jwz - the CSS peanut gallery [entries|archive|friends|userinfo]
jwz

  www.jwz.org
  userinfo
  archive
  rss

Links
[»| DNA (Log) (iCal) WebCollage (LJ) Mixtapes ]

the CSS peanut gallery [Sat, 25-Jan-2003 4:45 PM]
Previous Entry Add to Memories Tell a Friend Next Entry
[Tags|, , , ]
[music |Big Hat -- Dirty Old Dog]

So, as if I haven't been making enough friends lately... I have a question. Or really, some ridicule cunningly disguised as a question.

I seem to recall that the whole W3C peanut gallery, back when we released the Antichrist, I mean Netscape, totally lost their shit because HTML was supposed to be about MARKUP NOT PRESENTATION DAMMIT, and it was Fundamentally Wrong for people to do presentationy things in HTML: people should use <SHOUT> instead of <B> or whatever. And web page authors had better not complain if browser A laid things out differently than browser B, because all those tags are optional anyway dammit, and nobody made you any promises about how space would be allocated to table cells.

So, that's easy to understand. It's a consistent viewpoint, and while I don't agree that it's the most useful approach, it at least makes some sense.

    (These people also tended to pretend to care deeply about the blind and otherwise disabled. I am sympathetic to the needs of those users, but I can't help but think that those who claimed to speak for the blind were being more than a little disingenuous, just like those Hemp people who present their arguments in terms of their deep and abiding care for the textile industry, when their real motives are... something else entirely.)

And today, it's all about CSS. They gave up on trying to get people to use HTML the ``right'' way, and they invented this new meta-language for describing presentation. Perhaps it was an attempt at misdirection: ``hey, over here! Shiny! Use this instead of abusing <TABLE>! Free beer!''

So now we have this bizarre situation where the anal-retentive W3C crowd (who are now and always have been slaves to the de jure standards, regardless of what reality and the de facto standards are) have converged with the equally anal-retentive (yet philosophically diametrical) ``designer'' contingent (who are now and always have been trying to fit web design into the pigeonhole of the kind of paper-based print-layout design they learned in art school.)

It used to be we had two groups:

  • Those who only cared about the text, and if there were tags at all, they could be visually interpreted any which way, since they were primarily there for indexing purposes or something;

  • Those who wanted to make their HTML look exactly like the mockup they made in Photoshop or Illustrator or Quark.

And today we've got just the one group, and their battle cry seems to be:

  • HTML must only be used for semantic markup, but that semantic markup must rigidly adhere to the pixel-accurate positioning in the spec, so that we can still design our web pages in Photoshop!

This is, of course, a complete about-face. It's really quite bizarre. I must have slept through the point when the transition occurred. Either that or it happened in almost complete silence: I would have expected that kind of coup to have left behind rivers of blood and mountains of the dead.

linkReply

Comments:
[User Picture]From: [info]agent_smith
Sat, 25-Jan-2003 5:11 PM (UTC)

(Link)

I would have expected that kind of coup to have left behind rivers of blood and mountains of the dead.

We at the Agency prefer the quiet assasination of human filth, rather than the sort of wholesale slaughter you described.

Just so you'd know, Mister Z.
[User Picture]From: [info]bdu
Sat, 25-Jan-2003 5:28 PM (UTC)

(Link)

I would have expected that kind of coup to have left behind rivers of blood and mountains of the dead.

indeed, though I am not at all sad to see it happen.
[User Picture]From: [info]andrewducker
Sat, 25-Jan-2003 6:07 PM (UTC)

(Link)

I think that the advantage of CSS is that the formatting instructions aren't in the HTML itself. So the HTML remains a pure descriptive language, while the CSS says "and display the information like this"
From: billemon
Sat, 25-Jan-2003 6:50 PM (UTC)

(Link)

And mozilla still has that handy "use my colors" button ... though b*ggered if I can figure out how to make it ignore CSS completely :)
[User Picture]From: [info]jay
Sat, 25-Jan-2003 8:25 PM (UTC)

scripting...

(Link)

Hmm. I usually turn off Javascript and that also disables CSS as a side benefit.

I could be high on crack and imagine such things but you can try it.
[User Picture]From: [info]ralesk
Sat, 25-Jan-2003 6:53 PM (UTC)

(Link)

   Nice nice indeed, but web-designers' dreams fall apart when they meet different browsers :]

   However, these «things» W3C introduced instead of I, U and B… irk me to no ends.  Honestly, what is “easier to understand” for the accessibility programme in a EM or CITE or STRONG or whatever?  B would be just as fantastically simple, and it bolds the text (that is, by default, unless I go and override it in my CSS so it underlines instead), lust like STRONG SHOUT would and any decent program would should read it out strong.  “Easier to understand” my arse.
   Again do I love the idea behind a “well structured format” like XML, but again, what is easier to parse in XHTML 1.x than in say HTML 4.0?  Each have a nice predefined set of instructions and there.  There is not much trouble telling the parser, hey guy, here's a list of goo, these have no end tags, don't worry about them.

   And lots and lots of other irkage.  Among them a couple of stuff about XHTML 2.0 which is barely like any of its predecessors, does have a couple of good ideas – IMG tag gone! –, but it omits tags like – for example – H1..H6, though those have been the base of everything HTML ever since… well, long ago.
[User Picture]From: [info]netik
Sat, 25-Jan-2003 7:16 PM (UTC)

(Link)

One of the most fascinating aspects of the whole debate is the current proliferation of handheld web browsers and PDA's like the HipTop. Those users are very much like the bilnd in that they want text to work with. In their case it's for smaller screens, and not for text to speech, but ultimately, clear layout and order of words is still paramount.

Now that everyone's been a complete idiot and forced pixel-accurate tables and HTML down the throats of the world, the people are still dependent on text-based (and low-image content) browsing on wireless are having real issues reading the pages.

So, what's their solution? Instead of being smart and detecting browser, then changing content (or redirecting) based on browser type (text for lynx, hiptop, graphics for others, etc) they invent yet -another- language, WAP. WAP sucks though, and works with stupid menu style interfaces.

Idiots who use WAP, or don't want to use WAP, then make new web sites (i.e. mobile.ebay.com) for mobile users that noone can remember the URL for.

What's so hard about doing a browser detect and doing the right thing? Then people can do whatever they feel like with HTML or CSS and it won't really matter. Sure, it requires rewriting content, but that's what XSL/XML is there for. Write one, build presentation templates a few times, and deploy.
[User Picture]From: [info]jorm
Sun, 26-Jan-2003 1:27 PM (UTC)

(Link)


What's so hard about doing a browser detect and doing the right thing? Then people can do whatever they feel like with HTML or CSS and it won't really matter. Sure, it requires rewriting content, but that's what XSL/XML is there for. Write one, build presentation templates a few times, and deploy.


There is always a new browser-of-the-week that you have to pay attention to that doesn't work with your old stylesheets - and it will, of course, be kind to you and send a User-Agent string that matches a different browser for you. And guess what? You can't test for what a browser can *do*, because they make that impossible as well.

Here's an example: Opera is the *worst. browser. ever.* - partly for this thing. It *says* it supports both document.all and document.getElementByID(), but it really doesn't support document.all. And then, it really doesn't support document.getElementByID, either - so you try to see if it can with "if (document.implementation.hasFeature("Events", "2.0"))" because that *throws an error* in Opera. Oh! And you can't detect for "Opera" in the user agent string becuase it's user-definable!

It doesn't help that Opera is the *worst* as far as being able to handle CSS.

(At this point, stuff I build has *a* stylesheet and is css positioned and fuck 'em if the browser doesn't display it correctly. I'm slowly moving all my sites to new designs based on this but hey - it takes a LONG time, even with XML->XSL rendering).

XML/XSL helps make the idea of rendering for different browsers easier, but the simple fact is *you shouldn't have to* - designs should be readable in every browser (so doing pixel perfect positioning by tables is a sin - one I'm guilty of). However, most pages were built when this wasn't something that people thought about - and it's an *insanely* difficult process to translate old html 3.2 documents into XML to move forward - and the cost is prohibitive for most companies and organizations.

Sure, there are web "standards" - and in theory, if you develop for them you'll be "forward compatible" - but in practice that's bullshit. The W3C is *developer hostile*: html 3.2 was easy to understand becuase it was supposed to be accessible to the masses but as time goes on the standards become more and more pointlessly complex because the w3c apparantly decided that they didn't like the unwashed masses playing in the sandbox with them.

It's an author problem *and* a browser problem *and* a standards problem. But no one wants to think they have a share in the blame.
[User Picture]From: [info]flipzagging
Sat, 25-Jan-2003 7:49 PM (UTC)

(Link)

What I have to say mentions Netscape's failures a lot. Don't take it personally; as I see it, this is just how things turned out.

First, I think your premise is a bit off. It's not that there were ever two groups -- one all for text, one all for graphics. Only a tiny group of users is content with a pure text and markup web. A majority would like the headlines to at least be the right size and font. And a large minority wants sophisticated graphic design. (I thought Netscape's early success was partly due to recognizing this fact?)

Your conclusion also misses the mark, slightly. CSS-o-philes are ambitious designers, but those who want to work within the concepts of the web, not subvert them like Flash does. CSS proponents were and are trying to stop the web from tilting towards technologies like Flash and (back in the day) Microsoft ActiveX thingies.

I will agree, however, that there is a sort of alliance between a self-appointed web vanguard and the W3C on this issue.

This happened because of the chaos that occured in web design shops, particularly the high-end ones around 1995-1998, trying to make a client's web page look professional -- and around that time, you had to deal with many, many different browsers which were all broken in idiosyncratic ways.

Early on, CSS was touted as an answer. But Netscape 4's support was laughable, so pragmatic web programmers simply served different versions of a page to different clients.

Worse, at this point Netscape started to drop off the radar screen. That's when your orange-tinted-hornrim designer types started to become a standard-bearer for, erm, standards, and became deeply involved in the process. Nobody wanted keep on coding 10 versions of the page, and nobody trusted Microsoft.

To their credit they got Microsoft to support more vendor-neutral standards, even while there wasn't much serious competition. And CSS is genuinely useful for basic fonts and layout. But because of this group's own limitations, a lot of silly things got added to the CSS spec.

The really idiotic things in CSS aren't to do with graphic design, but with stuff that clearly goes outside pure styling issues. Take the :before and :after pseudo-elements. It becomes clear that the impetus for the standard is at least partially "wouldn't it be cool if...", and CSS 3 is even worse. This standard is now driven by graphic designers, and by standards bodies. Who's missing here? Oh right, programmers. Search for "rcss" in this commentary on raph's open source projects.

P.S. Are you saying people who resist de facto standards are foolish?
[User Picture]From: [info]jwz
Sat, 25-Jan-2003 8:26 PM (UTC)

(Link)

P.S. Are you saying people who resist de facto standards are foolish?

Yes. Yes, I am.

More specifically I am calling foolish those who point to a de jure standard and say "but but but it's in the spec!" while ignoring that which evolutionary pressures out in the real world have selected as the winner.

What's a "de facto standard?" It's what we call a standard.
What's a de jure standard?" It's what we call a wish.

[User Picture]From: [info]ladyada
Sat, 25-Jan-2003 8:47 PM (UTC)

FUCK YOU

(Link)

XSCREENSAVER FUCKING HANGS MY LINUX BOX ALL THE FUCKING TIME IM FUCKING PISSED NEXT BMAN IM GOING TO GO TO YOUR GODDAMNED AIR CONDITIONED RV AND BEAT THE SHIT OUT OF YOU YOU POSER!!!!!
From: iamblue
Sat, 25-Jan-2003 9:09 PM (UTC)

Re: FUCK YOU

(Link)

Was that sarcasm or are you really that much of a spaz?
[User Picture]From: [info]ralesk
Sat, 25-Jan-2003 9:13 PM (UTC)

Re: FUCK YOU

(Link)

   Oh, this was incredibly mature and full of information :D  I'm highly amused.

   Humans never cease to amaze me.
[User Picture]From: [info]yakko
Sat, 25-Jan-2003 11:32 PM (UTC)

Re: FUCK YOU

(Link)

Oh shit!

Bored with Slashdot, the trolls have found their new source of entertainment: the ALL-CAPS RANT in jwz's journal!

(thanks for reminding me I sorely need to troubleshoot my 3D problems and upgrade xscreensaver.)
From: zhixel
Sat, 25-Jan-2003 10:55 PM (UTC)

(Link)

I keep thinking about why making a pretty webpage is so much more tolerable now than it was five years ago and came a conclusion that it's so much the introduction of CSS and DOM. It's that Communicator 4.x is finally dead. d-e-a-d. As in 'I will not support this shit any longer.'
From: [info]naturalborn
Sun, 26-Jan-2003 12:25 AM (UTC)

(Link)

This is, of course, a complete about-face. It's really quite bizarre. I must have slept through the point when the transition occurred. Either that or it happened in almost complete silence: I would have expected that kind of coup to have left behind rivers of blood and mountains of the dead.

TBL is dictator of the w3c, so if he decides to do an about-face, the w3c decides to do an about face.

Every time I read an article about how the w3c is 'enabling the future internet' I completely lose it. Those assholes are shoving bloated, disgusting 'standards' which can take years to implement down everyone's throats, then are given credit for the actual implementing, which they have nothing to do with.
[User Picture]From: [info]vsync
Sun, 26-Jan-2003 1:29 AM (UTC)

schizo standards

(Link)

As I've said before, the problem now is that the W3C wants it both ways. They want HTML documents to include all kinds of useful semantic markup, but despite the fact that no one is prescient enough to foresee all the data/document types that will be needed in the future, they want to lock down which tags are allowed and which aren't. (Whatever happened to degradation, huh?)

Strangely, they have lately taken to removing some of these semantic elements, with the excuse that "user agents don't distinguish anyway". So they claim to produce the one standard to rule them all, but they are clearly showing that by ignoring them, software programmers can make them go away.

Their current kick seems to be CSS, as you've pointed out. I guess they figure that having documents that are nothing but a mess of <div class="foo"><span class="bar">...</span><span class="quux">...</span></div> is better than standardizing on a broad subset of common document elements, or allowing software to have custom degradable elements. Of course, we're all supposed to be using XML with custom DTDs (oops, I mean schemas today; what's it to be tomorrow?) and XSL and XSLT, so from their point of view, HTML has been relegated to the status of presentation language.

I note with some happiness that LJ does allow a lot of HTML's semantic elements through. Sadly, though, due largely to the W3C dropping the ball, these aren't enough. Pinot (my Web thingy where you can post comments and stuff) currently has custom semantic elements built in, with transforms to clean HTML. I'm planning to do some other stuff that should help a lot too (namespaced customizable entities for abbreviations and whatnot), but once again, all this will be lost when it is transformed to HTML for viewing. At least it's all preserved within the site, and if exported to a sane format will retain the semantics.

[User Picture]From: [info]jcurious
Sun, 26-Jan-2003 3:16 AM (UTC)

(Link)

D00dz My site is going to KICK! It's going to utilize HTML,XSL,CSS,XPath, XLink, XHTML, XForms, WebCGM, SVG, SOAP, RDF, PICS, MathML, be WAI compliant, and not be anything useful/entertaining/etc... people will come to it because it'll validate ;)
[User Picture]From: [info]dingodonkey
Sun, 26-Jan-2003 10:08 AM (UTC)

(Link)

You're a liar. Nobody uses MathML.
[User Picture]From: [info]blufive
Sun, 26-Jan-2003 4:25 AM (UTC)

but...

(Link)

It used to be we had two groups:
* Those who only cared about the text,[snip]
* Those who wanted to make their HTML look exactly like the mockup[snip]

And today we've got just the one group, and their battle cry seems to be:
* HTML must only be used for semantic markup, but that semantic markup must
rigidly adhere to the pixel-accurate positioning in the spec, so that we
can still design our web pages in Photoshop!


From where I'm sitting, it's a more complex than that.

There's a large crowd of "pixel-perfect" people, many (but not all) of whom write table-layout tag soup. Many of them are fighting tooth-and-nail against supporting anything but IE and NN4 (and will drop NN4 like a hot potato at the earliest opportunity).

Then there's a crowd going the semantic markup/CSS route, who don't care about the odd pixel.

There's a lot more of a continuum between these two extremes than you might guess from the arguments. There are people doing very pretty designs using semantic markup and CSS. There are still people doing the "just the text, Ma'am" approach. There are horrible text-only table-layout tag soup monstrosities, too, and a depressing number of sites that rely on fixed-size Pop-ups and Flash.

But many people have noticed that the W3C HTML4 and CSS2 are now de facto standards. Sure, there's the odd bit that doesn't work in all browsers (especially in CSS2), but most of it works in all the major browsers. Except NN4, of course, but that's dying a death now.
[User Picture]From: [info]jorm
Sun, 26-Jan-2003 1:37 PM (UTC)

Re: but...

(Link)


But many people have noticed that the W3C HTML4 and CSS2 are now de facto standards. Sure, there's the odd bit that doesn't work in all browsers (especially in CSS2), but most of it works in all the major browsers.


You're joking, right?

There are *two* browsers out there that have sufficient support for HTML4 and CSS2 that you can do things that are in the spec as a web page developer and be "relatively" sure will work correctly: IE 6 and Mozilla (and no, Virginia, this does not include Netscape variants, though 7 is a hell of a lot better).

Unless you're doing something extremely minimal, all the other browsers (Opera and Netscape 6 come to mind especially) turn your pages into Picasso paintings.

The "de-facto" standard is still HTML 3.2 with minimal font css. A lot of flash in that, too (using
[User Picture]From: [info]pexor
Sun, 26-Jan-2003 8:53 AM (UTC)

(Link)

Is it wrong to weep for the death of Display PostScript? Wouldn't a browser that could talk to your screen (within the boundaries allowed by the browser application and frame, of course, be Pretty Cool (TM)? As I think you imply, something like page layout is inevitable in browsers' futures. Why not simply take it to a possible conclusion and allow a "web page" to manipulate screen controls?

It's not like the browser is a thin client anymore, anyway. The World Wide Web is little more, these days, than a directory service for ever-multiplying Internet client/server applications. I begin to forget if it ever was.
[User Picture]From: [info]insomnia
Sun, 26-Jan-2003 3:01 PM (UTC)

(Link)

It is awfully strange, isn't it?

I am working on creating another site right now. I feel like if I were to use html, which I can use easily and efficiently to design the site, rather than CSS, which would take a lot more of my time and really test my patience, I'll be seen by some as some kind of caveman -- a throwback to 1996, back when the idea of actually putting content up on the web was the priority.

Obviously, it's not about what you have to say anymore, but how you say it.
[User Picture]From: [info]jwz
Sun, 26-Jan-2003 4:06 PM (UTC)

(Link)

I suppose it depends on who you're trying to impress with your site. Do you care what the CSS/W3C peanut gallery thinks? If so, perhaps you should get over that.

My dabbling in CSS has not led me to believe there's any practical advantage at all over nesting dozens of tables, given the drawbacks of each approach.

Hell, it's not even like this CSS crap is any easier. It's just differently complex.
[User Picture]From: [info]g_na
Mon, 27-Jan-2003 2:12 PM (UTC)

(Link)

(Not that you need another comment here, but here it is, anyway.)

Am I the only person who is happy with plain ol' HTML? I don't use CSS because it may or may not render properly on my browser (and I refuse to use that Inherently Evil browser), and I am a huge proponent of compatibility with older browsers. HTML works just fine, and using tables allows me to do all the formatting I would ever want.

Plain & simple is nice.

Also, Shockwave/Flash suck.