Does Microsoft give a damn about CSS?
Paul A. Scott
Pardon my obvious irritation in the subject line, but c'mon guys, what's up
with CSS support in IE? Every version of IE--whether Windows or
Mac--implements CSS differently, and every IE version is WRONG--damnit,
WRONG! IE 6.0.2600 under Windows XP is what pushed me over the edge. In
spite of everything else wrong, I find that the latest IE now flows <DIR>
blocks over each other in MID-BLOCK! Jeeeesus! Let's not fix the problems,
let's add new ones!
There are just too many problems and errors for me to detail here, and I
really don't want--or think it's necessary--to point out all the problems
and differences, but I would sure like a statement from Microsoft that (A)
you acknowledge the problems, and (B) you're going to fix all of them RIGHT
the very next release! I don't care if you support every CSS2 capability,
but please get it right what you do support. Currently, even the CSS1
support is .... well, it just plain sucks! Granted there's a lot of nitty
details when you dig into CSS and its implementation, but it's not quantum
mechanics for cryin' out loud. A little common sense and dedication will get
the job done right.
There. Well, I don't really feel any better--because I've still got'ta deal
with a screwed up CSS-- but at least I've got that off my chest, and I can
get back to bitchin' under my breath.
Paul
Erik Arvidsson
IE used to be the leader of CSS. Microsoft where one of the major players
when creating the CSS standards. That was with IE4 and now IE6 is at the
bottom of all modern brwosers when it comes to CSS.
1. Mozilla: 38 points
2. Konqueror: 30 points
3. Opera 5 & 6: 29.5 points
4. Explorer 5 Mac: 28 points
5. Explorer 6: 21.5 points
6. Explorer 5.5 Windows: 20 points
Explorer 5 Windows: 20 points
8. Explorer 4 Windows: 16.5 points
9. Explorer 4 Mac: 12.5 points
10. iCab: 6 points
Netscape 4: 6 points
Omniweb: 6 points
See http://www.xs4all.nl/~ppk/css2tests/
Too bad about 90% of the users use IE. This forces the industry to waste
millions of dollars on working around bugs and use IE proprietary solutions.
erik
"Paul A. Scott" <psc...@skycoast.us> wrote in message
news:B9F81AD1.1184B%psc...@skycoast.us...
Rowland Shaw
http://www.xs4all.nl/~ppk/css2tests/background.html
vs
http://www.meyerweb.com/eric/css/edge/complexspiral/demo.html
It's apparent that their is a difference in who comes out as compliant
testing the same feature.
I do not feel your pain. At the end of the day, HTML is a markup language,
not a page description language. In fact there's a very good article on this
point at http://hotwired.lycos.com/webmonkey/02/42/index4a.html?tw=design
"Erik Arvidsson" <er...@eae.net> wrote...
Paul A. Scott
> testing the same feature.
duh
> I do not feel your pain. At the end of the day, HTML is a markup language,
> not a page description language. In fact there's a very good article on this
> point at http://hotwired.lycos.com/webmonkey/02/42/index4a.html?tw=design
We're talking CSS not HTML.
Paul
BTW, it's 'there' not 'their'
Paul A. Scott
> blocks over each other in MID-BLOCK! Jeeeesus! Let's not fix the problems,
> let's add new ones!
Well, I have to eat my words on this one. Thanks to Fabrizio Giustina for
pointing to the CSS1 test suite:
http://www.w3.org/Style/CSS/Test/CSS1/current/
that indicates IE 6.0 exhibits correct behavior on <DIV> flow. This means IE
Win prior to 6.0 and IE Mac 5.2.2 behave incorrectly.
What a mess.
Paul
Rowland Shaw
> > It's apparent that there is a difference in who comes out as compliant
> > testing the same feature.
>
> duh
My point was more that different people treat the same results as different
outcomes -- the example given tested the same feature
(background-attachment: fixed), yet yielded different results.
> > I do not feel your pain. At the end of the day, HTML is a markup
language,
> > not a page description language. In fact there's a very good article on
this
> > point at
http://hotwired.lycos.com/webmonkey/02/42/index4a.html?tw=design
>
> We're talking CSS not HTML.
CSS still needs something to Style -- OK, it could be XML, but there isn't
Style without content.
Your original point is that there's not enough effort put into development
of CSS by Microsoft. I'd counter argue that they've put a lot of effort in,
but at present people aren't clamouring for the newer features. There's also
the band of people that want the HTML/CSS combination to allow for pixel
precise rendering, which just isn't practical (for many reasons).
My point from all of this is that content developers need to generate good
content that doesn't fall flat on it's face as soon as a client has a client
style sheet, that does degrade so the document is legible without CSS
support. Expectations are too high for what can be achieved, and web
developers are jumping through hoops to satisy marketing boys that don't
understand the technicalities, comments like "It shouldn't be too hard to
make that [highly DHTML dependent] application work with Netscrape 7, should
it?" and the like are heard too often.
If you take the view of looking at the market share of your users, Taking
Google's figures [ http://www.google.com/press/zeitgeist.html ], the Top 3
browsers are IE 6, 5.5 and 5. All have much better CSS implementations
compared to what's being used in the wild -- to take a page in the wild as
example, the UK's most visited website (according to a recent survey) was
bbc.co.uk, which does at least now do some CSS rendering, but relies mainly
on transitional HTML for it's content. I'd argue that if it was semantically
marked up, and relied more heavily on CSS, that it'd have better
accessibiliity (just slot in a different stylesheet for the lo-fi version)
and would also have encouraged user agents to fix their little niggles --
The whole of the UK going "this doesn't work" might be more encouraging than
a few people saying "this obscure testcase fails".
It's just an opinion, feel free to disagree with it, but like all opinions,
there is no correct or wrong answer.
Paul A. Scott
> to allow for pixel precise rendering
Yes, like W3C.ORG, which published this on their Web site:
"All 100%-conformant CSS1 agents should be able to render the document
elements above this paragraph indistinguishably (TO THE PIXEL) from [the]
reference rendering, (except font rasterization and form widgets). All
DISCREPANCIES should be traceable to CSS1 IMPLEMENTATION SHORTCOMINGS."
Paul
Rowland Shaw
Not what I meant.
Example of what I mean, in a document, I specify the user agent should use
"Verdana, sans-serif" as the font-family. User comes along without Verdana
installed -=> not pix perfect representation, per spec. User comes along
with client style sheet that forces everything into a high contrast colour
scheme with font-size: larger; again, different pixel representation to that
the author intended, but again the UA doing the right thing.