[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: Programming for Markup vs. Markup for Programming

• To: 'Didier PH Martin' <martind@n...>, "'Simon St.Laurent'" <simonstl@s...>, xml-dev@l...
• Subject: RE: Programming for Markup vs. Markup for Programming
• From: "Bullard, Claude L (Len)" <clbullar@i...>
• Date: Mon, 5 Aug 2002 08:54:07 -0500

I understand the advantages where one wants and 
needs aggregation.

Aggregation is not always useful.  There are lots of 
examples where it is better to keep the markup and the 
instance simple by reinventing the markup rather than 
attempting by namespace assignment to provide a means 
to identify the processors in the framework.   If the 
namespace is dereferenced and points to something like 
RDDL, que bueno, but that is just a level of indirection 
for choosing the interpreters to choose the referents. 

So, yes, useful, but not necessary.

To me, that means XML 1.0 + Namespaces 1.0 where the 
first choice is to use XML, then one chooses to use 
Namespaces.  The dependency goes from left to right, 
not right to left.

len

-----Original Message-----
From: Didier PH Martin [mailto:martind@n...]

Hi Len,

Len Said:
Yes, I realize that people want namespaces in core.  It's 
a bad idea and if these threads don't get that across 
clearly, people aren't listening.  There is a lot of 
basic work that can be done that never touches namespaces. 

Didier replies:
I can say from a practical perspective that namespaces are useful since
one particular feature of XML is precisely to enable the creation of a
new domain language by assembling other domain languages. I do not have
this advantage with other languages and this is probably the biggest
invention of XML. However, the inherent problem with assembling
disparate domain languages is that they may use the same word for a
different meaning. 

An other advantage but not actually in the spec is the capability to
relate the namespace to some documentation, a document identified by the
namespace URI giving more information about this
vocabulary/structure/semantics construct set. I cannot easily send a C++
or a smalltalk spec and related document with a C++ or smalltalk program
but I can link a namespace to an on-line documentation. If the whole
community including W3 would simply, for a moment, stop the Byzantine
fights and think in "practical" terms of what can _really_ help the XML
framework users whatever them call themselves programmers or XML
authors, we would progress in the right direction.

• Prev by Date: RE: Re: URIs, concrete (was Re: Un-ask the qu estion)
• Next by Date: Re: Re: maps
• Previous by thread: RE: Programming for Markup vs. Markup for Programming
• Next by thread: RE: Programming for Markup vs. Markup for Programming
• Index(es):
  • Date
  • Thread