Will XML eat the web?

[Tyler Baker]

Pavel Velikhov wrote:

> Paul Prescod wrote:
> >
> > Pavel Velikhov wrote:
> > >
> > > Paul Prescod wrote:
> > > >
> > > > You need to store data that is efficiently maintained relationally. You
> > > > need to *transmit* XML. Why not use a relational database and create XML
> > > > when you need it.
> > >
> > > Transmitting XML in a textual representation is not always a good idea.
> >
> > I don't follow you. XML is, by definition, textual.
>
> I would like to view XML as a logical data model. I hope the actual
> storage model
> will be transparent in the future. I.e. the next generation xml "parser"
> that
> implements a DOM interface should be able to talk to an xml source that
> physically
> is say an OODB and fetch small pieces of the document as they are
> requested by
> the application.

This gets down to the whole argument of what should XML be _primarily_ used for?
The great majority of enterprise/e-commerce applications seem to want to use XML
merely as an object protocol for use in returning query data and passing
transactions.  The internet crowd wants to use XML for presentation content.

I personally agree with you here that the XML spec is merely a standard
serialization format for XML Documents.  In fact you could argue that XML is just
a standard serialization format for any sort of tree based data structure that has
become familiar because of its almost identical syntax to HTML.  If XML did not
have such a familiar syntax, I doubt XML would have the level of interest it
commands today.

> > > When the user does the 'select *' query on an XML database, sending him
> > > a 20 megabyte XML file that he will need to parse, apply an xsl
> > > stylesheet
> > > to and display in the browser is not the best solution. IMHO, the result
> > > of the query should be shipped on demand, when the user is actually
> > > 'looking'
> > > at a piece of an XML file.
> >
> > I don't think I said anything to contradict this, though I would say that
> > by the time you have a "20 megabyte XML file" you've probably already done
> > something wrong. There is almost never a good reason to generate files
> > that large.
>
> I agree, generating 20Mb XML files is bad. However it will happen. If
> you make
> a lot of data available in XML by wrapping a relational database for
> instance
> users/applications will be able to request large XML files.

Yes it may happen for reasons we cannot predict.  But I think the real problem
here is that the DOM itself is just a tree based structure in and of itself.
Databases rarely only model data that can be expressed in a tree, but rather model
data that is a set of complex graphs which XML and the DOM is not geared towards
(minus creative uses of XPointer).

Tyler


xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)