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

RE: Object-oriented serialization (Was Re: Some questions)

• From: "Reynolds, Gregg" <greynolds@d...>
• To: 'James Tauber' <jtauber@j...>, xml-dev@i...
• Date: Fri, 10 Dec 1999 18:06:22 -0600

Sorry I'm a bit late with this...

> -----Original Message-----
> From: James Tauber [mailto:jtauber@j...]
> Sent: Sunday, December 05, 1999 7:25 PM
> To: xml-dev@i...
> 
> > The semantic constraints I am
> > talking about are one step away from these "ultimate" 
> semantics; they
> > tell you that an integer contained in a given element 
> cannot be greater
> > than 100, but they don't tell you why. These are still 
> semantics to me
> 
> Ah. This is why I have have some difficulty understanding 
> some of what you
> are saying. To me, the constraint that an integer cannot be 
> greater than 100
> is not semantics. It's syntax.
> 
> MyInteger ::= ( '100' | digit{1,2} | '-' digit+ )
> 
> or in some more perspicuous grammar:
> 
> MyInteger = Integer x : x <= 100
> 

But this won't work outside of Europe.  You have to have a clean distinction
between syntax and semantics, and an explicit, rigorous mapping from one to
the other.  The symbols used in your example have no intrinsic meaning, just
as numbers have no intrinsic form available to our perception, so the syntax
can only constraint the formal properties of expressions using those
symbols.  Sure, we have conventional meanings for constant symbols like '1'
and '0'; but they're still symbols pointing to something else, and as soon
as you start writing expressions with a different symbol set - in Sanskrit,
say, or Ethiopic, or you name it - then you're out of luck without a formal
semantics.

On the other hand, the cultural interpretation of the denotatum is beyond
the scope of language definition.  Doesn't matter what the user intends to
model using integers; be it widgets, fingers, or planets, the best the
language designer can do is provide a consistent language that accurately
models the integers.

-gregg

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@i...
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To unsubscribe, mailto:majordomo@i... the following message;
unsubscribe xml-dev
To subscribe to the digests, mailto:majordomo@i... the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa@i...)

• Prev by Date: RE: Appending to an XML document
• Next by Date: BizTalk JumpStart Kit - File missing
• Previous by thread: RE: Object-oriented serialization (Was Re: Some questions)
• Next by thread: Rocket framework for creating Web sites
• Index(es):
  • Date
  • Thread