W3C home > Mailing lists > Public > www-style@w3.org > June 2010

Re: [css3-text-layout] New editor's draft - margin-before/after/start/end etc.

From: John Daggett <jdaggett@mozilla.com>
Date: Thu, 3 Jun 2010 23:30:41 -0700 (PDT)
To: "MURATA Makoto (FAMILY Given)" <eb2m-mrt@asahi-net.or.jp>
Cc: www-style@w3.org
Message-ID: <1606718560.348416.1275633041296.JavaMail.root@cm-mail03.mozilla.org>

Makoto Murata wrote:

> > Can someone summarize what the advantages of the proposed
> > margin-start, etc. properties are for internationalized text?  Is the
> > idea simply that one can flip the writing-mode or direction and
> > margins "just work"?
> Yes.  Especially when major implemetations does not support
> horizontal-writing, fallback from vertical writing to horizontal
> writing is a must.  If such fallback typically leads to
> miserable results, vertical-writing of CSS will probably die.  
> What is needed is a smooth migration path.

I think a number of the alternate proposals are trying to propose just
that, graceful fallback behavior.  One proposal is for a new
pseudo-selector that older clients would not recognize.  Authors would
put fallback rules in style rules without the pseudo-selector and
vertical-text-specific rules in style rules with the pseudo-selector,
that way only user agents that recognize the new selector would use
the styles defined within.

Other than graceful fallback are there other reasons you feel logical
properties are required?

> Some of you might think that vertical writing is not really required
> for Japanese web pages.  Might be true.  But e-books are different. 
> EPUB relies on CSS.  Most of the physical books I read is written in
> vertical writing.  

Whoa there cowboy.  How does criticism of a proposal to add lots of
difficult to implement new properties constitute "thinking that
vertical writing is not really required"?  Several times you've made
very broad assertions against those who criticize the margin-start,
etc. proposal without making the effort to understand their criticisms.
It would be better to focus on explaining more clearly what problems
your proposal solves and why it's better than other approaches.

> It is true that changing the writing direction while reusing the
> same stylesheet does not provide a perfect result.  To really
> provide good results, we have to rewrite the content.  Arabic
> numbers are appropriate for horizontal writing while kanji-based
> representations of numbers is often better for vertical writing  (no
> tate-chu-yoko). However, nothing is perfect.  For example, reflow of
> web pages does not provide perfect results for typographers.  Not
> being perfect is not a good reason to reject margin-start, etc.
> properties.

A lot of folks at browser companies are spending time and effort to
make typography better on the web.  If a proposed change requires a lot
of work and doesn't really solve underlying problems completely, why not
consider a different approach that does a better job?

> > I think I agree with Hakon, the :rtl, :lrt, :ttb proposal seems better
> > suited to language-sensitive design.  With these an author could
> > specify the exact desired styling for tate-chu-yoko numbers without
> > affecting horizontal display.
> I do not understand that :ttb proposal.  When the HTML family disallows
> "ttb"as a permissible value of the dir attribute (rightly so), how
> does it work?  Or, does :ttb apply when the user or reading
> system chooses vertical writing?

It seems to me some form of pseudo-selector or pseudo-element
dependent on writing-mode or whatever should be a simple solution
here, one that matches how authors would actually use styles specific
to vertical text layout.

> > For supporting Japanese layouts with a mixture of horizontal and
> > vertical text, I think it would be more interesting to consider a
> > better grid model and how to do "multi-column" vertical layout
> > rather than convoluting the box model.
> I have spoken with many people about the requirements on e-books for
> Japanese text layout.  Nobody said that a better grid model is a
> minimal requirement.  Most people assume that the support of
> vertical writing is a must.

You seem to be specifying specific solutions, not requirements. 
Implementing vertical text layout affects other parts of CSS, like
multi-column layout.  And once it's available then it's a natural
extension to want a way to easily mix horizontal and vertical text
blocks on a grid, needs shared by non-CJK authors.

> It is true that vertical writing is not often used in Korea or
> mainland China.  However, when I asked "Do you need vertical writing
> for e-books?" to the Korean head of delegation in SC34 and the
> Chinese head of delegation, they immediately said yes.  The Chinese
> HoD once wrote "China has never given up vertical writing".

Examples of Korean vertical text layout would be interesting.  And
understanding the contexts in which vertical text is used/required in
Taiwan/China would also be useful.


John Daggett
Received on Friday, 4 June 2010 06:31:15 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:38:36 UTC