[
Lists Home |
Date Index |
Thread Index
]
Seairth,
I think you just restated the document vs. data conundrum. If you're writing
markup for a document for humans, all you may need is:
> <poem type="limerick">
> There was a young lady named Bright
> Whose speed was much faster than light.
> She set out one day,
> in a relative way,
> And returned on the previous night.
> </poem>
>
But if you're doing automated limerick validation, that's not enough
information. I would say the data is unnormalized for that intent. (There
are conflicting definitions of "normalize"; I mean it in the data modeling
sense of the term.) Too many compound data fields, in other words, forcing
the validation engine to do its own parsing for syllabilification.
Your other example is more like it:
> <poem type="limerick">
>
> <line><foot><syllable>There</syllable><syllable>was</syllable>
> <syllable>a</syllable></foot>...</line>
> .
> .
> .
> .
> </poem>
>
There's not quite enough information in the above, either, but it's closer.
In my opinion, XML is just a syntax for describing things. It may not be the
most efficient or intuitive, but I think it has enough universality to
handle just about description you care to write in it, to any level of
detail.
Yesterday someone told me he didn't think XML was a good language for data
modeling (I have a feeling he doesn't like SQL DDL, either). What is 'good'?
There's no universal 'good' that can be applied in every situation. Good is
what fits the constraints of the system.
Later,
Jeff
|
