Non-Validating XML Parsers: Requirements

Michael Kay M.H.Kay at
Tue Aug 4 11:26:50 BST 1998

>*sigh*  I do wish people wouldn't review things without
>them.  I happen to agree with you about MAY NOT, but that's
>what RFC 2119 says.  The RFC is about 600 words long, BTW,
>here's a link: .

Thanks for the reference. I've read it now. I'm relieved to
discover it does not recommend or assign a meaning to the
phrase "MAY NOT".

Wherever "MAY NOT" appears in a (so-called) spec, it either
means "MUST NOT" or it means "MAY OR MAY NOT", which is a
synonym for "MAY", and which, as I remarked earlier, is
formally equivalent to omitting the sentence.

>> I don't much like "may" either. Everything is permitted
>> unless the specification prohibits it, a sentence whose
>> verb is "may" therefore says nothing.
>*Everything*?  So if a specification for a C compiler
>*say* that compiling a strictly conforming program does
>make demons fly out of your nose, then the compiler is
>to do that?

Absolutely. It might not succeed in the market, but it would
conform to the spec.

(As did an early Algol68 compiler I once used whose only
error message was "<filename> is not a program". Which, come
to think of it, is not that far removed from the behaviour
of some XML parsers I have used...)

Mike Kay

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as:
To (un)subscribe, mailto:majordomo at the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at

More information about the Xml-dev mailing list