RE: Language declaration tests

> I had a look at this. Is this one test, or more than one test?
> "This is one of a set of tests" says 'one test', then "Tests 
> on this page"
> says 'more than one'.

I'll clarify.

> 
> I like "conclusions must be drawn with care"; maybe that 
> should be in <strong>.
> 
> I think it's a bad idea to do <html lang="ko"> for a page 
> that is almost all in English. If that's really necessary for 
> testing purposes (which I'm not sure it is in this case), 

Well we need to use something where font is changed automatically for given
languages.  CJK ideographic text is most likely to exhibit this behaviour
and the ideographs show the differences well. The Chinese varies in
application between SC and TC, and Japanese is sometimes used as the
default, so Korean is a good choice.


> then at least make sure that all the English stuff has 
> lang='en' on the relevant subelements. [in short, this is 
> called 'eat your own dogfood]

OK

> 
> It was unclear to me why only one or two lines were green, 
> rather than the whole doc, even if the whole doc is declared Korean.
> I'm not at all sure that it makes sense to make the tests 
> that complicated.

Because the other text has explicit style properties assigned to it.  It
helps people looking at a number of tests to keep the surrounding text the
same. It also helps focus attention on the part that is being tested.  For
further information, read the CSS test guidelines.

> 
> The ideograph labeled 'ko' turned out green, as well as the 
> non-labeled ideograph at the end. Looking at "If there is 
> variation in the above characters, and language information 
> is picked up for this feature, the following ideograph should 
> use the Korean font.", I'm at a loss to understand whether 
> the test was successful or not.

I'll add a note about this possible side effect, but the black text tells
you what to look for to understand whether the test is successful.

> A general reader, not being an expert on font/glyph 
> differences for ideographs, might say: yes, Korean uses a 
> green font, so the test is successful. As it turns out, 
> language info is indeed picked up, but not in the way the 
> test was specifically addressed to test. This points out that 
> rather than having two tests, only one test should be enough; 
> if either the color, or the glyph variant, or both, are 
> different, then the test is successful.

Read the telecon minutes for why we realised that that approach doesn't
work.

> 
> "Each line in the box is marked up as <p lang="xxx" xml:lang="xxx"
>  >雪</p> where 'xxx' is shown after the Han character."
> Why not just say "Each line in the box is marked up as <p lang="xxx" 
> xml:lang="xxx">雪 xxx</p>."? Simple and easy.

Because you don't know how it's marked up otherwise. And that is crucial to
the understanding of the test. (You need to identify the Korean one).

RI
> 
> 
> Regards,     Martin.
> 
> 
>  >Results are at http://w3.org/International/tests/results/lang-decl
>  >
>  >Deborah, I introduced some differences in style throughout, 
> and added an  >interesting test about the 2nd value in a 
> multiple list.
>  >
>  >After you guys have had a chance to comment, we should send 
> this out for  >wider review on www-international.
>  >
>  >We now need to repeat the tests on base UA versions and Mac Uas.
>  >
>  >RI
>  >
>  >============
>  >Richard Ishida
>  >W3C
>  >
>  >contact info:
>  >http://www.w3.org/People/Ishida/
>  >
>  >W3C Internationalization:
>  >http://www.w3.org/International/
>  >
>  >Publication blog:
>  >http://people.w3.org/rishida/blog/
>  >
>  > 
> 
>

Received on Tuesday, 19 October 2004 08:45:07 UTC