RE: draft-ietf-html-style-00.txt & class as a general selector

cwilso@microsoft.com
Wed, 6 Dec 1995 16:42:39 -0800


preece@predator.urbana.mcd.mot.com (Scott E. Preece) wrote:
>Well, it doesn't address the issue I raised in earlier mailing-list
>discussion. When I really am doing styling of unique elements, where what
>I am doing to each element is specific to that element and there is no
>class of similarly-styled elements, I really want to be able to use
>anonymous styles rather than having to create pseudo-names and indirect
>through a stylesheet. I don't want to lose that notational concision in
>the interest of forcing other people to do the right thing.

Agreed. For stylesheets to overcome the urge to add tons of presentation
tags to HTML, they need to be as easy for the hand-author to use. There are
still lots of people who will set the font for each section of text in
Microsoft Word, instead of using the stylesheet support, because it is
easier for them to author that way.

>My model for stylesheet adoption is that once stylesheet support is in
>the major browsers, the net will rapidly evolve a relatively small
>number of standard stylesheets (the browser vendors are the most likely
>sources for those standards). Once a few common stylesheets are
>available, the examples people learn from, in books, net tutorials, and
>common source code, will deomstrate the use of those standards and
>*most* new authors will simply learn that as the way to write. Ths will
>be supported by organizational authors who are required by their
>organizations to write to house styles.

Hmm, now there I'd disagree... I think it will be popular to have
corporate-wide stylesheets, especially on external marketing Web sites, for
corporate branding. In addition, LOTS of authors will author style sheets
for every document, for a unique look.

>I don't really think the STYLE attribute is going to be all that tempting
>to the vast bulk of authors, once they have a few examples to look at of
>putting all the document-wide styling information in one place. Most
>people don't want to type "<STYLE font=Palatino size=14pt>" on 83
>headings if ther's a way to type it once and have it apply to all
>headings.

Definitely. I think it's primarily a good idea to embed style information
directly for those who wish to author documents by hand - if an authoring
tool produces the document and stylesheet, it's just as easy to embed the
info into the header and use the ID.

I'm a little unclear on why it is desirable to separate CLASS and a
classification of style - Michael, could you explain this? In the limited
instances where stylistic information should be set on a class of objects
that already belong to a larger CLASS (denoted by attribute) or to disparate
CLASSes, it would seem that ID could be set.

I *WOULD* like to see, however, the ability to set style properties on a
class independent of the element name. E.g.: [CLASS="foo"] { color:blue }

Should this be allowed in CSS level 1? It is similar to the
TAG[ATTRIBNAME="value"] selection described in the sketch of Level 2 in the
fifth draft spec. It would seem an intelligent thing to add to Level 1,
since without this, it would seem the only way to set a global "class" is to
use the ID parameter (which *should* be unique.) I also thought of
supporting " .class = {}" (selector without an element name), but that seems
kind of ugly. Comments?

-Chris