Re: xml:lang [was Re: Outstanding Issues ]

Martin,

Martin Duerst wrote:
> 
> At 10:08 02/03/07 -0800, A. Vine wrote:
> >And to add to Misha's comment:
> >
> >There is no tag for locale.  Anywhere.
> >
> >The reason often stated is that locale is a client concept, not a data
> >concept.
> >Of course, in the example below, it is a data concept.
> 
> In the example of 1.234,56 appearing in text (e.g. HTML), we as humans
> see it as data. 

It is locale-formatted data, which has nothing to do with the client it's on.

> But for the computer, it's just part of a string.
> Marking it up, e.g. with something like
>      <number value='1234.56'>1.234,56</number>
> (not available in HTML) will identify it to the computer as data.

Whether it has a value to the computer depends very much on the nature of the
processing going on.  You can call it a string, or you can call it text, or you
can call it data.  What it is _not_ is locale-neutral.  It has a locale.

The point being that this is an example where only having the locale info at the
client is not sufficient.  Even if you have a value parm in the tag, the data
itself still has a locale - and you, the creator of the processing software, may
very much want to know what that locale is.  You might want to change the way it
is displayed, or put a pop-up tag for a mouse-over to alert the user, or who
knows what.  There are times where the locale tag would be ignored, just as the
lang tag is sometimes.

I recognize that there are 2 aspects to locale-specific info, syntax and
semantics.  This could simply be syntactically locale-specific.

The value parm would be useful for calculations and comparisons done on the
actual value.

> 
> This is similar to having "This document was authored by
> Ora Lassila" in plain text; RDF doesn't automatically capture it,
> it has to be transferred into RDF. Going up the semantic ladder
> is difficult and requires work.
> 
> >Another reason is that
> >locales are not standardized - and this is actually a bigger problem.
> 
> This has come up at the recent I18N Workshop
> (http://www.w3.org/2002/02/01-i18n-workshop/). If you are interested
> in further discussion on what W3C should do in this area, please
> subscribe to www-i18n-workshop@w3.org (sending 'subscribe' in the
> subject to www-i18n-workshop-request@w3.org).
> See also the announcement e.g. at
> http://lists.w3.org/Archives/Public/www-international/2002JanMar/0088.html.

I assume this message is for other people on the list - as you know I'm already
subscribed.

> 
> >In order to determine equivalency between 2 different locale-based formats,
> >standard internal representations would have to be agreed upon, which are not
> >necessarily US format.  In fact, most internal representations of numeric
> >values
> >are usually more cryptic than a locale-based format, for efficiency.
> 
> The W3C has a collection of standard formats for data exchange, defined
> in XML Schema Part 2: Datatypes (http://www.w3.org/TR/xmlschema-2/).
> Some of these are close to US formats, but some of them are quite different,
> but all of them are textual.

Right, I'll make sure to document this info for my groups here, thanks.

Andrea

Received on Friday, 8 March 2002 13:51:47 UTC