[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

discussion notes, and more list members

so, first off, welcome Andrew Wilson (Andrew.Wilson@cm.cf.ac.uk) and
Gregg Vesonder (gtv@research.att.com) to the mcp-dev mailing list.
you guys haven't missed much.  if you want out of the list (i know i
didn't consult you, net, before adding you, filfre to ask out) let me

next, in response to comments on the list, i've updated the spec at
http://www.ccs.neu.edu/home/dave/mcp2.html to reflect the suggested
changes to cord-ID generation.  i've decided to go the middle road
(the one most likely to annoy EVERYONE): i use "I" to mean "server"
and "R" to mean "client".  why?  well, the only useful definition of
either "initiator" and "receiver" or "client" and "server" we have is
based on who sent the first "#$#mcp" message.  that person is, in a
sense, the initiator of the session.  feel free to tear my head off,
but it doesn't really matter.  as long as the two letters are
different, the IDs will be unique.  i also fixed a few more random

i still haven't fixed up the bit about messages; erik's right, the
definition should include multilines (erik, maybe you should forward
your message to networker and gregg, just so's they know what i'm
talking about).

as for the local-editing stuff, i'm all for it.  hooray.  woo-woo.
we need to start generating proposed protocol docs, i think, as well
as a statement about namespace reservation.  i'll start on the
local-editing one RSN, and maybe come up with something about how we
could allocate namespace.