W3C home > Mailing lists > Public > www-style@w3.org > October 2003

Re: CSS21 16.5 Capitalization text-transform

From: Martin Duerst <duerst@w3.org>
Date: Thu, 16 Oct 2003 20:46:41 -0400
Message-Id: <>
To: Tex Texin <tex@i18nguy.com>, www-style@w3.org
Cc: W3c I18n Group <w3c-i18n-ig@w3.org>

Hello Tex,

Some great work, thanks. A few comments below.

At 05:22 03/10/16 -0400, Tex Texin wrote:

>A couple of comments on text-transform-
>The first is a pet peeve of mine, which I hope can be addressed.
>I find in a number of browsers, that if text-transform is used to change case,
>and a user highlights and copies the text to the clipboard (on Windows 
>the text that is placed in the clipboard is the original text (ie without case
>changes) not the properly cased text displayed in the highlighted region.
>I think this behavior is very contrary to user expectations.
>It would be good if the CSS spec would specify cut/copy clipboard behavior 
>respect to functions that change text.
>Should the original text or the resulting text be placed on the clipboard?

I agree with Richard on this one in what the spec should say
(at least as long as copying means downgrading to plain text),
even though I agree with Tex re. the expectations of some users.
Richard in his example is an author, not the end user (reader),
and thus knows more and has different expectations.

>Second, the spec says:
>"...consider the value of 'text-transform' to be 'none' for characters 
>that are
>not from the Latin-1 repertoire and for elements in languages for which the
>transformation is different from that specified by the case-conversion tables
>of ISO 10646 ([ISO10646]). "
>a) I suspect you mean EITHER/OR not AND.

I suspect it indeed intended to say AND. When CSS2 was created, Unicode
was much less prevalent than now, and many people were afraid to
require anything outside Latin-1. This restriction is outdated and
should be changed.

>For example the letter i uses different conversion values in Turkey, and I
>assume that even though it is a latin-1 letter, you intended for the 
>to not be required.

>However, this seems silly, since then implementors need to have a list of
>characters that have different conversions for certain languages and insure
>that no conversion is performed. At the point you are detecting these
>characters and language contexts, you may as well implement the correct

Fully agree. Why require a very specific exception for a halfway job?

>b) From an international perspective, I don't see why you mention latin-1 at
>all. Why should latin-2 users for example not have the benefit of
>text-transform? Why not simply require support for the 10646 case conversion
>table, since Unicode character support is more generally required elsewhere?
>Given everything else needed to implement CSS, the Unicode case conversion
>table seem to be a very small burden to implementers.

Yes, I agree this is the right thing to do. Keeping the latin-1
restriction would be completely outdated.

>3) The reference to 2070 should be upgraded to rfc 3066.
>Also, what should occur if the empty language tag is specified for the 
>in XML? Use the default conversion or perform no conversion? What if the 
>tag is

I don't think we need to mention 'UND'. For missing/empty language tags,
the default conversion is fine.

Regards,     Martin.
Received on Thursday, 16 October 2003 21:25:07 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:27:09 UTC