[mythtv] MythSOAP Expressions Of Interest

[gettler at acm.org]

> > You seem dead against this, so I won't push it.
> 
> Even if that were the case I'd like to see this continue. I've gone down
> this road a few times already (usually when someone would add a new field to
> the protocol that mythweb didn't care about, but mythweb would still need to
> be updated). So far I've always given up (I consider the limitation on new
> dependencies a large time waster plus I always lump in my secondary goal of
> doing the socket stuff differently).
> 
> I wouldn't recommend going into the dark and coming back with a new
> protocol. It would be really nice if it were planned out beforehand.
> 
> -- 
> Anduin Withers

Yes, planning out a new protocol would be a very wise idea.  Making choices
that raise the memory, processing, or network bandwidth over what they are
now may lock out low end hardware which would otherwise make for a great
frontend.

I've been running the mvpmc project (http://mvpmc.sourceforge.net/) which,
among other things, acts as a very nice mythtv frontend on the Hauppauge
mediamvp.  Currently the code to deal with the myth protocol in mvpmc takes
up less than 50kb.  Obvously I would expect that to be larger if the current
protocol supported more features (ie, scheduling, commercial skip, etc).
But with 16mb total, there isn't much room to spare.

Anyway, it would be great to see a protocol that could provide everything
a remote frontend would need.  But it would be a shame if embedded systems
like the mediamvp were unable to use it.

Jon