RE: [Bldgcomposer-research] relative reality!

[Zwicky, W. R ERDC-CERL-IL C. <Wil...@er...>]

Not very deep.  I realized this about XML when I first learned it.  The
distinction between attributes and sub-elements is weak, and not terribly
useful for real modelling.  In other words, the difference exists only in
the XML language, it has little value to the data I try to model.  When I
design XML and face this problem, my criteria are usually technical rather
than logical.  i.e.:
 - Is the value one line or several?
 - Need to embed sub-elements within the value?
 - What's easier for a hand-coder to type?
 
And the point about "iscoid" vs. "knee back" is obvious too.  XML can be a
highly discriptive language, you need to think about how far you need to go.
Sometimes it's better to say <class> ... </class>, sometimes <object
type="class"> ... </object>.  Sometimes it's
  <object> <subobject/> </object>
and sometimes
  <object/> <subobject/> <relate first="object" second="subobject"/>
 
The issues about spoken language aren't so strong in USA, we invent and
steal new words all the time.  The open-source movement uses the French
words "libre" and "gratis" all the time 'cuz they mean different things, but
in English both translate to "free".
 
-Bill

// William R. Zwicky
// USA-CERL  (217) 352-6511 x7405
// Champaign, IL