RE: Publishing the flexible box model

Hi David,

Excuse me if this suggestion doesn't make sense, or has already been
discussed. Anyway, reading the flexbox spec showed me a remarkable
resemblence with some work I did back in the '90s when implementing
custom layout managers in X-Windows (Xt Intrinsics) :-)

One thing that the users of these layout widgets found quite useful
was the ability to say that certain laid-out child widgets should
have the same size, typical example being buttons (Ok/Cancel/Apply)
getting the same width. The algorithm would base it on the preferred
widths (for orientation=horizontal) of the children and use the
largest one for all.

We exposed the functionality with these properties ("constraints") on
the children:
  sizeGroup: default -1
    all children with a sizeGroup >=0 would form a size group with
    other children with the same number and establish a common size
  sizeGroupFactor: default 1.0
    factor multiplied with the established size for the group to
    get the resulting size for this child (f ex makes it possible 
    to say that one child should be twice as big as the other in 
    the same group)

Best regards
Mike Wilson

> -----Original Message-----
> From: 
> [] On Behalf Of L. David Baron
> Sent: den 4 juni 2008 06:48
> To:
> Subject: Publishing the flexible box model
> I'd like to move forward with publishing the flexible box model as a
> working draft.  There have been specs floating around for years, but
> it's never been published on the TR page (despite, I believe, a
> minuted group resolution to do so at one point).
> I think the newest draft is
> .  (There's also
> an older draft at
> .)
> I propose we use the shortname css-flexbox; I'm willing to import it
> into space and prepare it for publication if needed.
> There's a good bit more technical work needed in writing up the
> details of the layout algorithm in mathematical form (and also in
> adding more examples, or readding dropped ones from the older spec).
> However, I don't think that prevents us from publishing it as a
> working draft so that we can get wider feedback on the material we
> have so far.
> I'll briefly summarize the reasons for this specification:
> The flexible box model module adds features to the CSS box model
> that provide some of the basic formatting concepts often used in
> user interface layout.  Much of the specification is implemented in
> both Gecko and Webkit (with prefixes), and this implementation forms
> the basis of the formatting model of XUL, the language used to build
> the user-interface of Firefox and a number of other applications.
> These formatting concepts have a good bit in common with other
> user-interface layout systems.
> Although many Web pages are essentially user interface, or documents
> in the middle wrapped in user interface on the edges, they are laid
> out using formatting primitives designed for formatting documents,
> causing much frustration for authors.  The goal of standardizing this
> material at the W3C is to provide Web authors with formatting
> primitives that are more capable of representing the layouts that
> they want to achieve while effectively using the space provided by
> browser windows of varying sizes.  (I talked about this in a good
> bit more detail in my talk at XTech two years ago; see
> .)
> -David
> -- 
> L. David Baron                       
> Mozilla Corporation             

Received on Thursday, 12 June 2008 11:47:59 UTC