RE: [xml-dev] Re: determining ID-ness in XML

["Bullard, Claude L (Len)" <clbullar@ingr.com>]

Ever inscribe Arabic numerals into stone?  Requirements 
count.

Weigh costs and returns before investing.  We are 
stuffing more things into the XML system vocabulary.  
Every few months we get yet another "gotta have" from 
the XML cognoscenti that makes the tools churn.

I understand why one wants an ID. I don't accept "overkill" 
as a requirement. Too nebulous.   What we are actually 
up against are the limits of well-formedness, so we are 
starting to mix in validity requirements because we don't 
like DTDs or schemas.  Nyet.  Propose the reason why 
the standard means don't work and I don't think anyone 
will take DePH seriously at this point.  Saying SOAP 
deprecated it simply says SOAP hasn't followed the 
spec and that says a lot about the W3C specs: no, 
they are not standards. They are lots of little 
"our way or the highway" manifestos.

len

-----Original Message-----
From: Champion, Mike [mailto:Mike.Champion@SoftwareAG-USA.com]

> I think we should not add features to a language for which 
> features exist unless we can prove we didn't create the 
> problem to begin with.

I'm sure that someone made a similar argument 1000 years ago.  "The Roman
Numerals are an international standard, and we shouldn't go with these
new-fangled Arabic things unless you can prove that arithemetic is
impossible using the standard."
<grin> 

Seriously, IDs are the ONE THING that you need to define in a DTD that other
specs -- which otherwise simply refer to well-formed XML -- keep stumbling
over.  The DOM does, Xpath/XPointer does, and so on.  I'm as we speak
writing a little "extract data from an HTML form, put it in an XML instance"
app, and stumbled over the "I can get the data from HTML via an ID but cant'
find the XML element to update without searching for it" problem.  Sure I
could define a DTD or schema, but that would be overkill for me and
seriously limit the consumers of my little app.  (It's an "educational"
example, so it's gotta be very, very simple and generic).  xml:id or one of
the other proposals would make this use case much easier to implement and
understand.