Hello Ketty, 2006/10/15, ketty . <[EMAIL PROTECTED]>:
This modification would be incompatible with current version am i right?
Yep.
But as for protocol additions that don't brake anything, are they a big deal too?
No, one can easily add new methods (if care is taken to use a different method identification number).
Should we try to gather as many protocol changes as we can when we are at it? Or would that be stupid?
I don't know: * Should we create a new protocol version number, copy all the old methods and add the new ones? (or) * Should we add a new protocol version number, just adding new methods and modified one? If I don't know the answer, I can at least list the objectives: * keep the protocol clean, removing unecessary methods when no longer used (think that the protocol could be published as an IETF RFC in the future); * don't try to keep at all cost protocol upward compatibility. I consider that until we reach 1.0, we can break the protocol whenever we want. That's said, as the software is already used a little (we have running servers), one should take care to make such changes carefully. ;-)
Things i think is missing from the protocol: * Deletion of tags / questions / responses. * Delegation.
Don't say that! Frédéric is asking me to add that for years, day after day! :-D More seriously, the issue in delegation is not much the protocol than the way to handle delegation on the server. If you are interested, I have given in the past an example that shows the issues in delegation. See the link in the wiki: http://demexp.org/en/doku.php?id=delegation Moreover, Jérémy DUBREIL (he should be on this list) is working on security issue and has some ideas on delegation. I hope he could explain to you his ideas. Best wishes, d. _______________________________________________ Demexp-dev mailing list Demexp-dev@nongnu.org http://lists.nongnu.org/mailman/listinfo/demexp-dev
