Properly formatted XML will always be self describing.
Example:
<300><100/><200></200></300>
Self describes just as:
<PurchaseOrder><ShippingAddress/><PurchaseOrderNumber></PurchaseOrderNumber></P
urchaseOrder>
self describes.
Humans will not be sitting around a year from now reading this stuff. It's all
going to be processed behind the scene by applications and systems. A human
interface will only be needed IF there is a breakdown. At this point, an
application can display the tags to the human in an appropriate language.
Troy
[email protected] on 9Mar2000 06:50 AM
To: [email protected]@INTERNET
cc: [email protected]@INTERNET (bcc: Troy R.
Lowe/0A/Caterpillar)
Subject: Re: languages and tags
Retain Until: 04/08/2000 Retention Category: G90 - Information and
Reports
Caterpillar Confidential: Green
Bernd,
one of the main goals using XML is that the result should be a
self-describing structure as far as this is possible. My personal opinion is
that coded tag names with description attributes included would semantically
violate XML guidelines and philosophy?! Such an approach might be used for
schema definition (for technical reasons). Concrete XML instances however
should follow strict semantic guidelines.
regards
Achim
"Eckenfels. Bernd" wrote:
> Hello Achim and List,
>
> given one possible solution to take the whole EDIFACT Data Dictionary I
> think it would be a very painfull job to translate all the Segment,
> Element and Code Entries into Memonic english. Some automated jobs
> (first three words of the official EDIFACT description for example) wont
> produce very good tag Names in all parts of the system. Therefore I
> think coded elements are the much better solution here.. after all we
> are talking about data processing not printing. It is quite easy to
> store the Full Text Name of every Element in an attribute and use it in
> a XSL Stylesheet for Visualisation.
>
> This means:
> generic XSL+language
> selector
> coded-XML-EDI-Message + DTD-with-fixed-attributes) ---> HTML with
> localized output
>
> Greetings
> Bernd
>
> > -----Original Message-----
> > From: Achim Gasch [mailto:[email protected]]
> > Sent: Thursday, March 09, 2000 10:26 AM
> > To: [email protected]; [email protected]
> > Subject: Re: languages and tags
> >
> >
> > Matti,
> >
> > I have also made experiences in multilingual environments for
> > years and I
> > completely agree to what you say. Using a 'language' for
> > describing semantic
> > relationships between semantic entities has no need for a
> > NATURAL language,
> > english would more be used as an semi-artificial
> > metalanguage, just because it's
> > te most common one on the global level
> >
> > regards
> >
> > Achim
> >
> > ------------------------------------------------------------------
> > Achim Gasch - SAIOS SA
> > mailto:[email protected] / http://www.saios.com
> > Tel: + 352 40 55 40 415 / Fax: + 352 40 55 40 6180
> > 90-A Rue de Strasbourg, L-2560, Luxembourg
> > ------------------------------------------------------------------
--
------------------------------------------------------------------
Achim Gasch - SAIOS SA
mailto:[email protected] / http://www.saios.com
Tel: + 352 40 55 40 415 / Fax: + 352 40 55 40 6180
90-A Rue de Strasbourg, L-2560, Luxembourg
------------------------------------------------------------------
