RE: CSS 2: priorities in cascading order

On Thu, 18 Dec 1997, Charles (Chuck) Oppermann wrote:

> If we add more attributes, what's to prevent someone from always using the
> "!SUPER-DUPER_INCREDIBLY_IMPORTANT_ULTIMATE" attribute on their site simply
> because it's a guarantee of appearance.

The purpose of my proposed change to !important; (or the proposed
!absolute; directive) is to prevent this from happening. It does so by
ensuring that, once a style rule has been marked as important, no later
rule (whether supplied by the author or not) will change the properties in
question, irrespective of whether the subsequent rule is marked as
!important; or not. This is what is meant by saying that all of the
author's rules will carry normal weight, regardless of whether they are
marked as important. The effect is that the !important; directive, or the
!absolute; directive would have no effect if included in an author's style
sheet.

As to the suggestion that the problem is best dealt with by a browser's
user interface, I disagree with it for the following reasons. Firstly, by
allowing a directive to be included in the style sheet which gives
precedence to certain of its rules at the expense of those of the author,
all that the reader must do is download (or write) the style sheet. No
user interface changes are required. Secondly, my proposal enables the
precedence of the reader's vs. the author's preferences to be controlled
at the level of the individual style rule, which user interface options
may not do. The reader may have certain essential presentation
requirements which are indispensable to accessibility, and others which
are not. The style sheet language should enable these to be distinguished.
Thirdly, if the mechanism whereby the reader's needs are met is built into
CSS, it becomes part of the standard itself and is hence mandatory, rather
than being a user interface guideline which software developers may, or
may not, choose to implement. Fourthly, the two approaches are not
inconsistent. The user interface could allow the user to set various style
options, and implement this by making changes automatically to a default
style sheet and marking relevant items as important. Thus I do not think
there is any need to choose between the two approaches, but rather to
insist on both. Fifthly, it is a central tenet of the Web Accessibility
Initiative that where possible, W3C protocols and formats should
facilitate access to documents in different media. As it stands, CSS 2
fails to do this in as much as it offers no means for the reader to ensure
that her/his presentational requirements are met. For this reason it is
desirable that changes be made at this stage of development so that access
requirements are satisfied, in a convenient and predictable way.

Received on Thursday, 18 December 1997 19:55:12 UTC