W3C home > Mailing lists > Public > www-i18n-comments@w3.org > June 2006

RE: [Comment on xml-i18n-bp WD] 2.3 meta vs. non-meta

From: Richard Ishida <ishida@w3.org>
Date: Wed, 7 Jun 2006 10:16:43 +0100
To: "'Elliotte Harold'" <elharo@metalab.unc.edu>, <www-i18n-comments@w3.org>
Message-ID: <00d701c68a13$18949440$6601a8c0@w3cishida>

Hello Elliotte,

This is a personal response, not a response from the Working Group.

> Section 2.3, "Avoid translatable attributes" and "Do not put 
> translatable text in attributes" bothers me a little. There 
> are many valid reasons to put human readable text in 
> attributes that dio not involve translation. For instance,

I think one problem is that the schema designer doesn't know whether a particular document using their format will be translated or not.  They have to design for the general case.


> The inability to have mixed language attributes is perhaps 
> not so big a problem in practice. Attributes don't carry 
> substructure. Anything that's really complex enough to have a 
> lot of substructre tends to go in an element anyway.
> 
> The alt attribute is a really good example. This really 
> should be an attribute for every reason *except* 
> internationalization. Secondly, since img is an empty element 
> it can carry its own xml:lang attribute that applies only to 
> its attribute text.

As it mentions in the text [1], one reason for caution with attribute text is that in bidirectional Arabic or Hebrew documents the alt text may indeed need to carry structure to ensure that punctuation, etc, is rendered in the correct place. For people working in these languages, the fact that attributes don't carry substructure can constitute a major problem.

In addition, multiple language ranges within attribute text can't be labelled.

Also, there doesn't appear to me to be any good reason why replacement text for an image, or text in other attributes, would not benefit from the ability to markup up abbreviations and acronyms like elsewhere, or apply emphasis or special styling to the text. I've often found myself wanting to do that kind of thing.

Hope that helps,
RI


[1] http://www.w3.org/TR/2006/WD-xml-i18n-bp-20060518/#DevAttributes


============
Richard Ishida
Internationalization Lead
W3C (World Wide Web Consortium)

http://www.w3.org/People/Ishida/
http://www.w3.org/International/
http://people.w3.org/rishida/blog/
http://www.flickr.com/photos/ishida/
 

> -----Original Message-----
> From: www-i18n-comments-request@w3.org 
> [mailto:www-i18n-comments-request@w3.org] On Behalf Of Elliotte Harold
> Sent: 23 May 2006 11:09
> To: www-i18n-comments@w3.org
> Subject: [Comment on xml-i18n-bp WD] 2.3 meta vs. non-meta
> 
> 
> Section 2.3, "Avoid translatable attributes" and "Do not put 
> translatable text in attributes" bothers me a little. There 
> are many valid reasons to put human readable text in 
> attributes that dio not involve translation. For instance,
> 
> * In narrative documents, when all markup is stripped what;s 
> left shoudl be a legible plain text document
> 
> * Sometimes meta-information that belongs in an attribute is 
> human readable.
> 
> These don't negate your reasons why you don't want to put 
> human readable text in an attribute; but they are in tension with it.
> 
> The inability to have mixed language attributes is perhaps 
> not so big a problem in practice. Attributes don't carry 
> substructure. Anything that's really complex enough to have a 
> lot of substructre tends to go in an element anyway.
> 
> The alt attribute is a really good example. This really 
> should be an attribute for every reason *except* 
> internationalization. Secondly, since img is an empty element 
> it can carry its own xml:lang attribute that applies only to 
> its attribute text.
> 
> --
> ´╗┐Elliotte Rusty Harold  elharo@metalab.unc.edu XML in a 
> Nutshell 3rd Edition Just Published!
> http://www.cafeconleche.org/books/xian3/
> http://www.amazon.com/exec/obidos/ISBN=0596007647/cafeaulaitA/
ref=nosim
> 
Received on Wednesday, 7 June 2006 09:16:51 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:19:00 UTC