W3C On How To Fix Browsers

kellan1 writes: "The W3C has published a note Common User Agent Problems, aka why do web browsers suck, and what can you do about it?"

Re:Mozilla

Keep the TEXTAREA, add a new one <FORMATTEXT> , or something similar.

Just a quick correction: Use TEXTAREA, but add an option to it, eg. <TEXTAREA FORMAT=HTML>. That way old browsers still support it, only the formatting is missing. Also the format would be extensible.

Re:Don't hide 404 messages!

"This user violated our Acceptable Use Policy and has had their account terminated. The page you are looking for is gone for good."

In that case, the proper status code would probably be 410 Gone [w3.org]

netscape!

im a web designer currently living up in san jose. on my way to work at our clients site, i pass netscape's headquearters every day.

we take great joy in yelling a different fix for their broswer every day out our window;
"how about rendering tables correctly?!"
"how about filed widths being consistent?!"

id be happy to yell any fixes anyone would like to see implemented.
oh the headaches of netscape. keeps me with a job though.

Re:How about following the DTDs?

The problem is that HTML was originally intended to be a relatively abstracted specification of the content, but this wasn't what people wanted. TABLE tags weren't forced on people by browser makers, but were embraced enthusiastically by web site creators. Basically the mistake was thinking that content creators wouldn't want to control how that content was presented to the user. Of course standardization is a good thing, but people need to admit that the intention of HTML has changed from a way to specify text abstractly (like DOCBOOK), to a way to specify a layout for a page in a flexible and robust manner.

--

Slashdot's text area

Would be immensely more useful, in spite of the abyssmal formatting, if it was larger than 50 characters by 10 lines. I don't know if this puny size is a conspiracy to keep comments short and pointless or some vague attempt to keep it "useful" for Lynx or 640x480 display resolutions. Either way, it just seems kind of ridiculous.

Re:hmmm

I guess it's biting and insightful then, because Netscape6/Mozilla have better support for the basic W3C standards (HTML, CSS, DOM) than WinIE. Check out www.richinstyle.com or some other independent site if you don't believe it.

Re:What I find completely amazing...

You're wrong for a number of reasons.

There is no definition of how tags are supposed to be rendered. That is explicitly left open by the HTML standard, and for good reason. CSS specifies more of that, but it's still not complete, and again, there are good reasons not to fully specify the rendering.

It is simply not true that the rest of the work is "a piece of cake", not when you're dealing with something as complex as the W3C standard definitions and all their interactions. If you don't believe it, try writing a browser yourself.

But the biggest problem is that despite the fact that HTML 3.2 and 4.0 are specified, it doesn't matter because Web page authors DO NOT stick to the standards. They write buggy pages which more or less render OK in the browser they happen to be using, and then they're done. There are almost no pages which adhere strictly to the W3C definitions, and that's why results vary from one browser to the next.

I think W3C missed one...

When the user agent clicks on a broken link, return the user to their current page and inform them of the error. How many times have I clicked the BACK button because nobody has put this in? How hard can it be?

w3c's guidlines

The W3C does have a tool you can use to see if web pages are compliant with their html specifications. Which, of course, almost no one's are.

As far as I am concerned, if this guy's [helsinki.fi] web page is not html compliant, I am not going to worry if mine isn't.

hmmm

This is not a Microsoft document.

the behaviors described are not protocols officially accepted by MS. (just look at the behavior of the browsers)

Given the dominance of MS in the market, is this document even relevant? [even though it is brilliant, insightful, and written by people who care about what is going on]

I am just glad we haven't progressed to the point where Microsoft "red" is a shade between black and blue.

Re:Good luck to them

I know, I wish Slashdot had the same type of WYSIWYG editor that my weblog [editthispage.com] does. It uses Manila [userland.com], from Userland [userland.com], to allow WYSIWYG in IE 5+ for Windows, at least. I know that isn't the Slashdot audience, but I'm betting 30% of the hits here come on IE...
---

A good solution

1. Have the web browser render correct pages correctly, as the #1 priority.

2. Have the web browser try to compensate as well as possible for mistakes, as long as doing so doesn't interfere with #1, above.

I.e. follow the Internet maxim, "be conservative in what you do, and liberal in what you accept"

3. Issue a web page quality feedback to the user so if a site has bad HTML, the user knows, so she can fix the site if it is hers, let the web master know if someone else is responsible for the site, know that a company can't do web pages right (great if you are browsing a web page design company's site! ;) or at least know that the reason the page looks bad is because of the web page being poorly done, and not a browser issue.

A good implementation would be, for example, an icon which shows quality, and when clicked shows any errors in the page. E.g. if the page is good there would be a smiley face and a tool tip would say "No errors - high quality page", and if the page was bad there would be a frown face and when you clicked it you'd get a window opening with a list of the errors in it. A really good implemenation would have a whole site of icons for perfect, good, ok, bad, and horrid HTML. Of course, the lower the quality, the more likely there are to be problems rendering the page...

How about following the DTDs?

Browsers started going to hell in a handbasket when they forgot why HTML was around in the first place - to make a platform-independent system for sharing information. Thus, a web page in Netscape *should* look different than a web page in IE, *however* the content should be the same.

The DTD merely says that this text is in a paragraph. Unfortunately, most browsers have embraced and extended this to assume that all browsers have the exact same layout. Thus, changing font sizes or types in your browser makes the page look just plain wrong.

Back when the DTD was being followed, *everyone* built web browsers, and all was good with the world. The content was similar, and no matter what the platform, you could still browse. Then came . And . And and all hell broke loose.

Now we're in an IE world. One browser for everyone. Netscape is flailing, Mozilla is close, but MS has free run of the DTD.

If you really want browser wars to heat up, you have to make usre that the browser followed the DTD properly so the display is not driven by the content, but is driven by the end user, as it should be.

Re:What I find completely amazing...

...then there is no reason that the page should not display identically on each persons browser. I find it unfathomable that this is not the case.

There are many reasons why two correctly-functioning browsers will display the same page differently. The <p> tag, for example, without style information, just means "paragraph" - there's no reason different browsers might not have (compiled-in or user configured) different default fonts, default spacing between paragraphs, default paragraph indentation, etcetera. Even with style information, the user can override the author's preferences.

Web pages are not Postscript or PDF documents. HTML authors who try to make pages that look exactly the same in all browsers Just Don't Get It.

Tom Swiss | the infamous tms | http://www.infamous.net/

Good luck to them

Here I am, typing into a <TEXTAREA> , a widget so abhominably broken, it only understands the barest rudiments of text editing (hit key, print letter), and they're worried about broken or missing <A> tags.

Come on, people... one of the most common uses of the web these days is to post messages on a weblog-type site (like, oh, Slashdot) -- and there isn't a widget that we can use other than the <TEXTAREA> so normal people can type text in, highlight a few words, and hit a BOLD button? They have to learn to use <B> tags? What is this, 1983?

I'll tell you what -- when there is a web browser that is isn't brainsick, then I'll care about the UI implementation of broken <A> tags...

Bring back verbose loading!

Here's one I'm surprised to see didn't show up: verbose page loading. What I mean by that is have the browser actually tell you what it's doing as it goes through the process of loading a page.

I remember old browsers used to display information like:

• Resolving name..
• Contacting server
• Negotiating connection
• Downloading xxxx.html
• Closing connection
• Done.

The trend I've seen in modern browsers simply say "Loading" or "Opening" without telling me what's happening. Having the extra information would help when troubleshooting what section of the content isn't loading.

Re:w3c's guidlines

Why? What's so wonderful about Linus that his homepage should be the de facto standard for HTML? Sure he's done lots of great things in computing, but he's hardly an authority on HTML and web authoring standards.

A little less hero worship, I think, would serve well.

for you browser writers out there

If it matters, I personally consider this one to be the most important:

1.8 Provide a mechanism to allow authentication information to expire.

Many browsers allow configuration to save HTTP authentication [RFC2616, RFC2617] information ("remember my password"). They should also allow users to "flush" that authentication information on request. For instance, the user may wish to leave the user agent running but tell it to forget the password to access the user's bank account.

Wrong: Most user agents consider that authentication information (e.g., password) provided by a user for a server/realm pair during a session is immutable for the duration of the session.

I don't think I'm the only one that finds it quite annoying to have to exit and restart my browser in order to make it forget my HTTP authemtication information. I believe Netscape and IE both have this problem.

All your dangifiknow [dangifiknow.com] are belong to us.

Other useful changes

More flexibility in caching algorithms - for example, cache images but refresh content would speed many browsers. IE5 caches based on time periods, which are either too long or too short for many sites. Allow users to define behavior of typing in a name into the url text box. Some browsers assume that this means that you want to search, or go to a database such as realnames. Defaulting to www..com (or .org or whatever) would simplify and speed browsing.

Simpler browsers

These days, your browser is a tool you use with your computer, in the same way a text editor or ftp client has been for a while. This is especially true for many users of free software, since the documentation is, by and large, on the web. Browsers, on the other hand, have grown much like a microsoft product -- more complex, more bloated, and with more features rather than simple and functional. I would love to see a browser with the html rendering abilities of mozilla or internet explorer, but without all the other functions. Such a browser might still need frames support, and possibly javascript, but it wouldn't need to be your chat program, your html editor, or your kitchen sink. It seems like most browsers fall on either end: lynx which is stable but which can't see many of the sites which are written these days, and mozilla which takes up a lot of RAM and does everything you'll ever need to do on your computer.

-thinmac

Re:How about following the DTDs?

Content is not the most important part of a website...if either are lax your website is useless.

Baloney. This is the information age, not the presentation age. I never go to a site to see what it looks like today. I want to find out what information is there today. If I'm interested in a 'user experience' I'll watch the commercials during the Super Bowl.

If the presentation is poorly done (the definition of 'poorly done' is different for different people) then it might be useless to some people. If there is no or low quality content then the site is useless to everyone, except perhaps to those who visit sites just to see what they look like.

Consider how cleverly crafted presentation will be lost on someone who is blind, color blind, epileptic, deaf, accessing the sight from a mobile phone, or some low bandwidth connection, etc.

Most people don't go to the internet just to look at the pretty pictures.

--

Re:netscape!

How about not refusing to draw anything because an element isn't closed?

This is part of the reason that web browsers suck: because people demand that they render broken content "nicely". I would much rather they render correct content properly, than do a half-assed job of rendering everything. "Best guessing" is precisely what leads to the non-deterministic behaviour the original poster complained of.

If malformed content doesn't show up correctly, it's the author's fault, not the browser's.

The browser wars at at fault.

We're always arguing... "NS supports the standards." "IE supports the standards." Well, the thing is... it was a rush to become more popular with the standards that your browser introduced. In the case of IE vs. NS, the two biggest players when a lot of our standards were being formed, IE introduced the features that made webpages more flexable and dynamic. Netscape stagnated however, not really coming up with anything new. Internet Explorer won out and the W3C had to conform to it. And this is not necessarily a bad thing. Think about how far behind web technologies would be if NS was the dominant browser? Sure, NS and IE can do a lot of the same things - problem is, IE does them more cleanly whereas equivalent NS implimentations are essentially big ugly hacks. And well, of course, there's a plethora of useful features that are in IE that you could never do with NS. And when is the last time NS introduced anything new? *pfft*

What makes browsers suck now? Everyone who makes a browser is very reluctant to follow the path of the ones who won the war. I can think of a few examples - specifically a few that IE introduced to CSS standards that Netscape rejected, Mozilla implimented then broke or removed, and no other browsers will do. I'm referring to the :hover property. Is this feature a bad thing? Absolutely not! Better than making a big hacky JavaScript solution to do the same thing. But will any other browsers support this? No. That's just stupid - and it's not the only example.

Well, this post may jump around a lot and lack consistency, but I'm rushing to the point and I think the point is clear. Browsers suck because web developers want to use new technologies being developed by companies who build web browsers that introduce them. Then the other browsers refuse to adopt the new technologies - in several cases simply because the authors hate the company *coughMicrosoftcough* that introduced them. That's just stupid.

Re:hmmm

This would be a biting and insightful comment if Internet Explorer didn't have the most comprehensive support for W3C standards of any browser in existence.

Re:How about following the DTDs?

Ooh....my tags got taken literally. Rats. Should read:

Then came <blink>. And <center> . And <font> and all hell broke loose.

What I find completely amazing...

is that so many browsers don't render pages correctly.

- There is a clearly articulated definition of what tags are supported
- there is a clearly articulated definition of what the tags are supposed to do and how they are to be rendered

My question is why do so many browser not render correctly? What is so hard? One of the hardest parts of programming is requirements definition for the software being written, and getting the customer/boss to clearly articulate what they want the software to do and how it should perform. Once you have the iron-clad requirements written down, the rest of the work is a piece of cake - it is implimentation and testing against the requirements.

In the case of HTML, where there are clear definitions for 3.2 and 4.0, a list of what is deprecated and what is supported, if web page authors stuck to the STANDARD and not use browser specific tags, then there is no reason that the page should not display identically on each persons browser. I find it unfathomable that this is not the case.

Re:Head up its ass.

This attitude about Web design is one of the reasons the Internet went broke. (another one is lame, hopelessly flawed business plans)

Web designers need to remember that Web design is not print. Documents will be changed by users and users should have that freedom. Worry more about information architecture, it's the content that ultimately matters. Design shouldn't just be decoration that viewers have to conform to, design should help viewers to better comprehend the information they're seeing.

What about handicapped users? Or those who need larger type to see? What about celphone browsers? PalmOS? WebTV? Crappy WindowsCE appliances in the airport business lounge?

Web design extends traditional design towards architecture and engineering. Good Web design is flexible. It doesn't matter what the building looks like if it falls on your head.

Strict adherence to standards is the best thing we've got. Letting go of the bells and whistles is better for your clients, your audience, your bottom line and your sanity.

I dont understand why a browser hasnt done this

What I dont understand is why browsers don't have better support for downloading and resuming of files. Why do programs like GetRight and Gozilla have to exist, catching clicks in browsers and downloading for us? And it would be twice as good if the browser could crash, and your download window keep downloading. I personally run mozilla and I'll use wget for huge files and resuming. It would be nice if I could do it right in mozilla.

Re:hmmm

Now don't get me wrong, I'm a mozilla supporter, but it really pisses me off (and wastes time) everytime someone says that, because this is what happens:

• Someone says IE/mozilla is the most standard compliants brower
  
• Someone else says that mozilla/IE is the most comliant
  
• Someone quotes a webpage that confirms their view of IE/mozilla being the most compliant
  
• Someone accuses that page of being written by IE/mozilla supporter/bigot and posts a page to another unbiased view showing that mozilla/IE is the most compliant
  
• Someone else discovers that that page is written by a company or supporter of mozilla/IE, and posts their own rant, quoting pages where IE/mozilla won't work.
  
• Rant about pages where IE/mozilla won't work.
  
• Name calling starts.
  
• The sane people continue to surf on, and accept (grudgingly) the fact that no matter how standards compliant something is (or isn't) there are still pages that won't render right, and until everyone gets to the same point of standards compliance/un-compliance, anyone writing web pages is still hooped as far as making everything look the same for everyone
  
• If we're lucky, everyone reverts to tableless pages with grey backgrounds.
  

Re:Bring back verbose loading!

Actually, most users are confused by verbose details. If you present the information, they assume it must be important and that they should understand it. "It says it's 'negotiating connection.' Is that good or bad?" If you tell the user to ignore the messages, you're reinforcing the perception that computers are very complex and the user isn't really smart enough to use them. Hiding the unnecessary complexity makes the experience more comfortable for the user.

That said, having an option "[ ] Show me details when downloading a page" would be great for those of us who can use the information.

Re:...and braille?

Braille? How's that supposed to work? Little pimples raise up on the screen?

...pimples on a screen reader, actually. Most screen readers plug in the serial port and display one or two horizontal lines of text. Linux supports this for the console, in fact SuSE 7 automatically runs its installer in text mode if it detects a braille screen reader.

More advanced software for Windows writes what line is currently under the mouse pointer, provided it's text and not a graphic.

Re:hmmm

Which is nice and good, but the problem is now the web is not driven by content any more. If the web were driven by content, (as a poster above notes) it wouldn't matter that netscape didn't follow w3c standards perfect, or if IE actually was the best browser in the entire universe, because I'd still be able to view every page I wanted to, and get the content therein.

Too bad now that 99% of the surfers out there don't care about content, but instead want their flashy, bullshit "user experience" to make it easier for them to read the bullet points of the information they're looking for. When was the last time you actually *read* a page full of content, that wasn't marked up to hell and back. I'm not saying everything out there should be block text for pages and pages, of course :) Just that that sort of page is slowly dissapearing.

Sometimes I *do* want the web to go back to the netscape 1.0 days (tables! wow!) where everything was grey (well, 1.0 had bgcolor I guess, so pre-1.0 days) when you surfed the web for information, not "experience". If I wanted experience back then I'd go outside and take a walk, watch a movie, or whatever.

Don't hide 404 messages!

It's dangerous to second guess the web designer. The 404 page may contain useful information for the user. Sure, most pages unhelpfully state "404 File Not Found" and little more, but it's possible for a page to be much more helpful. A site could have their 404 page automatically do a search to find the requested resource. Here is a good example at Wizards of the Coast [wizards.com]. A site could present a list of resources the site does, since one of them is likely to be helpful. Perhaps the page is gone because the client's account was terminated. I'd like to receive a message like "This user violated our Acceptable Use Policy and has had their account terminated. The page you are looking for is gone for good."

Saving content

I found section 1.3 interesting (``Allow the user to retrieve Web resources even if the browser cannot render them''). In my experience, it's often not the fault of the browser maker, but the site designer. How many times have you tried to view a video or audio clip, only to be diverted away because the javascript on the web pages can't tell if your browser has a plugin for the media type (as if you really need a plugin in the first place...)
--