[Standards-JIG] File sharing JEP

Michal Vaner (Vorner) michal.vaner at kdemail.net
Sun Apr 9 10:05:01 UTC 2006

Dne neděle 09 duben 2006 11:16 Jose Carlos Garcia Sogo napsal(a):
> El sáb, 08-04-2006 a las 09:34 +0200, Michal Vaner (Vorner) escribió:
> [...]
> > > I don't think a JEP-65 transport for Jingle would buy us very much.  It
> > > might help as a technical/political transition to getting rid of SI,
> > > but what about user-visible improvements?  I guess the question is:
> > > what can a Jingle negotiation bring to File Transfer that SI cannot?
> >
> > I think there is one big problem with file transfers and Jingle - Jingle
> > is UDP. You need to have your own* way to take care of lost packets, flow
> > control and everything. You are unable to optimize the size of one packet
> > to the maximum nonfragmenting size on the whole way, since you can not
> > operate with most packet flags as a normal user. TCP does this all, but
> > the problem is, it can not be STUNed and must be relayed trough the
> > proxy. However, I think it is a bad approach to the situation using UDP
> > for file transfers, we should instead want IPv6 and direct connection
> > access for everyone.
>   Sorry, but I would like to know where in JEP-0166 (jingle) says it is
> UDP. At all. More over, Jingle is "a framework for initiating and
> managing peer-to-peer session", which can be used for VPN (as other
> people suggested), VNC, file trasfers, ...

I want to see you using STUN for TCP, as you suggest ;-)

>   Perhaps we will need to build a TCP transport, but Jingle does only
> stablishes which parameters both ends need for making a connection from
> one to the other.


Ostatně soudím, že uzavřené protokoly a formáty by měly být zničeny, stejně 
jako Kartágo.

Michal Vaner (Vorner)

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 191 bytes
Desc: not available
URL: <http://mail.jabber.org/pipermail/standards/attachments/20060409/bc3eb789/attachment.sig>

More information about the Standards mailing list