[TYPO3-content-rendering] No more inline styles in css_styled_content / Do we care about IE7

[Christian Hennecke]

Am 28.07.2011 16:48, schrieb Patrick Broens:
> Hi,
>
> Some of you might have heard of the BLE project [1]. The project is
> called Incubator on Forge [2] and part of it is the Accessibility
> project [3].
>
> The reason why I want your attention is the following:
>
> One of the demands of this project is to get rid of all inline styles
> produced by the TYPO3 core. Css_styled_content is one to blaim in this
> case.
>
> Except for a few small issues, inline styles are especially used in
> image blocks, with the content elements Images and Text&  Images. The
> reason for this was that long captions exceed the width of the image. To
> preserve this, the width was added. Externalizing the width to a dynamic
> stylesheet is no option. We will end up with a huge amount of those with
> big websites with a lot of images on each page.
>
> Therefor I've searched for a solution, which does not need any width
> declaration. There are two ways to solve this:
> * Using CSS declaration: display: table-caption;
> * Using tables with the<caption>  tag, only when the caption is needed
>
> With both options there are obstacles, especially related to IE6 and
> IE7. Both don't know the CSS declaration, which makes this option not
> usable. The table<caption>  is recognized by both, but (yes of course
> there is a but) the caption is always displayed at the top of the image
> in these two browsers. For other browsers, including IE8+, I use the CSS
> declaration caption-side: bottom; which is a CSS2.1 property.
>
> In 4.7, where this is targeted, we still need to support IE7.
>
> So here comes the question ;-)
> What's your opinion about this visual (not functional) drawback for IE7?

Personally, I couldn't care less but many customers of web agencies 
will simply not accept it. That said I find it quite astonishing that 
usage of tables is considered as a work-around in an *accessibility* 
project. One could argue that tables would be semantically correct if 
images and captions are placed in corresponding columns, but the usage 
you point out seems to require one table for each image. This is a no go.