Abbreviations and language
I'm using <abbr> and <acronym> to mark up ... well, abbreviations and acronyms, of course. 
The WCAG guidelines also state that we should indicate changes in natural language, which I also do. (I've caught myself on the verge of marking up "Lorem ipsum" as latin in my wire frames! LOL)
However, when it comes to the aforementioned tags, there's something of a problem. Let's say I'm writing about web design in my native language (Swedish). I want to mark up the first occurrence of the term CSS with a title attribute. Now, the title (Cascading Style Sheets) is in English, but I would like a screen reader to read the actual abbreviation (initialism, really) in Swedish.
This would force me to write kludgy markup like this:
| Code: |
| <abbr title="Cascading Style Sheets" xml:lang="en"><span xml:lang="sv">CSS</span></abbr> |
Apart from being messy and causing code bloat, there's another small issue: Opera won't behave the way I expect it to. If I do it like this, Opera 7.11/Win won't display the abbr's title when the mouse pointer hovers over it. Mozilla does, and I think that's correct behavior. (IE doesn't even know what <abbr> is, so forget about that.) I don't worry too much about non-compliant browsers, but in this case I'm not quite sure which behavior is actually correct.
What do you think about this? Am I overly pedantic if I mark up abbreviations and acronyms like this? Is there a better solution?
/Tommy
Tommy has left the building
Share: 
I know it's not a lot of help, but I recall a post on Minz Meyer's blog about this issue:
http://minzweb.de/en/pages/sitetopics/archive_permalink.asp?id=27
Michael
Share:
Thanks for the link.
Like you said, it doesn't actually help, but at least I know that I'm not alone in wondering about this.
What I do at the moment is this:
| Code: |
| <abbr title="Cascading Style Sheets" xml:lang="en" lang="en">CSS</abbr> |
I don't bother with the <span> to make the 'CSS' part be read in Swedish, but I specify both 'xml:lang' and 'lang' to be on the safe side. The disadvantage, of course, is that a screen reader will read the initialisim in English, but that will have to do.
I agree with the author you referenced: there should be a 'titlelang' attribute for this.
Since the interest for this matter seems to be virtually non-existent (judging from the lack of replies here), I guess I won't put too much of an effort into this.
/Tommy
Tommy has left the building
Share:
| Quote: |
| Since the interest for this matter seems to be virtually non-existent (judging from the lack of replies here), I guess I won't put too much of an effort into this. |
I can't speak for others but thought I'd mention that I've been watching this thread with interest but as I don't have a solution I haven't posted.
Have you considered testing this out in some screen readers to see how they handle the different options?
Cheers,
Nigel
Accessify Forum Administrator ~ Nigel Peck / MIS Web Design
"Everything I say is not meant to be set in stone" - Van Morrison
Share:
Nigel - I thought that was the case, since there has at least been a number of views. The passage you quoted was intended to be kind of tongue-in-cheek.
I have never tried a screen reader myself. After reading some of the heated discussions that followed Dave Shea's well-intended plea for distributing JAWS to all web developers, I realize that this isn't just something you can install and run. You need to use it for some time - with your monitor switched off - before you can begin to understand what it is like for a person with limited or no vision to access the web.
Using the 'Text Browser' style sheet that comes with Opera 7.11/Win gives me some idea of what a screen reader might 'see' when it encounters one of my pages. It does indicate changes in natural language. OTOH, I know nothing about the level of support in, say, JAWS for the 'lang' and 'xml:lang' attributes. Besides, won't a user have to have multiple language modules (or something) installed for this to work?
The site I'm working on is purely personal, and I don't really expect to have (m)any visitors using assistive devices. However, since I'm interested in web design, web standards, and accessibility, I thought that building my own site as compliant as possible would make for good practice. (I'm a software developer, although I don't usually create web applications at the present.)
The current version of my site is everything but accessible, but I think I've learned from my mistakes there, so the new version will be a lot better. I hope to learn a lot from this forum.
Tommy has left the building
Share:
I see the quandary though does both the abbreviation CSS, and title have to be written in English if it were a word like "Jagermeister" (roughly translates to Huntmaster), which is a German word then yes you would use those conventions for the word; though not necessarily for the title attribute.
Since CSS would be pronounced C.S.S. furthermore it would be of little use to a Swedish person who couldn't understand English, though from experience I know many are very fluent in English.
Though interestingly enough that leads to some more intriguing hypothetical scenarios
};-) http://www.xhtmlcoder.com/
WVYFC chose the Yorkshire Air Ambulance as the main charity to fund raise for in 2006
Share:
CSS is an abbreviation for Cascading Style Sheets (we all know that of course). Without a doubt, the phrase could be translated into Swedish and any other language but then the abbreviation would likely not be CSS but something else. My point is, CSS is an abbreviation for an English word so the title should have the English words.
In the same way that we must use the American spelling of "color" and "center" in HTML, CSS is a technical term.
Just my 2 cents,
Jules
Share:
| Jules wrote: |
| My point is, CSS is an abbreviation for an English word so the title should have the English words. |
I think he's happy with that. It's just that he doesn't want software that would be capable of vocalizing the text to take that as a cue to say: See-Ess-Ess.
I guess many words have naturalized pronunciations in different languages; and while one might (for whatever reason) want to indicate the native language for the word, one mightn't want the word read out using the sound-system of the original language when the word is read in a passage which as a whole is written in another language.
That might sound weird or confuse people.
Examples in English would be "Caesar" or "Don Quixote" both of which have naturalized English pronunciations ("See-zer", "Don Kwikz't").
I can see that these abbreviations/acronymns from English could cause this problem. Swedes spell it CSS, but they obviously don't say "See-Ess-Ess".
Michael
Share:
| Michael wrote: |
| I think he's happy with that. It's just that he doesn't want software that would be capable of vocalizing the text to take that as a cue to say: See-Ess-Ess. |
I appologise, I completely missed on that point.
Jules
Share:
Hi Folks,
I would like to point aou an other matter related to this discussion.
We shall also consider that most of the UA do not have the chance to render properly all the languages because of the lack of dictionaries.
I am just experiencing this in Hungarian, which can't be considered a main worldwide language. in Hungarian pages there are a lot of English words, and this is pretty easy to be rendered by UA as far as English dictionary are provided with most of the UA.
But what about the Hungarian websites' pages provided also in English with many Hungarian definitions and words?
just a few percentage of the world population could understand the proper pronunciation of the definition.
But who could understand the real meaning of these Hungarian citations if the webmaster have "forgot" to explain it clearly in English beside the original version?
Could we take into consideration the possibility to insert something like the title element which could provide also semantic meaning to the foreign language abbreviations and acromyms proper rendering, so that information is clearer and even more accessible?
Any experience about that?
M2p
pat
Share:
| Pat wrote: |
| Hi Folks,
I would like to point aou an other matter related to this discussion. We shall also consider that most of the UA do not have the chance to render properly all the languages because of the lack of dictionaries. I am just experiencing this in Hungarian, which can't be considered a main worldwide language. in Hungarian pages there are a lot of English words, and this is pretty easy to be rendered by UA as far as English dictionary are provided with most of the UA. But what about the Hungarian websites' pages provided also in English with many Hungarian definitions and words? just a few percentage of the world population could understand the proper pronunciation of the definition. But who could understand the real meaning of these Hungarian citations if the webmaster have "forgot" to explain it clearly in English beside the original version? Could we take into consideration the possibility to insert something like the title element which could provide also semantic meaning to the foreign language abbreviations and acromyms proper rendering, so that information is clearer and even more accessible? Any experience about that? M2p pat |
Michael
Share:
| Pat wrote: |
| We shall also consider that most of the UA do not have the chance to render properly all the languages because of the lack of dictionaries |
Perhaps I'm misunderstanding this, Pat, but it seems to me that most of the rendering problems revolve around the encoding of the characters. Joel Spolsky has a recent article on this:
http://www.joelonsoftware.com/items/2003/10/10.html
So far as pronunciation goes:
If I want to indicate the pronunciation of words, the everyday practical method open to me is "respelling". Here's an example:
Usage varies across the country, but traditionally, the sequence of letters "L-A-T-H" has three different pronunciations in England:
Lather (foam, as on a sweathing horse): lather
Lath (a thin planed length of wood): laath
Lathe (a turning machine): laythe
But that respelling means nothing to someone who doesn't already speak the language. Instead of respelling, I could use comparison. But it doesn't necessarily help to say:
lather (a as in "man")
lath (a as in "father")
lathe (a as in "fate")
... unless one already knows the other words. And if I want to relate this to another language - well, maybe there is no direct equivalent to the sound in that language ...
The only real solution - and it's not a perfect one - is to use the IPA (International Phonetic Alphabet). But, of course your readers need the fonts installed:
http://www.sil.org/computing/catalog/show_software.asp?id=12
and you have to use unicode entities, and set the character set to utf-8:
http://www.phon.ucl.ac.uk/home/wells/ipa-unicode-test.htm
Michael
Share:
| Michael wrote: |
| I think he's happy with that. It's just that he doesn't want software that would be capable of vocalizing the text to take that as a cue to say: See-Ess-Ess. [---] I can see that these abbreviations/acronymns from English could cause this problem. Swedes spell it CSS, but they obviously don't say "See-Ess-Ess". |
Exactly. In this particular case, the Swedish pronunciation is similar enough to the English one that most listeners would get the idea. However, if we use XHTML as an example, the Swedish and English pronunciations differ more. The English 'aitch' sounds nothing like the Swedish 'haw'.
I think the standard simply isn't complete in this case. The question is how to handle that deficiency.
Tommy has left the building
Share:
<dyslexic thought tangent>
Lettered abbreviations are rarely language specific; they are only individual characters themselves, relational to the alphanumerical writing system used.
Acronyms however, are a different matter, as would be the expanded form of the abbreviation as written in full. Again, the only major issue occurs when the entity tries to interpret the abbreviation string as a free word.
XYZZY...
</dyslexic thought tangent>
};-) http://www.xhtmlcoder.com/
WVYFC chose the Yorkshire Air Ambulance as the main charity to fund raise for in 2006
Share:
Well, not quite. An abbreviation, or, rather, an initialism, is read letter by letter. The names of letters differ between languages. The letter 'H' is pronounced very differently in Swedish and English, for instance.
This doesn't matter at all in a visual environment, but it does matter in a screen reader or speech browser.
Tommy has left the building
Share: