XML to XML conversion

[Jure Sah]

Hello,

I vageuly remember a standard that allowed for an XML document using one
tag vocabulary to be translated to a diffirent XML document using a
diffirent tag vocabulary using the information provided in a kind of
dictionary XML pointed to in an URL present in the original XML file. I
am very interested in trying it out for various applications, but I
can't find anything on the internet about it, since I don't remember
it's name, only the general appearance. Can anyone help me with this?

What I remember, approximately:

The idea is that you have one XML file, in which the namespace is
defined in the standard way, using an URI. The URI, however, is pointing
to an online XML file. For example:

<x xmlns:abc="http://www.server.org/dictionary.xml">
<abcne>orange</abcne>
</x>

This dictionary.xml then contains, for example:

<root>
<one>
<format1>1</format1>
<format2>uno</format2>
<format3>ena</format3>
<format4>ein</format4>
</one>
</root>

So a program wishing to convert the original XML to, say format2, would
end up with the following XML:

<x>
<uno>orange</uno>
</x>

What I remember is that this dictionary trick was mainly used to allow
diffirent software to be able to interpret the same XML file, without
all of the software needing to be programmed for the same tag
vocabulary. The program would just have the translations of the tag
meanings prepared specialy for it in the dictionary file.

I remember(!) seeing this implemented somewhere and I don't want to
create my own standard here even if what I remember was just somebody's
bright idea and not an official XML standard. This little trick gives a
solution to creating globaly-compatible software when companies that
provide XML data sources can't agree on a common tag vocabulary. I need
more information so I can implement it's use compatibly.

Anyone?

Thanks in advance.

--
Model: INFJ
Primary function: Coprocessor
Secondary function: Cluster commander

Yes I'm a therian:
http://www.wikitherian.org

Creationism & Darwin:
"The bible says humans were
supposed to use animals to
do work for them and I like
to work so I must be an animal!"

 

[Tjerk Wolterink]

Hello,

cut

You should make the tags names language independent.
Only the content should be able to contain different languages.

But back to your question:
Use XSLT.

 

[Jure Sah]

You should make the tags names language independent.
Only the content should be able to contain different languages.

The tag names got nothing to do with what I can do, they are obviously
determined by the XML data provider, not by the processor.

But back to your question:
Use XSLT.

I've looked into it and the whole thing reminds me a bit of SQL queries,
obviously more suitable for transforming databases than interpreting an
XML files in various ways. Nevertheless it appears to be the standard I
asked for.

I am curious tho, how to provide an XML file and include (in a standard
way) all the intformation required to allow the file to be interpreted
by all the programs which's vocabularies we knew about at the time the
XML file was written.


--
Model: INFJ
Primary function: Coprocessor
Secondary function: Cluster commander

Yes I'm a therian:
http://www.wikitherian.org

Creationism & Darwin:
"The bible says humans were
supposed to use animals to
do work for them and I like
to work so I must be an animal!"

 

[hawat.thufir]

Jure Sah wrote:
....

I've looked into it and the whole thing reminds me a bit of SQL queries

....

Are you thinking of XQuery <http://www.w3.org/XML/Query/>?


-Thufir

 

[Patrick TJ McPhee]

[...]
% > But back to your question:
% > Use XSLT.
%
% I've looked into it and the whole thing reminds me a bit of SQL queries,
% obviously more suitable for transforming databases than interpreting an
% XML files in various ways. Nevertheless it appears to be the standard I
% asked for.

I don't know why you get that idea. A stylesheet that converts "one" to
"uno" is

<?xml version="1.0"?>
<xsl:stylesheet version="1.0"
xmlns:xsl="http://www.w3.org/1999/XSL/Transform">

<xsl:template match="*">
<xsl:copy><xsl:apply-templates select ="node()|@*"/></xsl:copy>
</xsl:template>

<xsl:template match="one">
<uno><xsl:apply-templates select ="node()|@*"/></uno>
</xsl:template>

</xsl:stylesheet>

This could be extended to read the names of the elements to transform
from some other file.

% I am curious tho, how to provide an XML file and include (in a standard
% way) all the intformation required to allow the file to be interpreted
% by all the programs which's vocabularies we knew about at the time the
% XML file was written.

I'm not sure it makes sense to do this. In any case, if you can depend on
all programs supporting XSLT, and supporting the xml-stylesheet PI, the
standard way to point an application to a stylesheet is to add something
like this:

<?xml-stylesheet href="http://my.machine/path/to/file" type="text/xsl"
title="type-of-conversion"?>

before the first element of the document.

 

[Peter Flynn]

The tag names got nothing to do with what I can do, they are obviously
determined by the XML data provider, not by the processor.


I've looked into it and the whole thing reminds me a bit of SQL queries,
obviously more suitable for transforming databases than interpreting an
XML files in various ways.

Er, no. XSLT is designed explicitly for transforming XML documents.
It doesn't have anything even remotely to do with databases or SQL.

Nevertheless it appears to be the standard I
asked for.

Actually the more I think about it, the more I'm sure there was some earlier
(even pre-XML) system that used translation files like you described. But I
may just be having a brain-fart.

I am curious tho, how to provide an XML file and include (in a standard
way) all the intformation required to allow the file to be interpreted
by all the programs which's vocabularies we knew about at the time the
XML file was written.

I don't think that's really what XSLT was designed for. If you're talking
about mapping the *interpretation* of markup, that's really a job for
something like ISO 11179 or Topic Maps.

///Peter