Mon, 24 Apr 2000 20:56:49 +0200
Paul Prescod wrote:
> Christian Tismer wrote:
> > It looks so, yes. But it's all XML, no DTD needed,
> > and it is much more powerful and easy to read.
> Well, those are both arguable. Somewhat more powerful, yes.
> Easier to read? If you don't know either language, yes, it is easier to
> read, as COBOL is easier to read than Python. But once you've learned
> both languages the more compact syntax means that you can read faster.
> It's a difference between having fifty one-line declarations on a screen
> at one time versus having ten five-line declarations.
Yes, but... I usually avoid as most as possible MS stuff,
but in this case I'm guilty: An XML schema looks just great
when you browse it with IE5. You can fold nodes, you have
a quite readable XML document.
Since I don't have such a tool for DTDs, browsing schemas
is in fact easier for me than DTD text files.
Let's see the schema structure under a different aspect:
It is in fact a higher level structure, which is expressed
somehow in XML. I agree that there might be better
visualizations of the model. Maybe this could be visualized
by an application on top of schema.
I also agree that links need to be supported, nut just trees.
> Schemas allow a few more types of validation and extension in exchange
> for complexity and verbosity. Judged entirely on this basis, the debate
> could go either way but political and historical factors are more likely
> to be the decision makers.
Ok, I think of it like of a Python structure which is serialized
as an XML pickle. Of course this doesn't show the real structure,
it is not perfect and verbose. But what I'm in fact after is the ability
to define types, have more control over my XML documents.
If the state-of-the-art happens to be schema, well I'll take
it, but I'm not excited about it. Give me something more
powerful with a good chance to get widely accepted, I'll
switch immediately. Schema is just better than nothing.
Has somebody looked into XMLAuthority? I'd really like to
know what their native data abstraction is. They seem to
be very flexible in reading and writing different formats.
Perhaps the *right* way to do it hides somewhere in their
ciao - chris
Christian Tismer :^) <mailto:firstname.lastname@example.org>
Applied Biometrics GmbH : Have a break! Take a ride on Python's
Kaunstr. 26 : *Starship* http://starship.python.net
14163 Berlin : PGP key -> http://wwwkeys.pgp.net
PGP Fingerprint E182 71C7 1A9D 66E9 9D15 D3CC D4D7 93E2 1FAE F6DF
where do you want to jump today? http://www.stackless.com