[
Lists Home |
Date Index |
Thread Index
]
> James Clark wrote: "RE: - they ought to have be able to structured
> values, just like elements (i.e. one could view the content of an
> element as a special unnamed attribute), .."
>
> Agreed. Attributes are highly useful as a means to attach metadata or
> annotations to elements. Metadata itself may have structure and may
> have metadata, hence the utility of structured attributes and attributes
> of attributes.
>
> The obvious syntactic device (in a future version of XML) for a
> structured attribute is something that looks much like an element, e.g.
[Interesting example snipped]
I can understand and respect the motivations for this, but I personally do not
believe the benefits would be worth the disruption of such a fundamental
aspect of XML.
I manage structured information through attributes all the time. The way I do
this is to make the attributes contain a term from a controlled vocabulary
maintained elsewhere: frequently in RDF, sometimes in just other general XML
formats.
I know that some folks here cannot abide such solutions, but in practical
work, they do very well for me, and for many others. In fact, I think this
pattern improves design, and that fully atructured attributes would tend to
encourage designs that suffer from excessive coupling. That said, potential
misuse is never really a solid primary argument for avoiding any feature.
--
Uche Ogbuji Fourthought, Inc.
http://uche.ogbuji.net http://4Suite.org http://fourthought.com
The open office file format - http://www-106.ibm.com/developerworks/xml/librar
y/x-think15/
Python Generators + DOM - http://www.xml.com/pub/a/2003/01/08/py-xml.html
4Suite Repository Features - https://www6.software.ibm.com/reg/devworks/dw-x4su
ite5-i/
XML class warfare - http://www.adtmag.com/article.asp?id=6965
|
