- From: Al Gilman <asgilman@access.digex.net>
- Date: Mon, 15 Jun 1998 10:05:57 -0400 (EDT)
- To: w3c-wai-au@w3.org
- Cc: w3c-wai-ua@w3.org
to follow up on what David Poehlman said: > isn't the title downloaded anyway along with the rest of the > source? so all we are doing is reformatting what's already > there? It is a question of when you GET the page that the link leads to. The idea is that if there is an un-explained link in page A which leads to page B, then why not pre-fetch page B and extract its TITLE which would then be displayed as the link text in page A where there is a link to page B. Maybe the best thing to do with this technique is to experiment with it under the WAI-ERT umbrella. It may be that it is a reasonable source of ALT text suggestions for authors. Or it may turn out that authors who insert image links without ALT, link to authors who post pages without TITLE. There are options in HTTP where you can get just the top n bytes of a page, so we don't absolutely have to pre-fetch the whole page. But the text of the page is not the bandwidth problem, anyway. For Author prompting, if there is a META SUMMARY tag which has been inserted to make search engine hits format cleanly, that is golden and the SUMMARY should be used as a clue if the author of the referring page can't quickly come up with link text. But simply a tiled over/under display of the "from" and "to" points of the link (positioned viewframe selected from HTML document) is what may make the most sense as an aid in authoring. For authoring, what is critical is to communicate the link [ALT] text question to the author as an issue of continuity. Everybody has had their composition teacher tell them they need to scrub up their continuity at some time in their life. Hope this helps. I haven't reviewed the full context. Al
Received on Monday, 15 June 1998 10:05:44 UTC