RE: [dita] Proposed changes to the xml:lang proposal

["Grosso, Paul" <pgrosso@ptc.com>]

OK, if we are removing this completely, I guess the point 
I'm about to make is moot.

But, no, you cannot "use the xml:lang attribute to identify 
the boundaries...."  An attribute cannot identify boundaries.
This is what I have been trying to say in this thread.  You
have to have a (non-empty) element--with a start tag and
end tag--to identify boundaries, and then you specify the
xml:lang attribute with the appropriate language value on
that element's start tag.

And I don't care if the audience is non-technical, you are
doing no one a favor by using imprecise language.  Attributes
cannot identify boundaries.

paul

> -----Original Message-----
> From: Gershon L Joseph [mailto:gershon@tech-tav.com] 
> Sent: Monday, 2006 April 10 13:16
> To: dita@lists.oasis-open.org
> Subject: RE: [dita] Proposed changes to the xml:lang proposal
> 
> Actually, the existing Recommended Usage section already 
> sufficiently explains how different languages are to be 
> marked up, so I don't see the need to add anything to this 
> section. I'll simply remove it from the Notes to Vendors section.
> 
> 
> Best Regards,
> Gershon
> 
> -----Original Message-----
> From: Gershon L Joseph [mailto:gershon@tech-tav.com] 
> Sent: Monday, April 10, 2006 8:51 PM
> To: 'Eliot Kimber'; dita@lists.oasis-open.org
> Cc: 'Grosso, Paul'
> Subject: RE: [dita] Proposed changes to the xml:lang proposal
> 
> We discussed this thread at length in today's translation 
> subcommittee meeting. The suggested rewrites are not in sync 
> with the W3C and Unicode specs (we have representatives from 
> both groups on the SC). However, we agree that since 
> applications cannot correctly read the language from the 
> characters, we should move this paragraph from the Note to 
> Vendors section to the Recommended Usage section, and 
> rephrase it to apply to authors and translators (the humans, 
> not the machines). Here's the rewrite:
> 
> When working on a document that contains strings of 
> characters from different languages, the author should use 
> the xml:lang attribute to identify the boundaries of each language.
> 
> 
> Best Regards,
> Gershon
> 
> -----Original Message-----
> From: Eliot Kimber [mailto:ekimber@innodata-isogen.com]
> Sent: Monday, April 10, 2006 6:11 PM
> To: dita@lists.oasis-open.org
> Cc: Grosso, Paul
> Subject: Re: [dita] Proposed changes to the xml:lang proposal
> 
> Grosso, Paul wrote:
> > That is why I think saying "the language...should be indicated via 
> > markup" is misleading, because is the string itself (as well as its
> > language) that needs to be indicated via markup.
> 
> And just to be a bit more pedantic, I think it would be 
> clearer and more accurate to say "text" instead of "string".
> 
> While the XML data is held as strings of characters, 
> semantically when you talk about *language" you are dealing 
> with "text" rather than strings of characters.
> 
> I would think that the clearest statement would be something like:
> 
> "As a matter of practice, text that is in a language 
> different from the base language of its container should be 
> contained in XML markup that uses the xml:lang= attribute to 
> indicate the text's national language."
> 
> Cheers,
> 
> Eliot
> 
> 
> 
>