RE: An alternative strategy
Jon Bosak writes:
>>Microsoft is to be commended for its willingness to push the
>>technology forward in this case. But I don't consider a
>>Microsoft-only solution to be a solution.
Well, I personally don't either - I have always *personally* (I'm not
talking about Microsoft here, so don't anyone bother flaming me) pushed
open discussion and resolution on standard solutions. In my opinion,
proprietary solutions are doomed on the Internet, as long as no single
entity controls the end-user software - this is a very good thing, in my
opinion, because it fosters innovation and interoperability. I was
responding to your comment:
>>I will stick my neck out here and predict that when Netscape and
>>Microsoft realize that everyone who really counts is getting by with
>>standardized embedded style attributes, they never will get around to
>>implementing CSS stylesheets, or will implement as little of the
>>standard as they can get away with -- leaving out the
>>context-sensitive selectors, for example.
I'd love to have all of CSS level 1 implemented in IE. That's my
development goal, at this point. As Bill Perry also said, CSS (and at
least two implementations of it) predates Microsoft's involvement in
stylesheets. Certainly, I've been involved in driving the spec for the
past few months, but I certainly didn't write it, and it doesn't belong
to Microsoft. We do support it, though - in concept AND in
>>| These rules only take up a few short lines, regardless of the number
>>| of items or length of text; that's not true if you use only embedded
>>| style attributes, and the content also ends up much larger with
>>| embedded attributes only.
>>This assumes that I as the publisher have calculated and inserted into
>>the stylesheet associated with each fragment only those style rules
>>that are relevant to that fragment. Not fun. The alternatives are
>>(1) supply the entire style sheet with each fragment; not a big saving
>>here (see the example style sheet in the current CSS proposal), or (2)
>>cache the style sheet; how far have you implemented this part?
(1) the example style sheet is rather large; my test sheets tend to be
around 1-2k at most.
(2) that's the idea; you would have a default set of "styles" -
presumably most companies would like to have a unified look to their
pages, and their "corporate style sheet" would be linked in to all their
pages. Support for linked style sheets is not finished in Internet
Explorer; it will be before we ship the final version (possibly not the
soon-to-be-released beta, since I'll be going to Paris at the end of
this week). It does (already) take advantage of our document cache.
>>| In my experience as a CSS implementor, the rule matching isn't the
>>| hard part, and in the long run, the work necessary for the rule
>>| matching (specifically, maintaining a representation of the document
>>| structure) will be necessary to support higher-end solutions anyway
>>| (so it's not throw-away work).
>>Are you sure that your solution scales to the general case? Does it
>>scale gracefully to handle the couple of examples that I posted here
(I'm referring to storing the document structure - not the actual
MatchRule() function.) If I recall correctly (in looking through my
inbox, I can't find your examples, but I do remember examining them.
The document structure representation I have would support them;
although currently there's no code supporting finding the second
paragraph following an <H1>, etc., it could easily be written without
any revision of the document structure.
>>I absolutely agree with you on this point. But I bet that not many
>>people will really want to code full-blown CSS stylesheets by hand,
I'd be delighted if authoring tools supported the creation of style
sheets ASAP so that users wouldn't have to hand-author them. The
reality for the last three years has been that most high-profile sites
(those likely to need style sheets) have been hand-authored, or
generated by scripts that are hand-authored. In comparing CSS and
DSSSL, it seems that CSS is easier for most non-technical authors to
understand, so if they HAVE to hand author a style sheet, I believe it
would be easier for them to do in CSS.