Drawing Back the Curtains on CSS Implementation

David Baron
, Mozilla
Elika Etemad, W3C Invited Experts
Molly Holzschlag, Developer Rel, Opera Software
Sylvain Galineau, Microsoft
Tab Atkins, Google Chrome

All members of this panel are part of the CSS working group.

Minutes from the CSS working group go out online every day… so keep up with what we’re doing and provide input.

How do you prioritize at the W3C?
What is the most interesting and fun to work on? Figure out where it is hard to develop websites in CSS, that’s where the time is spent.

Prioritize to specs that are really stable, or exists in other browsers already. IE doesn’t like to prototype in the browser. Also look at what people are asking for.

Specs need to stew a little while. They need to sit for a while on the heat before implementation. There should be time for collecting feedback. If there isn’t a lot of feedback, then it needs to sit a little longer.

You have a lot of opportunities to comment on standards. Tweet, or comment on the minutes. The working group is interested in feedback. http://www.w3.org/blog/CSS/

There are new features that make things easier, and new features to do things never done before. Sometimes more feedback comes from the things that you can do easier because they are fixing frustrations developers already have.

What do you think are the most damaging mistakes made by the CSS working group?

Fantasai: Making the width property not including border/padding. The current box model is not the best way.

Tab: Layout: This was short-sitedness, but CSS it was a document language, not for layout. Floats are being used for something they weren’t intended to do (they were meant for images and text).

David: Some of the most damaging mistakes were when the group made things more complicated than necessary. We hid some specs away to deeply in a document, a spec was interpreted in a way that lead to complexity (and browsers interpreting things incorrectly).

Sylvain: Floats are baffling… why not just position it like everything else? They are very fragile in actual use. Z-index is also a bit strange. Missing: at this point we don’t have a CSS substitution mechanism. Instead of adding to a style sheet, find a way to substitute… add more process information as it scales.

How come in 20 years of web we have not come up with a reasonable way for layout?

Tab: Layout managers are bad everywhere. It’s hard to do layout in a way that is easy to use. Flexbox, Grid, Position, Regions are coming soon. Soon these times will be available for CSS layout

Fantasai: Layout is hard to implement. Width and Height took a LONG time to implement. It’s a hard thing to add to webpages incrementally. On the web it’s not as easy as in print where everything is fixed. Websites are fluid, which adds complexity. The systems that people come up with for print don’t necessarily work in CSS.

David: It is important to look at different use cases for layout. Do we want one system for many use cases, or different systems for each use case? What are the use cases we need to solve?

Sylvain: Layout of what? An ad for a magazine are very different than say, UI. You can’t solve everything for everyone.

What’s coming in IE9

Border radius
Multiple backgrounds
Transitions (w/out javascript)
Image values (gradients)

What’s coming in Chrome?

Many types of typography improvements

Leave a Reply

Your email address will not be published. Required fields are marked *