10, 2008 at 01:57:51AM -0800, Jason Dorje Short wrote:
> > GGZ provides all the same features as auth with multiple backends 
> > support.  I'd see improving the GGZ support as a better
> > alternative than working on auth.
> Thanks, I'll abandon my effort on the auth front.

I would rather that if you have the time and would not mind
having your work possibly superseded at some unknown later
time when ggz has matured, to submit your proposed auth
improvements. With a clever design, there is also no reason
why ggz cannot become just another "backend" for authenticating
users (at least I assume the ggz api would permit that).

Not that I do not agree with the premise of ggz, it is just that
having multiple auth backends would help in the unification of
the various freeciv "flavours". For example longturn freeciv
uses a postgres library (pqxx of all things...) so that the
server auth can be tied in with drupal modules used on the
longturn.org site. Warclient freeciv implements a full extended
glicko rating system, which is quite a bit more complex than
the rating system provided by ggz given the complexity of game
types possible in freeciv.

I think server operators should have a choice in how they can
authenticate users. Perhaps they just want fine grained control
over which users are allowed to connect (e.g. banhammer), and
the connection statistics. Or the ggz master servers are down
that day.

Anyway, my point is that assuming implementing multiple auth
backends would not turn the auth code into a sprawling
unmaintainable mess, since nobody is actively working on the
freeciv-ggz integration at the moment I would that you make
your contribution rather than not.


Freeciv-dev mailing list

Reply via email to