W3C home > Mailing lists > Public > public-i18n-its@w3.org > July to September 2005

Re: Purpose specification mapping - additional purpose

From: Felix Sasaki <fsasaki@w3.org>
Date: Thu, 01 Sep 2005 01:04:50 +0900
To: "Martin Duerst" <duerst@it.aoyama.ac.jp>
Cc: "public-i18n-its@w3.org" <public-i18n-its@w3.org>
Message-ID: <op.swdqycchx1753t@ibm-60d333fc0ec.w3.mag.keio.ac.jp>

On Sat, 27 Aug 2005 14:18:50 +0900, Martin Duerst <duerst@it.aoyama.ac.jp>  

> At 22:36 05/08/26, Felix Sasaki wrote:
>  >
>  >Hi all,
>  >
>  >After reading something about the work of UBL, i.e. localization of UBL
>  >invoices
>  >
>  >http://lists.oasis-open.org/archives/ubl/200310/msg00010.html
>  >
>  >I thought that we could widen the scope of purpose specification  
> mapping.
>  >Currently it is mapping of existing vocabularies to ITS. Another aspect
>  >would be mapping from one language to the other, e.g.
>  >
>  ><para> to <absatz> or to <谿オ關ス>
> This either will result in simple one-to-one correspondences
> that can be implemented by a simple, straightforward XSLT, or
> it will end up trying to address the generic problem of mapping
> from a DTD/Schema A to another DTD/Schema B with correspondences
> that are not straightforward (but for which again most probably
> XSLT is the best answer, to the extent that the mapping can
> be formalized).

Another answer would be SGML architectural forms, or a solution like XLink  
1.1. - which is inspired from architectural forms.

> Seen that way, the internationalization/localization-specific
> aspect in this problem is very low. Expressed another way, if
> you want to map between two existing DTDs/Schemas, or create
> a new one based on an old one, you may need a translator and
> application expert to give you the corresponding element/attribute/...
> names, but not an internationalization/localization expert.
> This assumes that you don't e.g. want to add ruby when moving
> from German to Japanese, but we are already covering these
> things, so we should be okay.
> If I missed something, please tell me.

You don't miss something. It would be close to a simple one-to-one  
mapping. For some use cases that might be useful, e.g. the UBL example, or  
to ease the application of schemas written in language A for authors who  
only know language B.

Received on Wednesday, 31 August 2005 16:05:04 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:43:05 UTC