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

RE: [css3-writing-modes] Properties with horizontal/vertical values

From: Koji Ishii <kojiishi@gluesoft.co.jp>
Date: Wed, 13 Oct 2010 13:25:25 -0400
To: Brad Kemper <brad.kemper@gmail.com>, fantasai <fantasai.lists@inkedblade.net>
CC: David Hyatt <hyatt@apple.com>, "www-style@w3.org list" <www-style@w3.org>
Message-ID: <A592E245B36A8949BDB0A302B375FB4E0A779A8DC2@MAILR001.mail.lan>
> How about overflow of an image? I might have 'overflow-x:hidden /* or visible
> */; overflow-y:auto;' because I know the image will fit horizontally and
> don't want a horizontal scrollbar to appear (not even a grayed out version).
> The image would not normally get turned with vertical/horizontal text
> switching.

Very good point. I don't think we have clear and agreed resolution to this yet, but here's my opinion.

For the replaced elements, some of them should not rotate while others should. We need clear definition of that in the spec. Images will not rotate as you said. Form elements should rotate but the UA may choose not to rotate if it doesn't have the capability.

If the UA decided not to rotate a replaced element, it should assume that writing-mode: horizontal-tb applied to the element. Otherwise, for instance, logical width/height will be incorrect. This rule also resolves overflow-x/y issue you pointed out.

This resolution, however, brings up a new issue to the logical margin. Authors usually use margins to specify spaces between text and objects. If authors set margin-start: 8px to make some space between the label and textbox, and if the UA chose not to rotate the textbox, the margin appears at the wrong side.

To resolve this, my idea is margins use parent box's writing-mode to resolve logical directions. Since margins are outside of the borders and thus can be considered as outside of the box, I can think this as one of the way to understand the model. But I'm also feeling that there could be side effects, or there could be someone who feel strange. I need to think deeper on this, but if anyone has found any issues with this method at this point, I'd appreciate to know.

Koji Ishii
Received on Wednesday, 13 October 2010 17:23:18 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:38:39 UTC