W3C home > Mailing lists > Public > www-html@w3.org > February 2006

Re: hreflang

From: Henri Sivonen <hsivonen@iki.fi>
Date: Wed, 8 Feb 2006 22:32:46 +0200
Message-Id: <57248617-A8A8-4495-9193-A62245392938@iki.fi>
To: W3C HTML <www-html@w3.org>

On Feb 4, 2006, at 19:38, Jukka K. Korpela wrote:

> Consider, for example, the fact that there is nothing resembling  
> language negotiation on the European Union site

> However, as Mikko mentioned, the real problem is that user agents  
> do not send adequate information.

I think an even more real a problem is that users can rarely trust in  
the translations being up to date. Additionally, language negotiation  
may hide translations from search engines.

In fact, language negotiation depends on the premise that the  
translations are in sync *and* the user trusts them to be in sync. I  
think the premise is faulty. How often does the premise actually hold  
except for frozen EU documents? Even on the EU sites I tend to check  
out things in two or three languages just in case. How can I trust  
that eg. debian.org is up to date in Finnish? If I set my browser to  
prefer English, I can reasonable assume that I am seeing the latest  

> That's not the point. Many people use English-language browsers and  
> systems for several reasons - for example, because a browser of  
> their choice exists only in an English version, or because its  
> localization is awful (wrong translations, etc.).

Actually, the case where I think language negotiation actually works  
is for Web apps. Their UIs are not indexed by search engines anyway,  
so the search engine interaction is not a problem. Also, UI string  
bundles can be piece-wise translated so that if a translation is  
stale, the localization system can fall back to English for  
individual strings without preventing access to the latest version of  
the app.

For this purpose Safari's way of reflecting the OS X UI language  
preference order in Accept-Language is exactly the right thing to do.  
(It may not be for prose.)

And even then there's the problem with Internet cafés.

> Besides, even if the language of your browser happens to be your  
> native language, what about all the _other_ languages you might  
> know? In the WWW context, even languages you know just a little are  
> important in the preferences.


It appears that thoroughly expressing one's language preferences is  
hard even if one is aware about language negotiation. Considering  
your stated language skills[1], I would expect you to choose Danish  
if presented with Danish and Chinese language alternatives. (I am  
assuming that your Swedish parser is not Draconian. :-) Yet, your  
Accept-Language does not advertise Danish (or Norwegian).

Given the choices, a person can pick a language by following a link.  
However, it is likely that the person will not encode all the  
relevant data in the Accept-Language header even if the person is  
well aware of the header.

Considering how language negotiation is a permathread and how people  
feel the need to come up with things like XHTML 2.0 hreflang, perhaps  
language negotiation for content isn't such a good idea after all. It  
seems to me that all the architecture astronautics are not worth the  
trouble compared to plain links to search engine-friendly language- 
specific URLs.

[1] http://www.cs.tut.fi/~jkorpela/languages.html

Henri Sivonen
Received on Wednesday, 8 February 2006 20:32:55 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:06:12 UTC