Re: Action-Item: SML

Hi Yves, all,

Many thanks for your review. I have no additional comments, but I think we
should at least make the SML WG aware of what you said about the
localization related functionality.

Felix

>
> Hi all,
>
> I had the action item to look at SML from the viewpoint of ITS.
>
> SML's Sep-12 Working Draft is here:
> http://www.w3.org/TR/2008/WD-sml-20080912/
>
> As far as I can tell SML is more like a vocabulary that is meant to be
> used inside schemas written in XSD or Schematron, or inside
> other XML documents.
>
> With this in mind there is very little that seems related to localization
> and internationalization as most of the language-dependant
> content is actually hold in the host format.
>
> There is however one mechanism provided that is related to localization:
>
> The sml:locid attribute allows you to define the translation location for
> the text content of the containing element.
>
> For example: In this snippet of Schematron code, sml:locid hold the
> location of the translated text for the message of the
> assertion.
>
> <sch:assert test="smlfn:deref(.)[starts-with(u:ID,'99')]"
>  sml:locid="lang:StudentIDErrorMsg">
>  The specified ID <sch:value-of select="string(u:ID)"/> does not begin
> with 99.
> </sch:assert>
>
> While this is a nice feature. The implementation of it is application
> specific. SML does not define how the substitution should be
> don, nor what naming mechanism should be used, nor in what format the
> translation repository should be written in.
>
> There is an appendix here
> http://www.w3.org/TR/2008/WD-sml-20080912/#Localization_Sample that
> illustrate how sml:locid could work.
>
> There are a number of issue with the example, but they are not directly
> related to ITS.
>
> --- The value of the sml:locid attribute is a QName. In the example the
> namespace part is used to point to a directory where the
> translated files are located, and the local part is the ID of the message
> to read. The name of the file itself is made of the
> language tag a character underline, and the word "lang" (or the prefix of
> the namespace, the example is unclear if the fact that
> both are the same is accidental or not).
>
> --- The translation files are properties-like text file, where each entry
> can contains XML markup. This means to translate such
> file, one would have to
>
> It would be better for the translation to be stored in an XML document, so
> its content could be annotated with ITS, and process more
> easily. Having XML entries in a properties file makes thing difficult for
> most translation tools.
>
> --- There is an error/typo in the French message: "L'identifieur
specifie" s
> hould be "L'identifieur specifi? (an accent is
> missing).
>
>
> Overall, I'm not sure hat we should provide for feedback aside from the
> typo. Is there a generic way to associate language tag with
> a file name so the filename could be automatically inferred? I guess the
> Java properties names is something similar.
>
> Any ideas, suggestions?
>
> Cheers,
> -yves
>
>
>

Received on Tuesday, 21 October 2008 06:02:51 UTC