- From: Laurens Holst <lholst@students.cs.uu.nl>
- Date: Sun, 05 Feb 2006 16:18:37 +0100
- To: W3C HTML List <www-html@w3.org>
Oh, I came up with one additional use case, which is why I think the mechanism should not be removed entirely: Assume I have a site in both English and Dutch. Which of those is served to the user depends on the Accept-Language header. On the English version I have a link to the Dutch version, and vice versa. The link would contain /?lang=nl. If the user clicks on that link because he would contrary to his language preference settings rather like to read the Dutch version, all the internal links in the Dutch version would also have to contain ?lang=nl, or whenever the user clicks on another link it will switch back to English. I see two ways to avoid this. One is by setting a session cookie, which is then checked before the Accept-Language header in determining the language. This is probably the easiest and often-used method right now. The other way would be to add ?lang=nl to each of the Dutch version’s links (except for the one linking to the English version), that would work, however it isn’t exactly pretty; in that case acceptlang="nl" would both be prettier as it doesn’t ‘pollute’ the link with language information, *and* not require an additional server mechanism to check the ?lang GET parameter. So that’s why having an @acceptlang attribute (as @hreflang is currently specified in XHTML2) does have some value, when used to link to internal pages. In such a scenario, Oskar’s C.1., C.2. and C.4. concerns wouldn’t apply. (And I disagree with concern C.5. :).) Because of that, it might be valuable to mention that @acceptlang should primarily be used for links within the same domain, in the given scenario. Question however is whether this use case is worth having such an attribute, with all the possible abuse, or people should just use cookies, which does not require an attribute to be added to every internal link. ~Grauw -- Ushiko-san! Kimi wa doushite, Ushiko-san!! ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Laurens Holst, student, university of Utrecht, the Netherlands. Website: www.grauw.nl. Backbase employee; www.backbase.com.
Received on Sunday, 5 February 2006 15:20:43 UTC