- From: nir dagan <dagan@upf.es>
- Date: Sat Apr 25 23:43:13 1998
- To: jasonw@ariel.ucs.unimelb.EDU.AU
- Cc: w3c-wai-gl@sahara.upf.es
The markup used to move the user from page to page against his will is actually valid. It it done by the META element with http-equiv="refresh" and the content attribute defines a delay time and a URL. The syntax is valid, but the spec. does not require User agents to support it, and if they do it doesn't tell how. (it is mentioned in the chapter about HEAD and meta data.) Its a good idea to discourage authors from using it. It is terribly anoying and confusing even for a sighted user, not to mention a user of non visual medium. The guidelines may propose to authors to use server redirect if needed, or to write a regular link referring to the new URL. Since many authors have unfortuntely pages that are _blank_ beyond the "refresh" itself, browsers will continue to support it. Browser guidelines may recommend to allow the user to disable the feature altogether (like in Opera) or to configure the browser to display it as a normal link (even better, to avoid a blank page). Regards, Nir Dagan. > > During the past few months, I have noticed that some HTML documents > include a non-standard feature whereby a new page will be automatically > loaded after a predetermined period of time has elapsed. I can not > remember what the non-standard markup is which gives rise to this > phenomenon, but I am sure that other members of the GL working group would > be familiar with it. The guidelines should discourage authors from using > this feature, not only because it is, presumably, a proprietary HTML > extension, but more importantly, in recognition of the observation that > readers who are using screen access technology or for whom navigation is > difficult may not have time to familiarise themselves with the initial > page before the prescribed time period expires. > > Incidentally, Lynx 2.8 treats this construct as though it were an ordinary > link and does not load the new page automatically, regardless of the > timing which the author has specified. Perhaps this issue should also be > brought to the attention of the user agent guidelines group. > > > >
Received on Saturday, 25 April 1998 23:43:13 UTC