W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > April to June 1999

RE: PRI-9 LANG Attribute

From: Henk Wittingen <hwitting@inter.nl.net>
Date: Wed, 28 Apr 1999 10:16:29 +0200
Message-Id: <199904280816.EAA18139@www10.w3.org>
To: W3C-WAI-Markup <w3c-wai-gl@w3.org>
After reading the whole thread about the LANG-attribute I'm getting puzzled 
more and more. So please clear up my three main questions below.

1. Is there a clear advice not to use more than one language in a page?
2. Does the working group still think the LANG-attribute is redundant in 
relation to the characterset code?
3. Does the workinggroup think the LANG-attribute is important for 

Related to question number 1 it is wise to take in account that several 
European countries have more than one official language. Switzerland has four 
languages, Belgium has three and Finland uses Swedish beside the Finnish. This 
is only a limited example coming up to my mind when writing.
As Jaap pointed out the HTML standard supports the mix of several languages in 
one page. I think that web-content can be compared to a multi-lingual Tourist 
Information flyer in many cases. In the margins one can sometimes find little 
flags or country abbreviations (in graphics, no ALT of course!) indicating the 
language used in a part of the text. For selective reading in a visual world 
that can be of help, but it is far insufficient when using screenreaders and 
voice output.

Related to question number 2 it was Nir who made quiet clear that 
LANG-attribute and characterset are related mainly in one direction. To write 
a certain language properly, the characterset must support all glyphs used in 
that language. But many charactersets give multi-lingual support. UTF-8 is the 
best example for that while it supports English, Swedish, French, Greek, even 
Hebrew and much more. A page can contain more than one language but only one 
Let us not forget that making Greek or Russian names more readable in an 
English text they are many times written using the latin script although the 
original writing would be using the specific Greek or Russian characterset. 
Even though it would be much better if a screenreader could speak a Greek name 
in the Greek way and not the English way.
The LANG-attribute (in combination with the characterset information) is the 
only means defined yet to facilitate this in an acceptable way.

Related to number 3 can be added that the fact that nobody uses the yet 
defined standards is not the right (not saying: the wrong!) argument to lower 
a priority. Perhaps the conclusion could be that it is an Authoring Tool or an 
Education and Outreach topic. The whole problem is not mainly a matter of 
technique but more a matter of understanding the problems.

So please do not push this very important topic aside using the American way 
of global thinking that all information is in English. While on the other hand 
giving the impression that other languages don't carry information that need 
to be accessible.
I don't blame anyone for not knowing that many countries use more than one 
language. But take the advise and listen carefully, get well informed and 
think over all related facts before making any decision on this topic.

met vriendelijke groet / with kind regards,
Henk Wittingen - Ede gld - NL

  henk.wittingen@inter.NL.net (thuis/private)
  henk.wittingen@mail.ing.nl (werk/office)
  +31654605805@gin.nl (GSM-SMS/Subject only!)
Received on Wednesday, 28 April 1999 04:16:09 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 16 January 2018 15:33:29 UTC