Re: bindings, behaviours, CSS and aspirin (the one I need after this message)

On Thu, 30 Oct 2003, Daniel Glazman wrote:
>
> I don't think [BECSS] *is* CSS.

We (Daniel and I) discussed this on IRC, and came to the conclusion that
the reason he doesn't think BECSS is CSS is because, as I understand it,
he considers "style" to be a purely "look"-related thing, and never a
"feel"-related thing.

For a long time I've thought "style" included both look _and_ feel, and I
still think that today. The following rule:

   :hover { border: yellow solid; }

...is one that decides the _feel_ of the content -- it sets a behaviour --
but in my opinion it is still stylistic.

For me, there are only four aspects to a document:

   The definition of the markup language: The DTD or schema, and its
   related prose. This is what defines how the document itself is to
   be interpreted, and includes things like HLink definitions or RDF
   vocabularies. Documents use well-known definitions in order to
   guarentee that all recipients will be able to tell what they mean.

   The content: The semantics, the structure, the _data_: Anything that
   will be the same regardless of the medium, regardless of the abilities
   or disabilities of the reader. This includes forms, scripted logic
   (e.g. to validate the data entered on the form), and links to content
   resources.

   The content resources: images, sounds, videos, etc, that form an
   integral part of the content. Documents can provide multiple
   alternatives for each such resource (the canonical example is a video
   of the earth, a photo of the earth, and a textual description of the
   earth). This content is media-specific typically but there should exist
   alternatives for each media (even if it's just a textual alternative).
   External scripts or logic that is a key part of the content (such as
   form validation logic) but can be shared between documents is placed in
   this category.

   The style: presentation, look, feel, behaviour. This is media-specific,
   it is user-customisable, it isn't required to understand the document.
   It is possible to have different sets of these per media (e.g.
   alternate stylesheets). Stylesheets fall into this category, as do
   images used by the stylesheets. Scripts that perform purely stylistic
   purposes, such as making links open in a new window instead of the
   current one, are also style IMHO.

CSS is about the last of those four: it is about cascading _style_ sheets.
It doesn't, to me, matter whether the style is visual, aural, interactive,
static, or whatever -- it's still style, it's still CSS.

Thus, I think further behavioural extensions to CSS would continue to be
stylistic matters, and thus continue to be CSS.


> because of point 2, I think that we should have a new mimetype [...]
> then you could keep text/css for pure CSS. But I am not far from finding
> this counter-productive.

Yeah, personally I think the only real result of doing that would be
strange looks from the authoring community.

A new MIME type is _expensive_. Servers have to be updated, administrators
have to reconfigure, users have to be educated... and in the meantime,
existing UAs already accept CSS extensions as text/css, and for legacy
reasons are never going to stop doing so. I don't see any real benefits to
a new MIME type, and I see a _lot_ of disadvantages.

-- 
Ian Hickson                                      )\._.,--....,'``.    fL
U+1047E                                         /,   _.. \   _\  ;`._ ,.
http://index.hixie.ch/                         `._.-(,_..'--(,_..'`-.;.'

Received on Thursday, 30 October 2003 10:52:33 UTC