[isf-wifidog] Roadmap?

Mina Naguib webmaster at topfx.com
Mer 18 Mai 09:49:16 EDT 2005


Hello

Mina from the IleSansFil camp here.

I've just had a discussion about your thoughts with Philippe, another 
IleSansFil hacker.

As horrible as this sounds, we have no roadmap.  Even worse, I don't 
think we can have a roadmap.

The bulk of WiFiDog (client and auth) has been developed.  We've got a 
small to-do list (throttling/etc), but for the most part, the hacking 
is done.

What we're doing now is building the community instead of building the 
software.  This does entail changing the auth server quite a bit, but 
it's reflective of the community needs as they arise (profile, now 
instant messenger, now flickr, now.... etc.).

We *could* sit down and make a roadmap of the features that are coming, 
however 1. It will not be complete 2. It will be highly 
community-centric and 3. It may be highly irrelevant for other groups

Please keep in mind that IleSansFil has always been about the community 
and that's where our auth server/portal is headed (starting to show 
around the waistline too ;)  If a wireless group is heading in a 
similar direction, bulls-eye. If not, well, that's where we're at right 
now:)

The following are my personal opinions and Philippe's.  They are not 
reflective of the other hackers or IleSansFil as a whole:

1. We should re-write the client <--> auth protocol to be cleaner and 
more extensible.  This entails some modifications to both the client 
and server.
2. We should continue releasing the client, as well as a very clear 
protocol documentation
3. Along with the above, we should release a very light 
proof-of-concept auth server that demonstrates the use of the protocol 
from the server perspective (maybe in a couple of different 
languages/CGI frameworks if we're nice).  We should release it with a 
pluggable modules architecture and maybe a couple of simple modules.
4. We will continue releasing our existing auth server with all it's 
community-oriented stuff, in case a group has similar goals.  Hopefully 
we could modularize it a bit to make it lighter.

If a group has similar goals they'll work with #4, otherwise they'll 
use #3 as a platform for building the auth server so they can "have it 
their way".

Frankly I don't see the need for "the one, the only" definitive auth 
server to be used with the wifidog client.  As groups develop their 
needs, features could be shared to other groups as needed (similar to 
what the *BSD camps do).  The client is dumb and the protocol is 
simple.  The cleverness is all in the server.

I'd like to hear comments on the above, especially from IleSansFil 
volunteers, but also from other users of WiFiDog.



Plus d'informations sur la liste de diffusion WiFiDog