Re: hreflang

On Sat, 4 Feb 2006, Manuel Strehl wrote:

> I think it's kind of a vicious circle.

In a sense it is, but I think you missed the point in my message that you 
responded to (and quoted). Considering just the authoring side and the 
user side, it _is_ a vicious circle with no way out. The solution, if any, 
is in the software, in user agents.

If you install the IE 7 beta preview, you will enter "once-only" page 
(you can actually return to it), which asks some relatively irrelevant 
questions before letting the user start browsing. Instead of asking the 
user questions that may well be asked later or can be omitted (by 
defaulting settings and letting users change them if they wish), browsers 
should ask _essential_ questions, such as preferred default font size 
(offering a set of reasonable alternatives) - and language preferences.

As long as there is no move in that direction, language negotiation 
remains a specialty with marginal use.

(And the hreflang attribute probably always remains useless. Personally, I 
use it sometimes - as a documentation for myself mostly. But it's actually 
worse than useless when it is incorrect. And the destination of a link 
_may_ change language - perhaps switching to a language-negotiated 
version - without notifying people who link to it.)

-- 
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/

Received on Sunday, 5 February 2006 08:42:38 UTC