Main

• Home
• Post Atom Feed
• Comment Atom Feed
• Blogger Profile
• Contact
• About

Latest

• Backwards Compatiblity is NOT the Key
• I'm Back
• MSN Search
• GROSS User Group Meeting 2005-02-01
• 2005 South African Blog Awards
• Google's Country Search Analyzed
• Weblog Awards
• Online Advertising
• Media-specific @import
• Delicious and PEAR

Archives

• June 2004
• July 2004
• August 2004
• September 2004
• October 2004
• November 2004
• December 2004
• January 2005
• February 2005
• March 2005
• April 2005
• May 2005
• June 2005
• July 2005
• August 2005
• September 2005
• October 2005
• November 2005
• December 2005
• January 2006
• February 2006
• March 2006
• April 2006
• May 2006
• June 2006
• July 2006
• August 2006
• September 2006
• November 2006
• December 2006
• January 2007
• February 2007
• March 2007
• April 2007
• May 2007
• June 2007
• July 2007
• August 2007
• September 2007
• October 2007
• November 2007
• December 2007
• January 2008
• February 2008
• March 2008
• April 2008
• May 2008
• June 2008
• July 2008
• August 2008
• September 2008
• October 2008

The 'lang' Attribute

Since we're focusing on accessibility now (and rightfully so) I thought I would kick off with an issue that has long bothered me: multilingual text.

I find that one of the most practical, yet heavily underused attributes is the lang attribute (and by that I also include xml:lang obviously). Yes, many people like me have been using it for years now, but far too few IMHO. There has been far too little said about this issue in recent years, and I feel it's about time we should be starting to make people aware of this, why we need it and how to use it properly.

Marking up languages and using those attributes don't only apply to the root element and specifying the base language of a document (although that is of course also very important, especially for search engines). Marking up languages specifically and thoroughly is very important for speech-enabled UAs to pronounce the text correctly.

For example, let's say I want to throw a bit of common French: Au Revoir! Let's say an aural UA reads this out to a person with limited visibility. Just imagine if that French was pronounced as English! That wouldn't sound correct at all, would it? Even people (like me) with very limited skills in French would think it sounds rather weird! That's why it needs to be marked up appropriately as fr so that it can be pronounced correctly by a UA with a French engine (most should have one on a default install - French is a major western language).

Ideally speaking, you should mark up absolutely every single piece of text as the language it's written in accurately. This sounds pretty straight-forward, and usually it is, but sometimes it can also become quite tricky.

Some languages, like Dutch for example, have certain words that are taken from English and then adjusted somewhat to become Dutch. For example, the word downloaden (which means "download", naturally). Semantically, this is sort-of adjusted and adopted from English to Dutch, thereby making it Dutch. In other words, if the rest of the document is in Dutch you shouldn't really have to bother marking this up as another language.

When this word is pronounced by a computer, it will probably be pronounced incorrectly, because I strongly doubt that the "w" would be pronounced by it as it is normally pronounced in Dutch. Some people will therefore be tempted to mark it up as English. However, this is still incorrect from a semantic point of view. Downloaden shouldn't be pronounced in an English voice, but rather in a typical Dutch voice, otherwise it would sound completely unnatural in a Dutch piece of text.

Most languages on earth will have at least some words (some languages more than others) that are pronounced a bit funny. Therefore, it's the job of the creators of the relevant aural engine to cater for the various needs of their respective language. We, as web developers, can only aid them so far.

One of the facts that people often ignore (or want to ignore) is that proper nouns ("names of places or things" for those that don't speak geek) should also be marked up according to language.

For example, my name (pronounced in a typical English fashion) might not sound funny (well, probably it does, but they're expecting it to sound funny) to an American that doesn't know anything about South African names in the first place. However, for the average South African, to listen to my name being pronounced in English would sound totally ridiculous and often incomprehensible to them. They would expect it to be pronounced as in Afrikaans.

Therefore, you would like to have the average South African using an aural UA that might have an Afrikaans engine installed hear my name pronounced as it should be. Every time you mention my name or this weblog's title, you should preferably mark it up in the applicable language if possible. For example:

• <span lang="af">Charl van Niekerk</span> (for HTML)
• <span xml:lang="af">Charl van Niekerk</span> (for XHTML)
• <cite lang="af">Charl van Niekerk</cite> (for direct citations)
• <cite><span lang="af">Charl van Niekerk</span>'s Weblog</cite> (for citations from this weblog)

Note that I'm not listing every possible scenario and combination since that would just be a waste of time. You're not a total idiot, are you? ;-)

Marking up languages for names is indeed often very difficult. For example, when you want to mark up the language of the name of the author of a post or a comment. This will essentially mean that you need an extra textbox on your comment form to allow your users to type their name's language code.

However, sometimes you get people with their full name containing more than one language. For example, take a Russian woman marrying a Greek man. That woman will then have a Russian first name and a Greek surname. This essentially means that you'll have to start allowing span elements with lang/xml:lang attributes into the name field. This could get a little tricky, especially if you need to insert that name into the value of an attribute (for example, the title attribute of a rel element). Therefore, you'll just have to strip the markup in those.

Anyway, there are workarounds for everything, and we'll just have to live with the limitations. For example, you're out of luck in attributes and also in poorly-designed elements like the title element which doesn't really allow you to mark up anything properly. For example, no span elements are allowed, which means you'll have to mark the entire title up in one element. This is troublesome for titles as on this weblog, which I would have liked to mark up like this:

<title><span lang="af">Charl van Niekerk</span>'s Weblog: Some Post Title</title>

Very stupid not to be able to do this IMHO! But luckily, this will all change with XHTML 2.0 (thanks, Lachlan) since apparently in the current working draft at time of writing, span elements are allowed into the title element!

Yeah, this indeed goes to a whole new level of markup purism and software requirements... Who ever said life is going to be simple? :-)

Posted by Charl van Niekerk on Tuesday, February 15, 2005
Get TimesURL