[whatwg] Re: RFC2119 terms in uppercase

Malcolm Rowe malcolm-what at farside.org.uk
Sun Aug 29 11:47:26 PDT 2004

Ian Hickson writes:
>> [RFC2119 terms - uppercase or not]
> What the purpose may have been is not spelled out in the spec, and the 
> spec certainly doesn't mention requiring that they be upper case, though. 
> Which is my point. :-) They don't _need_ to be uppercase, it's just, IMHO, 
> a typographic convention.

I'd actually argue that it's a semantic convention, since it allows the 
reader to distinguish between 'must' (the word) and 'MUST' (the RFC2119 
term), which can be quite useful in a sloppy language like English. But 
that's a shakey argument at best, especially since it's not formally 
described anywhere (and RFC2119 doesn't count, since, as you've mentioned,  
it's not explicit enough). 

> Many, many specs (especially W3C specs) don't use uppercase.

I can't say I've noticed. Maybe I don't read enough specs? :-)
CSS 2.1 was the one I first noticed, and that has an explicit explanation at 
the top. I thought it was rather clever. 

> It now says:
> [snip]
> Is that ok?

Fine by me - it's almost identical to the CSS21 wording. 


More information about the whatwg mailing list