W3C home > Mailing lists > Public > www-international@w3.org > October to December 2003

[i18n-html-tech] some comments

From: Bert Bos <Bert.Bos@sophia.inria.fr>
Date: Thu, 9 Oct 2003 23:57:57 +0200 (CEST)
Cc: www-international@w3.org
Message-ID: <Pine.LNX.4.05.10310092106450.30554-100000@lanalana.inria.fr>

On Thu, 9 Oct 2003, Richard Ishida wrote:
> [1] http://www.w3.org/TR/2003/WD-i18n-html-tech-20031009/

Some comments:

0) This promises to become a very useful document. It is very readable and
looks nice. Please keep it like this when the missing sections are added.

1) Nice icons :-)

2) The outline view has "Editors' copy" in the title.

3) Section 2.1: I normally try to avoid putting META in HTML documents for
specifying the charset (preferring the HTTP headers instead), since I
often transcode my documents, in my case between Latin-1, ASCII and UTF-8,
to be able to do different things with it and depending on the software
I'm processing the document with. Transcoding of XML and HTML documents is
quite easy, but changing the META header is not. Most of my tools don't
even know whether they are working with HTML or some XML-based format.

4) Section 2.1: "xml:lang" and "lang": note that XHTML 1.1 (and most
likely XHTML 2) does not have a "lang" attribute.

5) Section 2.2: avoid CSS property values 'left' and 'right': this
guideline is a nice example of why this is bad advice: the equivalents of
left/right are not before/after, but start/end :-)

I think people should not try to use before/after and start/end, but stick
to top/bottom and left/right, unless they really know what they are doing
and have the necessary knowledge to test their style sheets in
left-to-right, right-to-left and top-to-bottom texts.

  - Left/right/top/bottom is easy to understand, always means the same, no
    matter what language your document is written in and avoids mistakes
    like the one in this draft.

  - A style sheet (without any rules that *explicitly* select on language)
    that works for left-to-right and right-to-left languages is either
    unlikely to be very interesting or very likely to be wrong in some
    aspects of typography. And that the same style sheet could be useful
    for vertical text is something I simply don't believe. A vertical
    document doesn't look like a rotated horizontal document. In fact, it
    is more likely that the left margins are the same than the "start"
    margins. The lines may go vertical, but the paper/screen didn't change
    and neither did the reader's body.

  - Most people will not write style sheets for documents in different
    languages, or at most in two very similar languages, like English and
    French (but even then the style sheets for the two documents are
    unlikely to be exactly the same). Asking them to write 'start' when
    they mean 'left' and when 'left' in in fact perfectly acceptable is
    cruel and doesn't make writing CSS very attractive.

  - CSS also uses the terms before/after to insert text such as section
    numbers or boilerplate text into a document, but that is a "logical"
    before/after, as if you inserted an element into the document tree; it
    says little about where the inserted text will be printed, because
    there are other properties for that. You can very well insert
    something 'before' an element and then use 'float: right' to make it
    appear on the right side.

I think there is a fundamental misunderstanding of what CSS is for.
Because device-independence, accessibility and language-independence are
looked-for characteristics for some W3C specifications, people think that
they must be characteristics of *all* W3C specifications. But if you can't
put the language-dependent aspects of a document in the style sheet, where
*can* you put them?

For an example of what I mean, look at the "CSS Zen Garden"
(http://www.csszengarden.com/). This is a site that shows the power of
clean HTML and CSS in the hands of professional designers. The authors
have thoughtfully provided Dutch, French, Russian and other translations,
but the style sheet designers didn't know, or didn't know how to handle
those languages. If you try to apply the various styles to translated
pages, you'll quickly see that the styles don't work: things are supposed
to be bold or italic, but the script doesn't have these concepts; the
titles are images in the style sheet and thus the translations disappear;
the layouts are clearly Western style, more specifically English: the long
words of Finnish already don't really fit in some of the narrow columns.
There is no Hebrew or Arabic translation, but you can try to imagine: how
would you do "Pret-a-porter", e.g.? And I have no doubt that a good
designer would be able to do an recognizably "equivalent" style with
vertical text, but it wouldn't be done by replacing some 'left' by
'start'...

6) Section 3.2: XML encoding declaration: same comment as (3) above.

7) Use the META element: this is a duplicate of Section 2.1.

8) META as early as possible: also a duplicate of Section 2.1.

9) Use hex escapes: In CSS there is no other way than to use hex and in
XML it is allowed, but in HTML you cannot use hex escapes, but must use
decimal (unless SGML changed recently).

10) Section 5.1: Hmm, I must be missing something: this guideline is
another duplicate. And in the following sections there are several more
duplicates...




Bert
-- 
  Bert Bos                                ( W 3 C ) http://www.w3.org/
  http://www.w3.org/people/bos/                              W3C/ERCIM
  bert@w3.org                             2004 Rt des Lucioles / BP 93
  +33 (0)4 92 38 76 92            06902 Sophia Antipolis Cedex, France
Received on Thursday, 9 October 2003 17:58:07 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 19:17:03 GMT