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

Re: [css3-writing-modes] horizontal-bt writing mode

From: David Hyatt <hyatt@apple.com>
Date: Wed, 27 Oct 2010 12:08:17 -0500
Cc: www-style list <www-style@w3.org>, fantasai <fantasai.lists@inkedblade.net>
Message-id: <9B072D2A-A50B-4ED3-B6A9-F98DCE0C8081@apple.com>
To: John Daggett <jdaggett@mozilla.com>
On Oct 26, 2010, at 9:04 PM, John Daggett wrote:

> David Hyatt wrote:
>> There is really no reason to leave it out.  Once you've abstracted
>> your engine, horizontal-bt just comes along for free.
> In terms of initial implementation sure, but you still have to have
> someone make tests for those extra values, and run those tests each time
> a test suite is run.  Running a minor test a zillion times means a real
> cost.

Right, although my point is just that the code for this mode will be in the engine regardless.

> Although the current CSS3 Writing Mode spec says nothing about this,
> writing-mode affects UI interactions.  For example, in vertical-rl mode
> IE8 alters the default viewport and how the mouse scroll wheel interacts
> with it, scrolling the wheel "down" scrolls left in the vertical case. 
> Oddly, the page-up/page-down/home/end keys are all "logicalized" but
> up/down/left/right arrow keys are not.
> To see this, view the examples below.  The Flash example mimics the IE
> UI behavior:
> Testcase for IE:
>  http://nadita.com/murakami/tests/wagahaiwanekodearu-vert-1.html
> Nice example layout using Flash:
>  http://macromarionette.com/#/p/25
> So in the horizontal-bt case what happens?  Using the model above the
> viewport would be set to the bottom left/right and
> page-up/page-down/home/end would all be reversed.  That's consistent and
> simple to implement but some poor tester in Beijing is going to have to
> test this each time some test suite is run.  Why waste people's time
> this way?

Again, the code path will be there if you're just using the principal direction of the document to key off of.  It's not like you write a bunch of unique horizontal-bt code.

I'd rather have tests for the mode, since the code will be in the engine anyway.  Also I have already had a situation where a horizontal-bt test caught a bug that I hadn't spotted in the same test in vertical-rl mode.

Received on Wednesday, 27 October 2010 17:08:53 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:38:40 UTC