Re: Summaries of issues around checkpoints 1.4 and 1.5

At 02:32 PM 2004-02-10, Joe Clark wrote:
>You simply have to declare your character encoding
>"legacy" encodings work better in real-world browsers

On the one hand, the clarification suggested by Richard, to wit

<quote cite=
>All text can be decoded into words, representable using Unicode characters.

would make it clearer (it's already true with the existing language) that duly
annotated use of Big5 etc. conforms.

On the other hand, what we should really do concerning Unicode, encodings,
etc. is, as Joe hints, defer to the Character Model; with a note-in-draft to
confirm it gets to Rec before we do.  We can remind people to observe it,
but should not abridge it in text that comes across as free-standing
requirements imposed by *this* document.

This requirement is important for access by PWDs but it's not our problem
alone and there is a document chartered elsewhere (i18n) to address exactly
this topic.

So if we want to claim credit for "plays well with others" we had better
avoid re-inventing their wheel.


> >     "All text can be decoded into words represented in Unicode."
>This isn't an accessibility issue in the first place. You simply have to
>declare your character encoding (either in HTTP headers or the <meta>
>element or both). Unicode, while desirable, is not the only encoding that
>works, and for some languages (like Vietnames, Thai, Chinese), the
>"legacy" encodings work better in real-world browsers.
>Could proponents of the current wording please point to people with
>disabilities alive and using the Web today who experience barriers or
>inaccessibility because some encoding other than Unicode was used?
>   Joe Clark  |
>   Author, _Building Accessible Websites_
>   <> | <>

Received on Tuesday, 10 February 2004 16:26:23 UTC