A question on nomenclature

Reynolds, Gregg greynolds at datalogics.com
Wed Dec 8 15:40:27 GMT 1999


> -----Original Message-----
> From: James Tauber [mailto:jtauber at jtauber.com]
> Sent: Tuesday, December 07, 1999 7:21 PM
> 
> > <name>
> >   <first/>
> >   <middle/>
> >   <last/>
> > </name>
> >
> > (James Tauber described this as a "schema-by-example") but 
> what I really
> > want is the name that I would call that "class" of XML documents.
> 
> In linguistic terms, you have a "grammar" defining a 
> "language" which is
> really just a set of "utterances".
> 
> In XML, a "grammar" is generally called a "schema" and an utterance is
> called an "instance". So what you are asking, if I understand 
> correctly, is
> what is the term corresponding to "language".
> 

I'm not familiar with this definition of "schema", but then I haven't been
able to follow the discussion on XML Schema Stuff very closely.  "Grammar"
to me suggests syntax, although probably it should mean the whole ball of
wax - syntax, semantics, lexis, etc.  But "schema" to me means (roughly)
"typed", and thus a mapping from syntactic structures to values, which is
extra-syntactic.  In fact I'd argue that XML _syntax_, strictly speaking,
determines only which sentences are legal in the language, and doesn't even
map (concrete) syntactic structures to abstract ones, which is a kind of
semantics.  Well, it does, but very informally and with some ambiguities.

> The term most consistent with the XML 1.0 REC would probably 
> be "document
> type".
> 
> So you would say you have a "schema" defining a "document 
> type" which is
> really just a set of "instances".
> 

I'm confused by David's example - it clearly can only be construed as an
instance in XML terminology.  One can infer any number of DocTypes
(=languages, grammars) from it, but there is nothing in the example to
support choosing one such language over any other.  Also, based on his post
from yesterday, it sounds like he's thinking of a set with only one member.

> 
> 1. Yes, people get confused between a schema and a schema 
> language and use
> "schema" to mean both.
> 

The whole complex of schema-related terms looks terribly ill-defined to me.
Naturally I've got my own little set of definitions, but can you point me to
what you would consider the clearest and most authoritative?  (Remember I'm
often unable to follow xml-dev closely, so please copy me if you respond.)

> 2. There is a distinction between a schema and the set of 
> valid documents
> for that schema (ie a "document type"). It is the distinction 
> between a
> grammar and the language it defines. So you could use the 
> term "schema" for
> the *definition* of the set of valid documents (whether its a 
> DTD, a W3C XML
> Schema or a schema-by-example), but the actual set of valid 
> documents is
> best called something else (like "document type").
> 

I'd suggest good old ZF set terminolgy.  An expression that explicitly
enumerates the members of a set is called an extension expression, and an
expression that logically describes the set is called a set comprehension.
So "{1, 2, 3}" is an extension expr., and "{ i : Z | 0 < i < }" is a
comprehension expression denoting the same set.  (I believe there are some
other terms in use, such as intension, but these two terms are common, and
both are used in Z.)  So the set of all documents that conform to a
particular DTD can be considered the extension of the  set defined by that
DTD, which itself is analogous to a set comprehension expression - call it a
Doc. or Lang. comprehension expression.

> Hope this helps

Ditto.

-gregg

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





More information about the Xml-dev mailing list