[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
practical matters (was Re: tired of waiting)
> Hell no. I still want to take tab out of the moo charset.
Okay, resolved, then.
> I'll have to stare at what we're doing and make sure our deployed
> mcp-1 implementations don't lose badly. I guess I dunno why YOU get
> to beat ME into compliance instead of vice versa....I suppose actually
> starting the process up again gets you some points.
Didn't min-version:/max-version: come out of the labspace/mitremoo summit back
in 1995?
Seriously, if we no longer have control over all the system out there that are
calling themselves MCP 2.0, then we should call this 3.0 and bag them. The
argument names for the versioning message are not the only difference between,
say, LabSpace's MCP dispatcher and the current spec.
Which is why I want to get something specified and public, so we don't end up
with another round of multiple conflicting implementations saying they
implement the same version of the same protocol. (Admittedly they say it in
different ways, which reduces the potential for collision, but then it reduces
it in the ugliest way possible.)
Which again brings up the idea of a reference implementation. If no one else
objects or (preferably) volunteers, I'll probably hack the LabSpace MCP
dispatcher into conformance, tracking changes so we can release it close to
simultaneously. I'll probably have to do this in my spare time, on a public
server, as I suspect trying to get this through AT&T release process is more
trouble than it's worth.
I may also (again, barring objections or volunteers) hack the module system
into shape for LambdaCore, if I can figure out a way to do it such that it can
spit itself out in script form suitable for either core.
I'm not volunteering to do anything about a client, though.
--Erik