- From: Adams, Glenn <gadams@spyglass.com>
- Date: Mon, 9 Nov 1998 18:35:01 GMT
- To: 'Ross Patterson' <ROSSP@ss1.reston.vmd.sterling.com>, http-wg@hplb.hpl.hp.com
- Cc: gadams@rafiki.spyglass.com
-----Original Message----- From: Ross Patterson [mailto:ROSSP@ss1.reston.vmd.sterling.com] Sent: Monday, November 09, 1998 12:07 PM To: http-wg@hplb.hpl.hp.com Cc: gadams@rafiki.spyglass.com Subject: Re: Comments (Part 2) on HTTP I-D Rev 05 "Adams, Glenn" <gadams@spyglass.com> writes: >74. Section 13.5.1, pg. 84, 1st para, 1st bullet, has "End-to-end >headers which MUST be ...". The use of MUST and SHOULD keywords in >relative clauses is problematic and should be avoided since it does not >state a requirement per se. I don't understand your meaning of "relative clauses", however this case in particular does state a requirement, that intermediaries forward end-to-end headers to the ultimate recipient. This requirement isn't stated anywhere else in the document, and must be retained. I use "relative clause" in the sense of English syntax; i.e., a subordinate clause introduced by a relative pronoun. Since such clauses are used to comment on or restrict the scope of interpretation of the preceding noun or noun phrase, it is not appropriate to embed an imperative in such a clause. It should be restated outside of a relative clause; e.g., "End-to-end headers MUST be ...".
Received on Monday, 16 November 1998 09:47:37 UTC