[
Lists Home |
Date Index |
Thread Index
]
Uche Ogbuji wrote:
> >
> > So yes, as long as it can be transformed ... to the extent that such a
> > transform provides a logical equivalence i.e. that we are talking about
the
> > same value with different lexical representations ... then I am in the
camp
> > that folks should choose whatever they prefer ** understanding that
certain
> > types of software, renderers, and a host of unpredicable issues will
guide
> > which lexical representation is appropriate for which application. More
> > succinctly: I don't care :-)
>
> I must have missed something. This is not what I understood you to be
saying,
> and glancing back at your recent posts, I'm still having trouble
reconciling
> those with the above.
>
> No matter. I'm happy to accept a satisfactory conclusion :-)
Well I had admitted I wasn't exactly sure what the fuss is about, so perhaps
I hadn't made myself clear enough.
Different formats each have their advantages. The clear advantage of
explicit tags, is that the meaning is less ambiguous. e.g.
<date>
<day>02</day>
<month>02</month>
<year>01</year>
</date>
vs.
"02-02-01"
on the other hand if the user _knows_ that the order is _always_ month, day,
year then the text string is passable (In the U.S. a medical prescription
dated such or a legal document dated such is accepted as such).
With XSLT it's pretty easy to generate the latter from the former, and with
regular expressions its pretty easy to parse the latter into the former so
both can coexist (understanding that for the purposes of dumping a document
into an XML database the former is preferable) I'd stated:
>
> I think that Simon's "regular fragmentations" is a perfect fit here. An
> author might quickly type something in which can be 'expanded' into a
> structure given a regular expression. Voila' ... best of both worlds.
Which I'll stand by :-))
Jonathan
|
