[
Lists Home |
Date Index |
Thread Index
]
On Thu, 6 Feb 2003 15:59:33 -0500, Elliotte Rusty Harold
<elharo@metalab.unc.edu> wrote:
>> Which ones are wrong today?
>
> I don't think any of them are wrong today, but 4 and 5 were not fully
> satisfied by the spec that was produced.
>
>> What's missing?
>
> A clear description of the information from the document that a processor
> must pass to the client application.
I completely agree! And amidst these points is where the permathreads are
born:
>> * It shall be easy to write programs which process XML documents.
This runs into the numerous strongly held but incompatible conceptions of
how one writes programs, e.g. static typed vs dynamic typed, event-driven
or datastructure-centric, big-design-up-front vs extreme programming, hand-
coded vs wizard-generated, and so on.
>> * The number of optional features in XML is to be kept to the absolute
>> minimum, ideally zero.
This runs into the fact that what is essential to one community (e.g.
docheads) is optional to another (e.g. dataheads) and vice versa.
The "clear description of the information from the document" is the InfoSet
permathread.
Since the namespaces cluster-^H^H^H^H^H^H^H^H permathread wasn't addressed
in XML 1.0, we can't blame that one on the original requirements.
|