On 11/29/09 6:47 AM, Mickaël Rémond wrote:
> Hello,
> 
> Le 29 nov. 2009 à 01:53, ThomasWrobel a écrit :
> 
>> Read only access maybe could be accessible without a true login for
>>  previewing wave data (I feel public wave data could one day be
>> quite a resource, just as wikis are today). For any user-specific
>> stuff I don't think a login is a problem. After all, isnt this all
>> analogous to Pop3/Imap? I don't think people will have much problem
>> using their password with clients.
> 
> Twitter shows that client ecosystem can be diverse and play various
> roles. I expect that many clients will be developed for wave, some
> with very specific purpose (notification of wave update, wave reader,
> etc). Some will be web based, some will be desktop base. If your wave
> account is only link to wave it might be acceptable to share it. If
> your wave account however is linked to a larger system it might be a
> problem. Sharing your Google login and password with a web wave
> client for example can grant access to many service and not only wave
> (wave, email, adwords, etc).

Agreed. There will be many ways to access the same data + service.

>> Incidental, I got nothing against XMPP one way or the other. But
>> the Pygowave guys now have their Json (/STOMP) c/s protocol up and 
>> running. They even have a standalone client almost ready. So for
>> people wanting to dive in with their own clients, that seems the
>> place to go.
> 
> We have also written clients using XMPP, talking to our own wave
> server and I would feel XMPP is the way to go because it is extremely
> similar to the federation protocol. XMPP client can thus process what
> is coming from the client and in the same way, without server side
> transformation what is coming through federated wave server. It makes
> sense to me.

Makes sense to me, too. Now we just need to get to work. :)

Peter

-- 
Peter Saint-Andre
https://stpeter.im/



Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Reply via email to