Re: Fwd: [xml-dev] Build applications using the "simplicity stack"

From
Stephen Cameron <>
To
Arjun Ray <>
Date
2014-04-02T01:15:22Z
ID
<CAG=ut6+6KLmdM=2eUc0So=>
Thread
Re: Fwd: [xml-dev] Build applications using the "simplicity stack"

On Wed, Apr 2, 2014 at 11:44 AM, Arjun Ray <[email protected]> wrote:
On Wed, 02 Apr 2014 06:35:48 +1100, Stephen Cameron
<[email protected]> wrote:
| On Tue, Apr 1, 2014 at 10:15 PM, Arjun Ray <[email protected]> wrote:
| > On Tue, 01 Apr 2014 12:35:36 +1100, Stephen Cameron
| > <[email protected]> wrote:

| > | a data-model in XML Format (schema)
| >
| > [...] XML schema formalisms (DTD, RNG, WXS) are for analytic document
| > structures, not - as far as I know or understand them - for arbitrarily
| > general "data models".
|
| [...] These formalisms have types and convey structure and relationships
| between types.

DTD don't have data types at all, besides strings and tokens.  RNG
defers data typing to external libraries.  WXS is fatally infected
with DB-think and thus has some bizarre mishmash.  Anyone remember
gHorribleKludge?  (Hi Amy!:-))

There was room to re-invent SGML notations, but no one (except RNG, in
a sense) went that route.  The ability to represent types _within_ the
formalism is quite limited.

However, the real nub of the matter is that not all data models are
naturally or conveniently reducible to trees, whereas a XML document
is in the first instance a tree, and schemas for XML are therefore for
trees.  The impedance mismatch may be great or small according to
context, but to discount it as inconsequential doesn't strike me as
wise.

This was covered in a thread about using XML as a data-model some months back, that true data-models are graphs, a valuable insight for me.
 
| Is there a more suitable formalism for data-models that you make use of?

UML comes to mind.

| An object-oriented formalism, in UML, uses XML for interchange.

With limited traction and success, if Wikipedia is to be believed (and
in this instance, I don't see why not).


| my concept of leveraging data-models in cost-effective ways, as I tried
| to give some examples of, is my key point.

Fair enough.  In my work, I've never found schemas for XML instances a
good starting point, or even a vital consideration, in system design.

In my reading to date the 'Object-Role Modeling' of Terry Halpin seems useful and soundly based, but some effort still needed on my part to get through his book 'Information Modeling and Relational Databases'  Details here: http://www.orm.net. As a pure data-model, separate from the behavioural aspects of UML, I think a useful exercise.


_______________________________________________________________________

XML-DEV is a publicly archived, unmoderated list hosted by OASIS
to support XML implementation and development. To minimize
spam in the archives, you must subscribe before posting.

[Un]Subscribe/change address: http://www.oasis-open.org/mlmanage/
Or unsubscribe: [email protected]
subscribe: [email protected]
List archive: http://lists.xml.org/archives/xml-dev/
List Guidelines: http://www.oasis-open.org/maillists/guidelines.php