On Mon, 25 Oct 1999, Klaus Moeller wrote:
<snip>
> > log. Or sent down the wire, for that part. I still say negotiation
^^^^^^^^^^^
> > is the only way to go, with the option for the sending party to
^^^^^^
> > blatantly ignore it. The recieving end could do the stripping of
> > fields locally, if the sending one does indeed ignore it.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> I strongly disagree here. In terms of scalability I dont want to strip
> out lots of unnecessary/unwanted information in a central place. I
<snip>
>
> Klaus
I can certainly agree that -desired- behaviour is to negotiate away the
data so it is not sent in the first place, but this might not always be
entirely feasible for embedded applications. I believe one of the main
ideas behind this effort was to make it implementable on almost anything -
if it's a -requirement- to support negotiation of contents, that'd make
life hard on the kind of people who do communications suites in a couple
of K worth of RAM. If it's optional, the beefier stuff can support it, no
sweat, the less-so stuff can just ignore it.
Kriss
--- .... --..-- -.-- --- ..- .-. . .- -.. -- --- .-. ... . --..-- . .... ..--..
Kriss Andsten <[EMAIL PROTECTED]> telnet slartibartfast.vogon.se 4243
