Re: Translation control in HTML5

Responding to this mail and others in various sub - threads ..., see >FS 
below

David Muschiol wrote:
I am not Ian – may I put my two cents in anyway? As mentioned above,
I do not consider XPath an option for the masses of developers either.

 >FS: I think for *rough* compatibility with ITS the most important 
parts are: having local markup (a "translate" attribute), a global means 
of selecting several nodes (XPath or something else), and the same 
behavior of these with regards to inheritance of Translate "yes" or "no" 
information, overrides, and defaults. Sure, it would be nice to get 
XPath for global selection, but I understand your argument stated above 
and think these rough compatibility points are mostly important.

Leif Halvard Silli wrote:
Since I first proposed to extend the language tags with a subtag
giving translation relevant information, I have become convinced
that we should not use @lang and/or the language tag to give
translation *commands*. The purpose should only be to used them as
selectors. Thus, if you wrote <em lang="de-q-notrans">Prost!</em>,
then you have, in principle, not given the command that this text
should not be translated.
 >FS: I have never seen a separation into "commands" and "selection" for 
"Translate" information, so I doubt that such behavior would be 
supported by localization tools working also with many other, non-HTML 
and non-XML formats.

Karl Dubost wrote:
I don't know any W3C Note or W3C Specs recommending implementation 
strategies for automatic translations.
 >FS: AFAIK the usage of markup like <code> is currently too irregular 
to come up with implementation strategies what to do about <code> etc. 
In a sense this is like a chicken and egg problem: if there would be a 
"Translate" attribute in HTML and people would use it and other markup 
more coherently, they might get better results from automatic 
Translation tools, and these might change their strategies etc.

Felix

Received on Monday, 4 August 2008 06:41:52 UTC