Le Friday 25 October 2002 19:43, Paul Grosso a écrit:
> url() isn't a function. It is not listed in section 5.10.
Absolutely.
I'm wondering why it is not, BTW.
The whole xsl-fo specification lacks of orthogonality :
- things that look like a function (url(), rect(), etc) are not a function
while others are ;
- things that look like a reserved names (function names, for example) do not
behave like reserved names (you need an opening parenthesis somewhere after
for it to be considered a function) ;
- things that look like a minus operator do not behave like an operator (you
need a space for it to be definitely considered as an operator an not as an
hyphen) ;
- things that look like a data type (<percentage>, <angle>) are in fact just
an initialization value while others are real data types (<number>, <name>) ;
- etc...
> I claim the expression 'url(xxx)'--when given as the value
> of an XSL FO property--evaluates, according to the XSL FO
> expression language rules, to the sequence of characters:
> url(xxx)
> which is a valid uri-specification.
You are perfectly right to say that. It's exactly what the specification says.
I'm just wondering why the whole specification is constructed that way. There
are well recognized notions around in computer science (a keyword, an
operator, a function, a data type, a constant, etc), and I'm wondering why
things have to be so non-standard in the xsl-fo specfication.
The 1.0 specification has many other great qualities. I know it is not easy to
write such documents, you all did a great work and we are all thankful, and I
don't want to spoil the first birthday party. But if you want ideas on what
could be good new research tracks for XSL-FO 2.0, I think that orthogonality
- and therefore simplicity - is really something that is lacking.
--
- Linux produces remarkedly less hot air than Windows: under
Windows, the processor gets hot after just a few minutes...
- Yes, but it never stays on long enough to burn out!
|
