- From: Jeffrey Mogul <mogul@pa.dec.com>
- Date: Thu, 23 May 96 13:03:59 MDT
- To: "Gregory J. Woodhouse" <gjw@wnetc.com>
- Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
This makes a lot of sense. The EBNF section is included in a number
of places besides the HTTP drafts. I think we've definitely reched
the point where a separate RFC is justified. I also agree with you
that this sort ofr change is more appropriate for future versions
so HTTP 1.1 isn't needlessly delayed.
Just to reinforce that: I'm sure Jim Gettys has far too much work
to do on the HTTP/1.1 draft before it is submitted as a Proposed
Standard (a step that is supposed to occur Real Soon Now). Asking
him to also do the surgery required to split the EBNF into a separate
draft seems likely to rob him of any remaining sleep time he might
manage to find in the next few days.
At any rate, the length of the draft is not in itself the main
problem we need to solve. It is far more important to make the
protocol as simple as possible (but, as Einstein said, "and no
simpler") so that people will be able to implement it. And it
is important to make the specification simpler and clearer, so
that people will be able to understand it. Clarity is often
served by removing words, but it is also sometimes served by
adding a few words.
So don't get too hung up on the length of the document. Splitting out
the EBNF might be useful because it is useful for other protocols, but
it's not going to make the HTTP/1.1 specification any easier to read.
-Jeff
Received on Thursday, 23 May 1996 13:11:04 UTC