E-mail List Archives

RE: Bobby WCAG checking

for

From: Jukka Korpela
Date: Oct 25, 2002 3:03AM


Freda Lockert wrote:

[quoting Bill Mason]
> >Bobby is looking for language identification in the HTML tag, e.g.:
> ><.html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en-US"
> lang="en-US">
>
> Bill, Joe, Emma, thank you all for the responses. The page passes all
> the checks now, and BBEdit has changed the other 42 pages for me.

Fine, the use xml:lang and lang attributes in the <html> tag is
recommendable. But it's actually just a part of using language markup, and
mostly the least significant part in practice. It may help e.g. a screen
reader read the text of a document properly. For example, IBM Home Page
Reader knows a few languages, and if the language is not indicated in the
document using the lang attribute, then it applies the rules of whatever
language has been selected the default in its settings - and I can assure
you that English text sounds horrible if read by the rules of the Finnish
language! But this isn't really such a big issue, since after hearing the
first few words spoken oddly, the user can change the setting of Home Page
Reader. (It's still odd that e.g. the WAI main page violates WAI rules by
not indicating the language.)

(I might state the obvious here, but I still wish to point out that "en-US"
means English as spoken and written in the United States, so-called American
English. For British English, "en-GB" should be used. You might also use
just "en", for English generically.)

More importantly, language _changes_ inside a document should be indicated,
since the user cannot (in practice) modify the settings on the fly in the
midst of a document. Thus, if you have French words or phrases in a
document, you should use lang markup for them, e.g.
<span lang="fr" xml:lang="fr">Loire</span>

That means a lot more work, and such markup helps a small amount of users
only, since very few user agents pay any attention to lang markup; however,
when it helps, it helps considerably, if the page contains text in more than
one language. As a practical move, one might use lang markup just for
_large_ pieces of texts, like sentences quoted in their original language,
book titles in "foreign" language, etc.

Anyway, it is an explicit and requirement in WCAG 1.0 that changes in the
natural language of a document's text be identified. There are no
reservations, so apparently _all_ changes should be indicated. Ref.:
http://www.w3.org/TR/WCAG10/#gl-abbreviated-and-foreign
And in practice the lang and xml:lang attributes are currently the only way
of doing that. (It even says that the language changes be identified for all
text equivalents, too; but we _cannot_ indicate e.g. that language changes
inside an alt attribute value.)

I wonder how many people have put an icon of WCAG 1.0 conformance onto their
pages without realizing that their page probably violates the Priority 1
requirement on indicating language changes. Of course, checkers like Bobby
don't check this checkpoint at all. (Theoretically, checkers _could_ do
that. In fact, the language of text can be reasonably well decided
automatically using statistical and other tests - if the text is
sufficiently long.)

--
Jukka Korpela, senior adviser
TIEKE Finnish Information Society Development Centre
http://www.tieke.fi/
Diffuse Business Guide to Web Accessibility and Design for All:
http://www.diffuse.org/accessibility.html


----
To subscribe, unsubscribe, or view list archives,
visit http://www.webaim.org/discussion/