[
Lists Home |
Date Index |
Thread Index
]
* Kenneth Stephen <marvin.the.paranoid@gmail.com> [2005-01-29 16:19]:
> Alan,
>
> I dont think it follows that the structure of information in the
> XML representation needs to match the lexical representation. In the
> loose coupling demonstrated, the information is structured
> heirarchically, even though lexically it is not. To represent the same
> data you show, in a loosely coupled fashion, one would go :
>
> <organization id="org" />
> <person id="john">
> <worksFor />
> </person>
> <person id="sue">
> <worksFor="john" />
> </person>
> <person id="stephan">
> <worksFor="sue" />
> </person>
>
> ....etc. This is pretty much the way the XML serialization of RDF
> works - the structure can
> be derived by an application / parser aware of the structure rules
> without depending on lexical structure.
>
> That being said, I would disagree with the original poster's
> assertion that flat structure is the preferred design. For things like
> XSL processing, a lexical structure that matches the information
> structure is desirable. For things like RDF, the lexical structure is
> not very important. So I think it is a matter of which technologies
> are being used.
I'm coming at XML in general through XSLT in particular, so I
find it cumbersome to have to construct structure.
However, I do see how the structure you provide is easier to
update. An initial transform can put it in the correct order
for, say, rendering an xhtml display of the tree.
To be told that it's application dependent is useful for me to
hear as a student, so that I don't approach problems with
prejudice against one form or another.
--
Alan Gutierrez - alan@engrm.com
|
