[xml-h] what would you like?

Gavin Thomas Nicol gtn@rbii.com
Sun Jan 19 14:35:37 GMT 2003


On Sunday 19 January 2003 05:12 am, Nico van de Water wrote:
> in my view, this also holds true
> for terminology-related solutions, whereby an XML-based repository is
> searched through an XSLT stylesheet and presented in XHTML format.=20

Usually the search happens through some other means. For example, when yo=
u=20
access some content the CMS will look up the URL, find the content, and t=
hen=20
transform the content into XHTML using XSLT, which is what the browser se=
es.

Whether this happens dynamically or statically (i.e you use ant or somesu=
ch to=20
batch convert) is one factor in the choise of CMS... and also, in your wo=
rld,=20
whether you can support multiple languages.

I've personally always been in favor of dynamic delivery coupled with a m=
eans=20
for putting together pages in multiple languages, and that *can* touch on=
=20
hypertext.

> Hence my
> notion that CMS solutions and hypertext are not that far removed from o=
ne
> another. Again, maybe I am wrong here, but then again, someone from thi=
s
> list could enlighten me...

If you think of linking as the problem of asserting a set of relationship=
s=20
over content, then CMS and hypermedia are similar. For example, if you ha=
ve=20
foo-en.xml and foo-fr.xml, you want to say they are linguistic variants o=
f=20
one another, which you can do in a linkbase, but most CMS do not expose i=
t=20
that way.

In our system (we're a CMS vendor) we have an XML file containing metadat=
a for=20
each asset in the system, and have the notion of both "derived objects" a=
nd=20
"related objects". Both of those relationships are maintained in the meta=
data=20
using UUID's in attributes to identify the resources, and we use queries =
to=20
treat the underlying database as a linkbase.





More information about the xml-hypertext mailing list