Date: Mon, 02 Jul 2001 20:02:49 +0900
From: JINMEI Tatuya / =?ISO-2022-JP?B?GyRCP0BMQEMjOkgbKEI=?=
<[EMAIL PROTECTED]>
Message-ID: <[EMAIL PROTECTED]>
| If we take the second approach, it might be possible to get a
| consensus in a relatively short period. So, it would be worth trying
| to keep the whole spec in a single document.
This is not the right way, though I found myself tempted to agree with it
for a while. As long as the read interface is set, so any random application
can deal (if it wants to) with anything that is handed to it over the net,
we could perhaps defer on maying a "standard" on how to send arbitrary
sequences for some time - allow various implementors to dream up their own
ways to handle this, and try them, and just wait and see which of those is
best liked by application writers. Since relatively few apps need the
flexibility, the costs in having to deal with multiple different methods
in the shortish term shouldn't be too great. Or that's the way I was
thinking.
But I always knew that wasn't really a good approach, and Francis' message
convinced me that it wouldn't be the right thing to do.
So, I'd prefer to complete & document the receive case, and just leave the
send case for a later doc to define. That is, not have any method at all
defined for inserting headers in the new doc for now (implementors can keep
implementing the current spec).
kre
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
