Hello Marcos,
Looking through step 6, I think I see an error in the second example.
For example, if the subtag is |"*-Ch"| and folders included "de-cH",
"fr-Ch" , and "IT-CH", then the folder "de-cH" would be matched, then
base URI would be
|"widget://f81d4fae-7dec-11d0-a765-00a0c91e6bf6/en-cH"|and the widget
locale <http://dev.w3.org/2006/waf/widgets/#widget11> would be "|en-ch|".
I could understand if the locale was any of "de-cH", "fr-Ch" , and
"IT-CH", to match *-ch, but en-ch is not mentioned.
Also, where multiple possible matches occur, as in the example above, is
there a deterministic way defined for which will be used?
This becomes relevant, if for example, each locale folder had separate
distinct resources.
e.g.
base URI contains the following graphic resources for the widget
a.gif
b.gif
c.gif
d.gif
de-cH contains the following graphical resources
a.gif
b.gif
fr-Ch contains
b.gif
c.gif
and
IT-CH contains
a.gif
c.gif
Which of these should happen for *-ch assuming de-cH is chosen as the
widget locale?
1) c.gif is requested
a) the base c.gif is delivered, as de-cH does not contain a localised
version of c.gif
b) c.gif is delivered from fr-CH and used as this is a better match to
*-ch descriptor than the base.
How should this be altered if IT-CH were to be chosen as the widget
locale and b.gif were requested?
Marcos Caceres wrote:
Hi Addison, All,
I've attempted to make the widget spec conformant with BCP437. In
particular, I've deferred searching for localized folders to RFC4647
using "lookup" against a language priority list composed of a basic
language range. WAF would be grateful if (i18n) people could take a
look at the current draft in the spec and make sure that it is ok (so
we can close issue 23 [1]). The i18n text is fairly short (only about
1 page), so it should not take much time to read, and, hopefully,
respond to.
In the widget spec, the localization definitions and author
requirements are described in section:
http://dev.w3.org/2006/waf/widgets/#localization
The i18n processing model (which makes use of RFC4647) is described in section:
http://dev.w3.org/2006/waf/widgets/#step-6
Hope it all makes sense.
Looking forward to hearing your feedback,
Marcos
[1] http://www.w3.org/2005/06/tracker/waf/issues/23
----
David Clarke
