[standards-jig] NEW JEP: Inband Bytestream (JEP-0047)
ben at blahr.com
Mon Sep 30 06:10:02 CDT 2002
On Mon, 30 Sep 2002, Richard Dobson wrote:
> > I want to remind the discussion of the karma settings, which can be a very
> > annoying issue for transmitting data as "inband" data.
> Ah yes almost forgot about that, thats another spanner in the works, this
> probably needs to be done over a different connection to the server then, to
> a port specially for it that has higher karma settings, that will also help
> with the problem of people sending inband when the server admin does not
> want to allow it because they can simply not supply this port and set the
> karma appropriately so the low levels of messages and presence etc will get
> though but as soon as someone tries to do a inband transfer it will bring it
> to a halt.
Don't you see though? This is exactly the reason the JEP is written like
this. The whole point of the JEP is to address inband bytestreams with
small payloads. In fact, I would argue that the previous discussion about
encoding mechanisms is a waste of time. If you want to send something
inband, then you will have to deal with the fact that it will get larger
during trnasport, as a result of BASE64-encoding, and that karma will slow
make it more painful. The point of the JEP is not the end-all, be-all file
transfer JEP, but to address sending small icons and such inbound, where
going through the trouble of setting up another socket isn't worth it.
In this situation, its more cost effective to just send the file inbound.
However, in most situations, its more cost effective to send the file OOB,
since you won't have things like karma limiting your speed.
More information about the Standards-JIG