- From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
- Date: Thu, 15 Apr 2004 10:39:34 +0200
- To: "Brian Kelly" <B.Kelly@ukoln.ac.uk>, "Alex Rousskov" <rousskov@measurement-factory.com>
- Cc: "Rajasekaran Deepak" <deepakr@students.iiit.net>, <www-qa@w3.org>
> I also agree with Olivier's points. +1 Also about "must" ... in the standards world a different word "MUST" has a well understood and common meaning as defined by RFC 2119 http://www.ietf.org/rfc/rfc2119.txt [[ MUST This word, or the terms "REQUIRED" or "SHALL", mean that the definition is an absolute requirement of the specification. ]] also section 6: [[ Imperatives of the type defined in this memo must be used with care and sparingly. In particular, they MUST only be used where it is actually required for interoperation or to limit behavior which has potential for causing harm (e.g., limiting retransmisssions) For example, they must not be used to try to impose a particular method on implementors where the method is not required for interoperability. ]] Since "MUST" and "must" are pronounced the same and easily confused, it is IMO generally best to avoid the word "must" when possible. This is particularly important where the phrasing sounds like a standard document. I also feel that section 6 is helpful for many imperatives even outside the narrow scope of that section. For example, I have avoided using an imperative in the paragraph above. Jeremy
Received on Thursday, 15 April 2004 04:41:56 UTC