tomhudson's Journal: Every browser is *still* broken.

After 15 years, we still don't have an un-b0rked browser. CSS 2.1 was done in 1997, and yet firefox, opera, chrome, arora - they all render differently for non-trivial layouts.

15 years, and they still can't get the basics right. It means that the problem is not the implementation, but the underlying concepts that are flawed in fundamental ways.

And there's no blaming Microsoft or Apple for this fiasco.

No, we did this to ourselves. We're all suckers. The people setting the standards did it wrong, and we didn't immediately stone them to death, salt their fields, enslave their families for the next 3 generations, and all that other "Carthage must die!" goodness.

So we have let ourselves become slaves to stupidity.

What a waste of time, energy, brain cells, and just general aggravation. Have fun with html5 + css3, folks - you'll never see it finished in your lifetime, not even if you live for another 100 years.

Apple has it right - apps, not a stupid one-size-fits-nada web browser. Just like they have it right about not releasing stuff until it's good and ready.

Stupid browsers. Stupid us.

Re:

[Barbara, not Barbie]

I finished up a project this weekend, all the little extra tweaks that make it look really polished (no, it's on a domain you've never heard of).

So, while I had been checking it in other browsers, updates just broke it in both Opera and Chrome.

I could have been spending the last month doing something more productive - like finding some paint to watch dry, or vegging out in front of the TV.

And of course, now there's a story on the front page about how Google is yet again trying to push the web browser

Re:

[kermidge]

I do wish you'd tell us how you really feel. [grin]

Years back I tried learning basic HTML and the then-fairly-new CSS; I figured, cool, some basic tools just like in carpentry and mechanics, a set that'd act as standards to get the job done, and everybody'd play with the same stuff.

So enough damn site owners want to make things different, or want to use the tools to do "neat, gotta have it" things for which the tools weren't designed; then, of course, each browser maker had to decide how much and to what e

Re:You mad sis?

[Barbara, not Barbie]

Thx. If you look at the change-list, it's obvious that a lot of the problems derived from the original spec not actually specifying something or doing so in an ambiguous fashion, so people either just assumed. But that doesn't fix problems like not being able to collapse whitespace between tags, and parsers not meing smart enough to realize that if there's only whitespace, it should be collapsed (if you wanted real single space, you'd use an &nbsp ; character entity). I put on my thinking cap, and came

Re:

[Bill Dog]

I've adopted doing this:

<div class=foo>
......... some stuff goes here
</div><div
class=foo>
......... some more stuff goes here
</div>

I.e. split the line *within* a tag, within which ws is not rendered. Not that bad because closing tags are short (being attributeless), so the rest of the following opener on the next line is geographically in very close proximity, just vertically nearby instead of horizontally. And no superfluous markup is being introduced!

But yes it would be nice if so

Re:

[Barbara, not Barbie]

You can do that ..., it works too.

I like using custom tag names instead of div class=whatever, because it also makes it easy to find the closing tag in a long list of tags (plus on average it saves a few characters).

Example browser mess-up: fixed position of certain elements when you only specify one of the 4 of top, bottom, left, right. When fixed is specified, the object is supposed to be taken completely out of the flow.

Some browsers set the other ones to whatever it would have been with respec

Re:

[kermidge]

Wow, nice stuff. Wish I'd had this kind of clear thinking and exposition available at the time (maybe it was, I just never found it.) The z tag looks neat.

Re:

[Barbara, not Barbie]

Thx. Actually, I was thinking, and as long as there are no PRE tags in the mix, just running the page through a perl script to do a "s/\>\s+\ If you've ever seen web pages with links that have the underline extending past the actual text, they have the same problem - whitespace (such as newlines) between the opening and closing tags.

Re:

[Bill Dog]

I only have IE 8 and FF 3.6 here at home, and I couldn't get their rendering to diverge, but they both do it wrong as you describe, with fixed div inside fixed div, where for the contained div if you only specify a top, it takes padding-left and margin-left relative to the containing div, but if you also then specify a left for the contained div, it switches its horizontal origin to that of the viewport. Weird. Super-strange considering I'm using the XHTML Strict doctype, putting both browsers in their leas

Re:

[Barbara, not Barbie]

Chrome does it right, opera and ff mess up.

Then again, chrome does scaling wrong (I figured out a way around webkit's bug, but it's a bit ugly-fugly), so you can't win for trying.

That being said, I've spent most of the last 48 hours banging away (my eye is sore, but it's holding out .... a good sign, I guess :-) and I've managed to get everything working just right, and to complete what I set out to do - which is a good thing, because I had an idea a couple of weeks ago that might just work out.

What r

Re:

[Bill Dog]

Well if Chrome does it right, they changed it in the last 8 months or so, because Chrome v13 rendered my fixed positioning test the same as IE 7 and IE 8 and FF 3.6. Maybe/hopefully during the move to HTML5, browser makers are cleaning up their support of companion standards.

Re:

[Barbara, not Barbie]

Well, I'm using 18.0.1025.151 - but you might try using the html5 <!DOCTYPE html> DTD-less declaration instead - it should trigger the "do not emulate any quirks" behaviour. Unfortunately, FF still suffers ... which is a real annoyance.

lol

[Captain Splendid]

Just like they have it right about not releasing stuff until it's almost, but not quite, good and ready.

Apple are notorious public beta testers. Original release always sucks in at least one non-trivial way. The first point release is always swift and much closer to a finished product.

Re:

[Barbara, not Barbie]

All I know is that everyone I know is switching en masse to Apple, because everyone I know who has already switched is very, very happy.

Maybe they've been lucky, but nobody seems to want a Windows machine any more, nobody wants an Android smartphone, nobody wants an Android or Windows tablet. It's all iPhones and iPads and iMacs.

If Microsoft thinks that being able to create "crapplications" in Windows 8 using html and css is going to change that, they're wrong - it's already at least 2 years to late.

Re:

[jawtheshark]

Indeed, and if you dare to tell an Android fan that he needs to stop thinking outside his bubble and look at what's happening in the world, you get gems like "but I am a nerd, I think like a nerd and I'm supposed to". I'd say, no, as a nerd you're supposed to have a certain amount of intelligence (I won't dare to say empathy), so you can see that what is good for you isn't for everyone else.

When I bring up the (manufacturer) support issue on iOS versus Android they tell me that you can upgrade a 3 year old

Apple and Microsoft both had a hand in this mess

[plover]

Apple didn't start out right either. They didn't want apps on the iPhone. They wanted everything to be web based. And it was universally seen by geeks as a stupid idea, but Apple didn't care. They did what Steve wanted.

But what Apple does best is learn from their mistakes, and they always deliver what their least competent users want. So Apple did a 180 and started promoting the hell out of apps. They did it out of greed (the app store is damn profitable, after all); coincidentally, it was technically