Version reviewed: http://www.w3.org/TR/2006/WD-UNDERSTANDING-WCAG20-20060427/
Lead reviewer and date of initial review: Richard Ishida, Jun 2006
Subject lead in: [UND WCAG2]
These are comments on behalf of the Internationalization Core WG, unless otherwise stated. The "Owner" column indicates who has been assigned the responsibility of tracking discussions on a given comment.
We recommend that responses to the comments in this table use a separate email for each point. This makes it far easier to track threads. Click on the icons in the right-most column to see email discussions.
|1||Understanding Guideline 3.1||Direction of text||
It is not clear to us why correct support of the 'direction of the text' is an accessibility issue. We recommend that you remove all mention of text direction from this document.
Furthermore, it is incorrect to include mention of directionality under a success criterion that talks about programmatical determination of the "primary natural language or languages", since directionality is not necessarily associated with language. Text direction should be it's own success criterion, if it is to stay.
(If you disagree with this recommendation, we will come back to you with a substantial number of additional comments based on the content of this document related to text direction, and probably recommend that i18n WG needs to be involved in drafting that text. For now we will hold all such comments until this one is addressed.)
|2||Understanding Guideline 3.1||Primary natural language||
The term 'primary' is used here in a different way than currently used in i18n documents. We use 'primary language' to mean the intended audience of the document - as specified in the HTTP header or the meta statment. Our equivalent for your term 'primary language' (based on the advice given in your techniques) would be 'default text-processing language'. This is what is expressed via the language attribute(s) on the html element. (Note that we are currently considering options for replacing our use of the term 'primary language' with something less ambiguous and more accurate.)
For a summary of our current usage see http://www.w3.org/TR/i18n-html-tech-lang/#ri20040808.100519373
i18n WG needs to discuss this section with WCAG WG to understand more clearly what the intent is with regard to 'primary' vs. 'text-processing' language, and to help formulate clearer guidelines.
|3||3.1 Additional techniques||dc:lang||
We would be interested in knowing what you will say about use of dc:lang
|4||3.1.1 Example 1||Content in two languages||
"A Web unit produced in Germany includes content in both German and English, but most of the content is in German. The primary natural language is identified as German (de)."
If the primary language is expressed using HTTP or meta tags, it is possible that both languages should be identified if this is a document aimed at a bilingual audience. If the primary language is to be expressed in the html element tag, only one language can be chosen. This example is too vague. This goes back to the question of what WCAG means by 'primary language'.
|5||3.1.2 Related resources||Liam's resources||
The links to Liam Quinn's documentation are not very specific. We are also concerned that the information at the end of these links no longer constitutes best practise - for example, if you look under SPAN there are recommendations to use the <i> tag, rather than <em>, attribute values are given without quotes.
|6||3.1.2 Related resources||No pointer to i18n language decl doc||
There should be a pointer to http://www.w3.org/TR/i18n-html-tech-lang/ when that document is published. Note that it is currently a WD, and that the title has recently changed in the editor's copy to "Internationalization Best Practices: Specifying Language in XHTML & HTML Content".
|7||3.1.3 Idioms||Lang value length||
The markup for the Dutch example says:
<span lang="nl-NL">Hij ging met de kippen op stok</span>
We recommend that you remove the -NL unless you really want to make the point that this is an idiom specific to the Netherlands. In general, language values should be kept as short as possible.
On the other hand, since the W3C uses American English spelling, you may want to change the lang attributes in the html element to "en-US" - which will help for spell-checking, and possibly also for voice browsers?
|8||3.1.3 Idioms||Missing xml:lang||
The markup for the Japanese and Dutch examples should include xml:lang attributes as well as the lang attribute, since this is XHTML served as text/html.
Please check this for any other phrases in non-English text.
|9||3.1.5 Resources||Non-English resources please||
Since this section says that this guideline can be applied to non-English text, it seems strange that there are no references at all to non-English assessment techniques in the Resources section.
Please provide some.
|10||3.1.3 Idioms||Repeated Japanese text||
"Example 3: In Japanese, the phrase "さじを投げる（どうするこ ともでき なくなり、あきらめること" literally translates into "he threw a spoon". But it means that there was nothing he could do and finally he gave up. "
The Japanese original text is confusing, since it contains both "he threw a spoon" *and* the explanation. The latter, that is どうするこ ともでき なくなり、あきらめること , should be deleted.