xml-dev - Re: Strong Typing in SGML and XML

[Peter@ursus.demon.co.uk (Peter Murray-Rust)]

In message <3.0.32.19970506102133.007a1bf0@pop.intergate.bc.ca> Tim Bray writes:
> At 09:44 AM 5/6/97 GMT, Peter Murray-Rust wrote:
> 
> >> 3. The syntax for dates and so on should match some ISO standard,
> >>    but I haven't found which one yet.
> >
> >Do you mean you there are several and you haven't decided between them?
> >I thought that people had converged on a single one (I can't remember
> >the number, it's something like 8601).
> 
> I mean I spent half an hour poking around the Web and didn't come
> up with anything right away.  If someone will send me a pointer to
> the standard syntax, I'll put it in the draft.

ISO 8601.  Being ISO it isn't on the WWW, but there is a very concise
summary which I found at http://www.mcs.vuw.ac.nz/  - just look
for ISO8601 in the search engine.

It manages timezones within the date, and dates and times both absolute and
relative. 
> 
> >I don't find SQLSIZE 'obvious'
> 
> OK, all of the types but one need a single parameter; each parameter
> is numeric, except for DATE, which is a boolean for timezone
> existence.  I didn't want to make up different attributes for each one.
> Yes, it's hopelessly overloaded.  Maybe it should just be called
> XML-SQLPARAM.  It is *not* the case that there is a single concept

I much prefer this.  OTOH some might require two?

[...]
> 
> Yes, this has to be supported.  Somebody else pointed that out too.
> 
> >Is equality defined/definable for floating point?
> 
> Yes, because in the real world, there are no real numbers [sorry,
> math joke] - what I mean is that floating point numbers exist either
> as fixed-size binary objects in computer storage, or as strings of
> digits, decimal points, and exponents, also in storage.  Either
> way, equality tests are meaningful.  Given good implementations of
> the IEEE rules, they are even useful.

As always this has to be precisely specified.  It should be clear whether
a number in memory is being compared as its IEEE representation od something
else.

> 
>  -Tim

A general point about validation which I keep labouring and not making 
much headway is where does all this happen?  It can happen at authoring, 
at parsing, or at the application.  My concern is that unless this is defined
it's likely to fall through the net.  And having built strong typing into
CML, it's not always trivial to implement (in fact I'm sure it's not 
correct in places).  For example, should the system always hold a string
value regardless of the original type?  And if it converts back to a 
string representation presumably it should use the original string rather
than reconvert.  What happens with transformations is certainly not trivial,
because it can involve precision and output format.

	P.

> 
> xml-dev: A list for W3C XML Developers
> Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
> To unsubscribe, send to majordomo@ic.ac.uk the following message;
> unsubscribe xml-dev
> List coordinator, Henry Rzepa (rzepa@ic.ac.uk)
> 
> 

-- 
Peter Murray-Rust, domestic net connection
Virtual School of Molecular Sciences
http://www.vsms.nottingham.ac.uk/

xml-dev: A list for W3C XML Developers
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To unsubscribe, send to majordomo@ic.ac.uk the following message;
unsubscribe xml-dev
List coordinator, Henry Rzepa (rzepa@ic.ac.uk)