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

RE: Names As Types

• To: 'Vladimir Gapeyev' <vgapeyev@s...>
• Subject: RE: Names As Types
• From: "Bullard, Claude L (Len)" <len.bullard@i...>
• Date: Wed, 31 Aug 2005 09:10:07 -0500
• Cc: 'Nathan Young' <natyoung@c...>, XML Developers List <xml-dev@l...>

I was talking about stacks that are themselves 
XML-based applications. 

On the other hand, where do the semantics come from?

http://www.newyorker.com/printables/critics/050822crat_atlarge

(Thanks Murray.)

Rick Jeliffe gets it.  As usual.  Although, 
again I ask, what is the impact of substitution 
groups on schema design (they introduce choices 
as a means to label and limit extensibility 
but are they just categories with the ability 
to provide category features)?  

In fact, it isn't all turtles, but that meme 
was presented here and when one starts to 
think it through, it becomes less powerful 
because the systems do work.  Why?  We make 
them work.  We had a very long long discussion 
on the TAG list about the http range issue. 
Eventually it was settled as the clear thinkers 
knew it would be: a system is defined in terms 
of itself.  Appeals to universality as almost 
useless. 

That is why the question is asked, "Are 
you a turtle?"  Mike likes the scientist 
story, but i use the NASA/flight test story. 
When you compare these two, there is a 
meaning because this is how 'clear, clean 
thinking' engineers work versus the campy 
philosophies taking root in the WWW as a 
side effect of architectural principles 
appealed to justify flawed history and 
analysis.  It is political superstition but 
given there are no tests to overturn it, 
it will be remembered and repeated until it 
becomes a turtle truism.  Truisms work until 
they fail.

XML doesn't have names as types.  We use 
XML generic identifiers to name types. 
It IS about communication, not rules. 
Math and logic as systems are NOT more 
real than the turtles, but they are 
reliable and replicable; but we have to 
tie them to XML, so is there a best way 
to go about doing that?  Maybe not, or 
maybe we teach turtle truisms that are 
ok until exceptional conditions emerge 
(No, the levees will hold and the Dome 
can withstand a Cat 5.)

As the article above points out, is 
it really truth one is after here or systems 
that demonstrably work?  True or not, what 
patterns of XML application language construction 
emerge that we can share and rely on?

Using Names for Types:  reliable, replicable, 
a turtle truism that holds?

len

From: Vladimir Gapeyev [mailto:vgapeyev@s...]

Hm, I am starting to understand something: you appear to be talking about 
stacks of turles that do {thinking about, specifing (and reading other's 
specs), setting goals, designing, implementing, using} OF applications / 
systems, XML-based or not.  And I thought the talk was about stacks of 
turtles that ARE themselves XML-based applications / systems!

Can one even hope to conceptualize any useful stacks of the first kind? 
That's just a mess of communicating individuals.  Only math and software 
coming out of this mess are real.  [Wow, I'll call this platonism of the 
3rd millenium!]

• Prev by Date: SV: Names As Types
• Next by Date: RE: Names As Types
• Previous by thread: SV: Names As Types
• Next by thread: RE: Names As Types
• Index(es):
  • Date
  • Thread