Lists Home |
Date Index |
attributes are essential for abstraction, in both markup and data models
eg <p class="emphasised"> or <table name="sales">
in fact we use relatively few abstractions in all our work. if anything
we should be encouraging xml design with fewer elements, and more
attributes - it would be much easier, and more natural.
eg language uses few abstractions - nouns, verbs, adjectives, adverbs,
articles - in a small number of loose syntactic constructions.
eg modern geometry is simple once you understand transformations -
compared to the nightmare of classical geometry
eg people can only work comfortably with about 6 - 8 "things". digits in
telephone numbers are a great example. therefore C is a good language -
relatively few things to remember to use it (classical is better than
standard in this sense), java is bad. simple html is pretty good (could
be better if the dom clutter was cleaned up and uniformly supported),
some xml is dreadful.
my conclusion - we can get a lot further using xml to get better
abstraction and attributes to bind the abstractions to specific
instances. then ordering of attributes is not important, but ordering of
elements might be - particularly if it has to represent structure as in
books etc, or processing sequences (i've got a good start to a
transaction processor running).
On Mon, 2003-06-23 at 06:12, Thomas B. Passin wrote:
> [Simon St.Laurent]
> > I don't find abstraction to be the essence of XML. I find XML to be a
> > text-based format with labels inside - and I do suspect that foundation
> > is what makes it popular, not the damn Infoset.
> I suspect that _metaphor_ is the "essence" of XML and predecessors. Take a
> piece of paper with text (and whatever) on it. What do people naturally do
> with it besides read it? They highlight passages (inline markup), they
> circle or otherwise mark blocks (block type elements), they move blocks
> around with scissors and tape (transformations), and they insert comments
> and instructions at points - like proofreader's marks (PIs and inline
> elements), and they write notes to themselves (more inline elements, though
> XML is weak in its support for annotations, IMHO).
> A "data-centric " view is still that of a series of blocks in a report, so
> it still works with the metaphor.
> I think that whenever some abstraction or feature strays too far from this
> basic metaphor, more and more people start having problems in understanding
> XML. I18n doesn't fit too well, but at least people know about other
> languages. Attributes sort of fit with annotating the purpose of an inline
> markup, so they sort of fit but still people are arguing about them because
> they do not fit all that well.
> Either we should try to stick within the metaphor or we should replace it
> with a more compelling one. Several can even be used together, but what
> does not work so well is to have a lot of features that do not fit any
> familiar metaphor.
> XML Schema, anyone?
> Tom P
> The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
> initiative of OASIS <http://www.oasis-open.org>
> The list archives are at http://lists.xml.org/archives/xml-dev/
> To subscribe or unsubscribe from this list use the subscription
> manager: <http://lists.xml.org/ob/adm.pl>