Re: ZIP-based packages and URI references into them ODF proposal

On Tue, 30 Dec 2008, Julian Reschke wrote:
> 
> Does the spec specify how to parse or serialize a future element called 
> "foobar"?

Yes.

(The parsing algorithm might change in a future version, of course; that's 
the problem with a having a syntax that isn't self-descriptive, like XML, 
or forwards-compatible, like CSS. There's not much we can do about that.)


> > > That's a very verbose way to state "must ignore unknown values".
> > 
> > It's a precise way of saying it, that leaves nothing ambiguous. That's 
> > the whole point.
> 
> I think the same precision can be reached with less verbosity.

Oh, well, I'm sure it can be said more tersely, sure. (Just out of 
interest, how would you phrase it?)


> > > So, if XML 1.0 *did* say that, how would you then introduce a new 
> > > value? Older recipients would ignore it, after all.
> > 
> > If older recipients ignore the value, then we're golden. Just 
> > introduce the new value, and then you have predictable down-level 
> > behavior and predictable up-level behavior.
> 
> Well, if it's ok for consumers to ignore a new value, then this is the 
> right approach. The problematic case is when that fallback behaviour is 
> not sufficient.

Yes. That's where language design comes into play. But that is a mere 
detail of design compared to the broader issue, that specifications should 
define handling for all inputs and all conformance classes in detail and 
unambiguously.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Tuesday, 30 December 2008 14:05:58 UTC