- From: David Woolley <david@djwhome.demon.co.uk>
- Date: Sun, 29 Oct 2000 10:08:38 +0000 (GMT)
- To: w3c-wai-ig@w3.org
- Cc: S.Vassallo@e-bility.com
> I have been surfing the net and came across an interesting site using > this > feature. If you want to have a look the address is: > http://www.betterhealthchannel.vic.gov.au/ (you will need to follow the > welcome link on the home page to view the drop down box with a list of > links to other key pages - found in the bottom left hand corner of the > screen). This wasn't what I thought it was. It's not a popup but a selection box that forwards on detecting an onchange event. They have done a much better job of a non-script fallback for this than I've seen done before - it works cleanly without script support. I have reservations, though, about this sort of technique. I think one should use links for navigations, so that people don't have to guess what to do (text only, there was no obvious label on this control). About the only advantage they have is that you can get everything on one screen. Most people copying the idiom will fail to realise that a fallback is needed and many will not have access to CGI to enable them to provide that fallback. You cannot use this as the only form of navigation if you want to be indexed by search engines, as they will not attempt to fill in forms. In passing, I noted that the site's multi-language support is broken. One should always use HTTP to specify character setsl if at all possible, and any character set specified in the HTTP takes precedence over any in the text (the text may have been transliterated by the server). Their Lotus server is explicitly specifying ISO 8859/1, so, for example, any conforming browser will ignore the BIG5 in the body of the Chinese (I can't read it, but it seemed a good test case). General abuse of character sets may mean that people in BIG5 countries use browsers that assume or are configured to always use BIG5 and ignore all sources of character sets, but Lynx definitely obeys the standard, in its default configuration, and so will recent IEs when configured for automatic detection. From an accessibility (wide definition point of view) the language labels should have been in the language in question (there are some issues with font and encoding for non-CJK languages, when taking real browsers into account) and the site ought to have done contents negotiation (like Google does++) to automatically serve a home page in the user's preferred language. > forms. The forms I have designed to date use tables for layout and a > submit button - I have included field names and a tab index to make it > easier to navigate - is this sufficient? It seemed to read sensibly text only, however, it is seriously abusing blockquote elements for some presumed presentational effect they have on whatever browser was used. The only legitmate structural markups in this context would have been using TH instead of TD (theoretically this allows a speech browser to associate the name with the data cell, but I suspect that none of the table accessibility features in the HTML spec have been implemented - you should read them though) and LABEL, to eplicitly associate the lable with the control. ++ At least for English, French, German and Spanish.
Received on Sunday, 29 October 2000 05:50:09 UTC