[Standards-JIG] File sharing JEP
Michal Vaner (Vorner)
michal.vaner at kdemail.net
Fri Apr 7 20:42:21 UTC 2006
Dne pátek 07 duben 2006 18:34 Peter Saint-Andre napsal(a):
> /me bites the bullet
> IMHO the long-term solution is to do all out-of-band media exchange
> using Jingle. That would include file transfer / file sharing.
> My reasoning is as follows:
> 1. It's not good to have two ways to do similar things, since it makes
> coding clients harder. The Stream Initiation (SI) framework is quite
> similar in concept to Jingle, though IMHO not as flexible.
But we should have a way how to move from one to another, keepeng some
short-term backwards compatibility. Thus, session initialization with jingle
would mean the 'old' clients would not find a difference and we had time to
upgrade them by adding the jingle. Wouldn't a MUST on jingle and MAY on out
of band stream be enought?
I use file transfers on a day basis. However, if we change the protocol now,
in the fawour of a better one, we have two incompatible protocols in the
wild. It would mean clients will be either incompatible and not capable to
transfer files fith everyone, or implementing both, which is not even
slightly better than having only the old one.
> 2. We never figured out a good way to do anything but file transfer
> (such as voice and video) using SI. Or at least I never figured it out.
The SI is slow, because it is always relayed, therefore it was not OK for it.
> 3. I think Jingle gives us a more flexible approach than SI (voice,
> video, etc.), thus enabling clients to support only one framework for
> all media exchange.
> 4. It's possible we could define a Jingle transport method for SOCKS5
> bytestreams, thus re-using JEP-0065 (though I haven't worked out the
> details yet).
> 5. File transfer / file sharing over Jingle would be defined in such a
> way that it enables the transfer of multiple files, rather than just one
> file. Thus no need for something like JEP-0105, which never took off for
> a variety of reasons.
> 6. Jingle should include all the retry semantics that people want to add
> to JEP-0096 (if not, we need to define how that would work).
> Or so it seems to me right now.
Another problem I see is that jingle will be a lot harder to implement than
just the simple tcp connection to some IP address used with SI. I'm writing
my own client for fun, and I fear the time I will have to do the jingle
Ostatně soudím, že uzavřené protokoly a formáty by měly být zničeny, stejně
Michal Vaner (Vorner)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 191 bytes
Desc: not available
More information about the Standards