which format will win?

[DawnFalcon]

Originally Posted by frabjous

I don't know whether or not a good WYSIWY-whatever editor exists for web pages. I have no need of one. But nothing you've written has provided a single reason for why there couldn't be one.

I'm not quoting argument direct no. It's pointless without the background you've admitted don't have - it starts arcane and gets worse from there. The major, single, most clear, visible and damming evidence is that said editor simply doesn't exist in this day and age. It's 2010, not 1990!

Anyway, CSS is a fait accompli. You're going to have resign yourself to that fact.

Why? ("Publishers are going to control book prices. You're going to have to resign...")

Aside from IE, CSS noncompliance is really not that bad... and certainly it's fine for the needs of the casual user.

So where's the causal user's editor? Oh, right. It is that bad.

[guyanonymous]

Are ANY of the browsers out there even 100% compliant?

I believe Safari has gotten 100/100 on the acid3 test....but that doesn't mean (according to the acid3 site) that the browser is 100% compliant:

Originally Posted by http://www.webstandards.org/action/acid3/

The main part of test is automated through JavaScript, a sort of test harness that runs 100 subtests. Getting a score of 100 is not the same as passing Acid3 – a common misconception, or perhaps an oversimplification.

[DawnFalcon]

Yea Guy, there's one fully compliant program - Amaya, which is primarily an editor (and not even a web editor per-se, it's primarily an XML editor).

However, the problem is you effectively need to support a number of layout engines: several different versions of Trident (Internet Explorer), Webkit (Safari, Chrome), Gecko (Firefox) and potentially Presto (Opera) - and that's before you consider mobile browsers, which tend to have far poorer support...

ACID is obviously a wider-ranging test, and I'm focusing just on CSS support for my argument

[JSWolf]

Would would be nice is a fully compliant WYSWIG editor that allows us to see our CSS changes as we make them.

[JSWolf]

Originally Posted by DawnFalcon

Yea Guy, there's one fully compliant program - Amaya, which is primarily an editor.

However, the problem is you effectively need to support a number of layout engines: several different versions of Trident (Internet Explorer), Webkit (Safari, Chrome), Gecko (Firefox) and potentially Presto (Opera) - and that's before you consider mobile browsers, which tend to have far poorer support...

ACID is obviously a wider-ranging test, and I'm focusing just on CSS support for my argument

MS IE 8.x got a whopping 20/100.

[frabjous]

Originally Posted by DawnFalcon

I'm not quoting argument direct no. It's pointless without the background you've admitted don't have - it starts arcane and gets worse from there. The major, single, most clear, visible and damming evidence is that said editor simply doesn't exist in this day and age. It's 2010, not 1990!

You do realize that your position is bound to be unconvincing without direct argument?

I don't see how I don't have the requisite background. What you need is an explanation of the nature of CSS and/or its relation to (X)HTML that explains the situation. I know CSS and XHTML, and I have plenty of experience using WYSIWYG editors for other tasks, so I can follow along. That I don't have that much experience with WYSIWYG web editors doesn't mean I can't follow an argument for why the nature of CSS and XHTML makes one impossible, at least for the needs of the kind of casual user you see yourself as championing.

So fill me in please.

[DawnFalcon]

The direct argument is the inconsistent layout engine CSS support and lack of an easy-to-use editor. You could get a reasonable looking website out of Frontpage 2k, even if the code was messy as heck. Now, half of what it does won't render properly in browsers.

And I didn't say "impossible", I said it was so difficult it hadn't been done.

Short form of the arguments, copy/pasted from elsewhere (I'm not going to argue about the details since I don't 100% agree, and this was written in the CSS 2.0, not 2.1 era):

*CSS 2 was designed by engineers with little reference to usability, design principles or interface management. The original design of CSS1 was improperly directly extended into CSS2 without basic considerations as to longer-term usability being done.
*Layout is typically (in printing, etc.) done using areas of the screen as boxes (hence the old methodology of frames!), which is something which is actually quite hard and counter-intuitive to do in CSS. Also, vertical element handling support is problematic compared to horizontal element handling, despite their being of equal importance.
*Limited inheritence. There are times you can't find a parent/ancestor, breaking the entire inheritance model. With these times being essentially arbitrary, it adds arbitrarily to the difficulty of using CSS. On the flip side, elements are not allowed to inherit positional commands, creating a lot of unnecessary positioning issues.
*No expression support. This means you need to be far, far too specific as to what you're trying to do in many cases rather than creating a layout which will work roughly-as-designed in many situations. In particular, this bites hard when you're doing columns, which you cannot define as boxes but end up doing complex and fixed elements.
*Overlapping properties. Properties can step on each other, creating further inconsistencies in inheritance, and requiring in many cases arbitrary decisions as to dominance: something different browsers handle in different ways.
*Limited nesting properties support: There are some properties which cannot be nested in another, despite the fact that there is no essential reason why the approach shouldn't work.
*The net effect of the above is that the layout us being handled by something designed for styling, with predictable effects in that how it's used is quite different from how other forms of documents are created, and has become increasingly specialised.
*There has been a general complete disregard for ease of generating the code in a programatic fashion. In essence, the assumption has been for hand-editing the code rather than allowing in any way for how an editor is going to handle behind-the-scenes decisions for a user.
*tldr version; There is not a good WYSIWYG HTML+CSS editor. There will not be a good WYSIWYG HTML+CSS, or anything resembling such. I hope you like hand-editing.

[guyanonymous]

I wish that 'the web community' had just decided: 100% compliant or screw you. That would have got the browsers going really quick. Ah well

Bunch of enablers!

[DawnFalcon]

Or for that matter if, in 2003-2004 they'd simply declared IE to be the standard way of doing things, since it had a ~95% market share at the time.

(Why yes, I'm big on "reality in standards")

[nerys]

The problem is the "web community" at the time was . that big

while the user base was

00000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000 00000000000000000000000000000000000000000000000000

that big. so it only took a company to make a browser people wanted to get that 0 mass to move to it.

the . was simply ignored as the corporations and developers did what they wanted :-)

so while ie had 95% market share it was FICKLE and would jump ship without hesitation and was not the . bit.

:-)

[DawnFalcon]

Fickle? supported by the evidence, really. IE displaced Netscape, and since 2004 IE's slooowly seen it's market share slide away. The user base has shown a remarkable degree of loyalty to IE.

Certainly the web standards bodies have been unduly swayed by "experts", which works fine for some things, but...not a mass market medium like the web.

[guyanonymous]

I think that, where information is concerned, standards are critical.

Just look at the discussions and concerns about old formats no longer being readable on current devices (whether it's drm servers going down or simple abandonment but current/future devices).

Look at the problem odd-ball standards can cause (cost, safety, functionality), such as Imperial measures...I believe it's down to the US and one small country in Africa that use Imperial...and even scientists in the US don't use it anymore. It was a standard developed with less, well, forethought, than metric. It's still around, but it's not doing anyone any good. Metric, thought out more by experts, is far superior in all measures (pun intended), including usability. I'm glad for those experts.

Accessibility, consistency across browsers/editors/etc, a hope of real web-based typography and design, a guarantee that anyone, anywhere, anywhen who knows the rules of the standard can access the information.... that stuff I'd happily leave to the experts to decide. Alas, the 'non-experts' or the selfish (businesses that develop un-needed proprietary standards that don't achieve the above goals) don't always listen.

OK...I doubt I'm making any sense. But I support standards in information sharing, measurement, etc.

[frabjous]

Originally Posted by DawnFalcon

Short form of the arguments, copy/pasted from elsewhere (I'm not going to argue about the details since I don't 100% agree, and this was written in the CSS 2.0, not 2.1 era):

Perhaps you should give the citation.

I see some general (really pretty vague) criticisms of CSS in there, and I see a lot in there having to do with the inherent difficulty of implementing CSS consistently (though frankly it seems to me that these problems would exist with any complex style-based implementation), but I didn't see much to explain the incompatibility with WYSIWYG interface or friendly GUI editor in particular. (Sure, this is claimed, but no specific support is given for the claim.) So I still don't think you've answered my question. I appreicate the effort, though.

Out of curiosity, what do you dislike about CSS? Just that you haven't found an editor you like, or is there something in particular? E.g., the way a certain property is handled, or inheritance in general, or ...? I'd be pleased with one or two examples.

[DawnFalcon]

Website it's from is long gone.

And if you want to ignore it, your call. But it's perfectly specific enough, and indeed XSL (also from the W3C) avoids many of the same pitfalls, even sensibly using XML markup itself. It's also a good explanation of why the difficulty of using a natural formatting model and CSS's many inconsistencies - especially in duplicate tag functionality and it's inheritence model - lead to conflicting and incomplete implementations and there being no good WYSIWYG HTML+CSS editor.

If you're simply going to dismiss it like that, then I don't intend to continue the discussion with or (or any discussion on this forum again, for that matter). The primary evidence - lack of said editor - remains.

[frabjous]

I've been trying to understand your position. This is why I keep pressing you for details. I don't dismiss your concerns; I'm just trying to understand them.

I personally have found the use of CSS quite straightforward and convenient. So it's not as if I am basing my attitude on nothing. It's hard for me to think that's it's the mess you suggest when I don't have any concrete reason to think there's a problem. Except in IE, inheritance pretty much always works as I would expect. Putting a style attribute inside an individual occurrence of a tag rather than using a class or defining a style applicable to all given tags of a given sort seems like sloppy usage to me, but I figured that was a concession for the ease of WYSIWYG editing.

What did you think of the kind of GUI for CSS creation I described earlier? Is that not possible? Is it not sufficient for most users?