xml-dev - RE: Multi-lingual experiment - a call for action

["Didier PH Martin" <martind@netfolder.com>]

Hi Alexander,

Alexander said:
a) translating the DTD doesn't help much, because then we won't be able to
produce one unified XML document that contains all the languages we need for
testing purposes

Didier replies:
You probably missed something in the original post. I said that the
experiment would be about the output of a data source. As we all know. Do
we? most actual data source fields and tables are already encoded in native
languages. If you do a request to these data sources, you'll get a document
having a native language DTD. So, I asked for translation in different
language because, if these are all translation of the same DTD (I mean same
semantics not necessarily same language ;-) people can pick the one they
want to use for their test and recognized what is meant. Also, any
experiment has to be with invariant and variables, we choused to have an
invariant syntax for a particular domain language (even if this represent
the output of a data source it is a kind of domain language) the syntax and
structure rules are based on the XML recommendations, an invariant semantic
(the fields meaning and structure are invariant) but we'll have variable
language encoding of the element, attribute and data content.

Alexander said:
b) it would be better to include longer test passages in the sample document
so that there is something that can really be translated

Didier replies:
It is not an experiment about translating content but more an experiment to
show a possible usage of XML as an output format for data sources. As we
know, data repositories contain native language fields and tables names,
therefore, any request to these data sources will result with a document
having a native language DTD encoding. The experiment or experiments will
show how we can use the XML framework technologies to manipulate these
documents and maybe, the by-product of these experiments will be that people
will discover that what is more important is the semantics not the language
used to encode the elements and attributes. Also, maybe discover that tools
like XSLT can also be used as translation tools not only as transformation
tools. And finally, maybe discover something I even didn't envision yet. The
whole purpose of this experiment is to _explore_ the implications of
multi-lingual environments, data source data exchange (between data sources
encoded with different schemas and data sources fields and tables encoded in
different native languages - as we'll encounter in the real world)

Alexander said:
c) due to the inherent unsuitability of UCS-2 (aka Unicode 16-bit) for
transmission over the Internet in anything but ZIPped or Base-64 encoded
form, I would suggest we stick to UTF-8, which any suitable XML editor will
process correctly

Didier replies:
Yes I agree I discovered that myself during the weekend. You are right, it
is a lot easier to encode the documents with UTF-8 than with UTF-16. And I
discovered that, for instance, Windows' wordPAD can be used to edit
documents encoded in UTF-8 but not UTF-16. So, yes, your point is well taken
Alexander.

[.. post about xml spy...]
Its OK for your own experiment to use XHTML since you wanted to demonstrate
the UTF-8 capabilities of your editor. The actually proposed experiment has
a different knowledge target and different purpose. It is not a commercial
demonstration, more an experiment to explore the XML usage for data sources
in a context as close as possible to the cultural environments experimented
in the day to day realities. It is not about knowledge encoding but more
about data source result encoding (even if some explicit or tacit knowledge
may be encoded in data sources - but this is an other story...).

cheers
Didier PH Martin
----------------------------------------------
Email: martind@netfolder.com
Conferences: XML Europe (http://www.gca.org)
Book: XML Professional (http://www.wrox.com)
column: Style Matters (http://www.xml.com)
Products: http://www.netfolder.com


***************************************************************************
This is xml-dev, the mailing list for XML developers.
To unsubscribe, mailto:majordomo@xml.org&BODY=unsubscribe%20xml-dev
List archives are available at http://xml.org/archives/xml-dev/
***************************************************************************