quirks mode and IE5 vs IE6

[Jeff Thies]

Well, it's probably past time for me to regularly include doctype in my
html.

What is quirks mode and how do I avoid it. I seem to recall a specific url
is required.

What is the suggested doctype for html4.0 transitional? That's probably what
I'm writing.

Also, what's the differences between IE5(windows) and IE6?

I don't have IE6, but usually what I write looks very similar in IE5, Opera7
and NS7.1. Sometimes IE5 mac looks a bit different although Safari is good.
What do I need to worry about in IE6?

Cheers,
Jeff

 

[Neal]

What is the suggested doctype for html4.0 transitional? That's probably
what
I'm writing.

See http://www.w3.org/TR/html401/struct/global.html#version-info

 

[David Dorward]

Well, it's probably past time for me to regularly include doctype in my
html.

Umm... very much so.

What is quirks mode and how do I avoid it. I seem to recall a specific url
is required.

Quirks mode is when a browser emulates the stupid mistakes of its
predecessors so code written to love their errors doesn't fall apart.

http://gutfeldt.ch/matthias/articles/doctypeswitch.html

What is the suggested doctype for html4.0 transitional? That's probably
what I'm writing.

HTML 4.01 Transitional isn't suggested, for modern webpages HTML 4.01 Strict
is the most apropriate.

http://www.w3.org/QA/2002/04/valid-dtd-list.html

Also, what's the differences between IE5(windows) and IE6?

IE6 has fewer bugs and fewer massive secuity holes.

 

[Martin Honnen]

Jeff Thies wrote:

What is quirks mode and how do I avoid it. I seem to recall a specific url
is required.

Also, what's the differences between IE5(windows) and IE6?

See

http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnie60/html/cssenhancements.asp
for details on IE6 and its two different rendering modes

 

[Jeff Thies]

What is quirks mode and how do I avoid it. I seem to recall a specific
url

Quirks mode is when a browser emulates the stupid mistakes of its
predecessors so code written to love their errors doesn't fall apart.

http://gutfeldt.ch/matthias/articles/doctypeswitch.html

This is what I've summarized. Let me know what I have wrong.

Quirks mode in IE6 and fullstandards mode in Mozilla (NS7 also?) is turned
on by include a full HTML4 doc type *with* a url. There is no full standards
mode in IE5.

IE6 quirks mode is mostly about nesting boxes and inheriting border widths
and padding.

Mozilla quirks mode is about emulating NS4!

There is no rendering difference between HTML4 strict and loose at present.

Now if I can just get rid of the occaisional center tag I have to use in
IE5, I'd be HTML4 strict!

Cheers,
Jeff

 

[Steve Pugh]

This is what I've summarized. Let me know what I have wrong.

Quirks mode in IE6 and fullstandards mode in Mozilla (NS7 also?) is turned
on by include a full HTML4 doc type *with* a url.

Not quite. With HTML 4.01, all Strict Doctypes trigger standards mode;
Transitional doctypes with a URL trigger standards mode; Transitional
doctypes without a URL trigger quirks mode.

Recations to HTML 4.0 and XHTML 1.0 doctypes are slightly different.

There is no full standards mode in IE5.

Correct, there is no doctype sniffing in IE5.x. It only has one
rendering mode, and that is what is being simulated in IE6 quirks
mode.

IE6 quirks mode is mostly about nesting boxes and inheriting border widths
and padding.

Mostly, yes. It also determines other things, such as whether
font-size: medium; is taken as the browser default or one size larger.

Opera 7 also uses doctype sniffing. Its quirks mode is an attempt to
emulate IE behaviour; hence it introduces bugs that Opera 6 did not
have.

Mozilla quirks mode is about emulating NS4!

No. See http://mozilla.org/docs/web-developer/quirks/quirklist.html

The quirks are mostly minor details of CSS and HTML that most browsers
have historically gotten wrong.

There is no rendering difference between HTML4 strict and loose at present.

This is why doctype sniffing is bogus - the rendering mode has nothing
to do with the HTML code used, just with what label is stuck on top of
it.
For example: The same code can validate as both Strict and
Transitional (as Strict is a subset of Transitional).
If the doctype is changed from <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML
4.01 Transitional//EN"> to <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML
4.01//EN"> then the rendering mode changes from Standards to Quirks.

Now if I can just get rid of the occaisional center tag I have to use in
IE5, I'd be HTML4 strict!

Centering with CSS is generally possible in IE5, though it does mean
taking advantage of a bug in IE's CSS support. What's the exact
problem?

Steve

 

[GD]

Correct, there is no doctype sniffing in IE5.x. It only has one
rendering mode, and that is what is being simulated in IE6 quirks
mode.

That applies to IE for Windows but IE5 for Mac does Doctype sniffing (I
think it was the first 'big' browser to do so?) and will get the box
sizes right in strict mode. It also supports the proposed CSS3 property
to select the method of calculating box sizes for each selector.

Although quirks mode in IE6 is supposed to make it compatible with IE5,
I have found that this is often not the case, at least not for IE5.0 on
Windows. I've often tested pages in IE6 in quirks mode and they looked
ok, but when IE5.0 loaded them they *seriously* broke. IE6 seems to
simulate basic things like the broken box model but there are other IE5
nasties it doesn't mimic.

I've never been motivated enough to look into it in detail. I just hope
people have learned from the nightmare of trying to support NS4 for so
long that it's not worth the hassle, especially when the users are just
too lazy to upgrade to IE6 or a better browser. I doubt anyone can cite
resource issues as the reason for sticking with IE5.

As has been the case with NS4 (or even still is for some), as long as
sites keep working with IE5 why will people ever be motivated to
upgrade? From their point of view there's no problem with IE5 - it
works.

 

[DU]

Well, it's probably past time for me to regularly include doctype in my
html.

What is quirks mode and how do I avoid it. I seem to recall a specific url
is required.

What is the suggested doctype for html4.0 transitional? That's probably what
I'm writing.

Also, what's the differences between IE5(windows) and IE6?

MSIE 5.x for windows does not implement any kind of differentiated
rendering mode. The good news is that with time people do upgrade their
browser. MSIE 6 for windows beta 1 was released in march 2001 IIRC.

I don't have IE6, but usually what I write looks very similar in IE5, Opera7
and NS7.1. Sometimes IE5 mac looks a bit different although Safari is good.
What do I need to worry about in IE6?

Cheers,
Jeff

I recommend you always use a strict definition because in all browsers
supporting standards compliant rendering mode, a doctype declaration
with the strict definition always trigger standards compliant rendering
mode.

DOCTYPEs that will trigger standards compliant behavior in Opera 7, IE6
for Windows, and Netscape 7:
http://www.opera.com/docs/specs/doctype/

There are other related benefits to using a strict definition.
The most important benefits you get when triggering standards compliant
rendering mode in MSIE 6 for windows are
- correct implementation of the CSS1 box model: MSIE 6 in backward
compatible rendering mode still incorrectly implements the CSS1 box model
- faster parsing and faster rendering of pages (assuming your markup is
error-free: best is to use the W3C validator)
- rendering is closer, more consistent to what other W3C web standards
compliant browsers render and this should stay that way or get further
improved with time as browser manufacturers are aiming to support web
standards and they work on eradicating browser bugs

DU

 

[Jeff Thies]

Although quirks mode in IE6 is supposed to make it compatible with IE5,
I have found that this is often not the case, at least not for IE5.0 on
Windows. I've often tested pages in IE6 in quirks mode and they looked
ok, but when IE5.0 loaded them they *seriously* broke.

Like how? Overwritting content?

IE6 seems to
simulate basic things like the broken box model but there are other IE5
nasties it doesn't mimic.

I've never been motivated enough to look into it in detail. I just hope
people have learned from the nightmare of trying to support NS4 for so
long that it's not worth the hassle, especially when the users are just
too lazy to upgrade to IE6 or a better browser. I doubt anyone can cite
resource issues as the reason for sticking with IE5.

As has been the case with NS4 (or even still is for some), as long as
sites keep working with IE5 why will people ever be motivated to
upgrade? From their point of view there's no problem with IE5 - it
works.

Well you know, we have to support the major browsers. I was just recentlay
able to give up on NS4. Last year I had to download. NS4.5 in addition to
NS4.7 because it worked differently in that browser!

I'd love to give up on IE5, any idea what the IE5 market share is?

Is IE5 Mac closer to IE6 windows or IE5 windows?

You know how tough it is to have two versions of IE on a windows box!

Jeff

 

[Toby A Inkster]

I'd love to give up on IE5, any idea what the IE5 market share is?

In my experience, about 5% and IE 5.5 at about 7%, but both are falling.

 

[GD]

Like how? Overwritting content?

Positioning elements in the wrong place or rendering the wrong width are
the main things. I've noticed that on some pages if I use absolute
positioning and set the element width to 100% then IE5 doesn't equate
that with 100% of the window. If there isn't enough text in the main
area to go accross the screen and wrap then the positioned element ends
up as wide as the text. There should be no relationship between a
positioned element and the length of a line of text in a normal element,
especially when the main box isn't a parent of the positioned box.

Well you know, we have to support the major browsers.

True, but there's a difference between making sure the site works ok,
and worrying about evey last pixel. For instance, adding padding and
margin to inline elements can be very useful, especially on navigation
bars where you want to space out links but not use a table. IE5 can't
render the padding or margin so the links will be closer together, but
it hardly breaks the site. If the user wants it prettier they should
have the decency to upgrade. I'm not even wanting people to change to an
alternative browser like Mozilla or Opera, just a version of their
'free' browser that came out in the last 3 years!

Is IE5 Mac closer to IE6 windows or IE5 windows?

Closer to IE6 but better They're totally different layout engines so
IE for Mac has its own set of bugs. Floating is broken but in a
different way from IE5 on Windows and it doesn't seem to combine line
height, padding and margin correctly on inline elements. But apart from
that it renders closer to Opera 6 than any other browser, at least in my
experience. Maybe it's different with tag soup It does fixed
positioning, child and adjacent CSS selectors and even allows the
contents of positioned elements to break out without clipping them,
something IE for Windows does to annoy me, personally!

 

[Eric Bohlman]

I've never been motivated enough to look into it in detail. I just hope
people have learned from the nightmare of trying to support NS4 for so
long that it's not worth the hassle, especially when the users are just
too lazy to upgrade to IE6 or a better browser. I doubt anyone can cite
resource issues as the reason for sticking with IE5.

If "anyone" means "any autonomous individual in complete control of his
computer setup" you're probably right. But consider a large corporate
setting with tens of thousands of users all using IE5. In that case, an
upgrade is a pretty massive undertaking, especially considering all the
Intranet application that would have to be tested with the new browser to
make sure they don't break. There's a good reason why large organizations
seldom have bleeding-edge technology.

 

[Karl Smith]

That applies to IE for Windows but IE5 for Mac does Doctype sniffing (I
think it was the first 'big' browser to do so?) and will get the box
sizes right in strict mode. It also supports the proposed CSS3 property
to select the method of calculating box sizes for each selector.

Opera 7, and I believe Mozilla too, implement the "box-sizing"
property. IIRC, it is no longer officially a CSS3 proposal, more
recent CSS3 drafts saying the need for such a property is obviated,
now that "all major browsers" implement the CSS box model similarly.

 

[=?ISO-8859-1?Q?Andr=E9s_Sedano?=]

Well, it's probably past time for me to regularly include doctype in my
html.

What is quirks mode and how do I avoid it. I seem to recall a specific url
is required.

What is the suggested doctype for html4.0 transitional? That's probably what
I'm writing.

Also, what's the differences between IE5(windows) and IE6?

I don't have IE6, but usually what I write looks very similar in IE5, Opera7
and NS7.1. Sometimes IE5 mac looks a bit different although Safari is good.
What do I need to worry about in IE6?

Cheers,
Jeff

May I ask you why don't you use IE6? When I have to use IE5 in some
client's office, I suffer a lot. It renders incorrectly lots of pages
and I can't stand that... I can't understand why people doesn't use IE6
if it's safer, faster, more compatible with standards...

So, please, can you tell me why are you still using IE5?

Or, if anyone has reasons, I'd greatly appreciate if you share them with me.

Cheers,
Andrés

 

[Jeff Thies]

May I ask you why don't you use IE6? When I have to use IE5 in some
client's office, I suffer a lot. It renders incorrectly lots of pages
and I can't stand that...

Gee, could that be it?

I can't author something that may be badly broken in a mainstream browser,
even if it is that browsers fault.

I can't understand why people doesn't use IE6
if it's safer, faster, more compatible with standards...

By the same token, why use IE at all? If I'm browsing for my own purposes, I
don't use IE.

So, please, can you tell me why are you still using IE5?

Because my clients would go balistic if 18% of their audience had serious
problems with their site and I wasn't even aware of it! It's really the
difference between authoring for yourself, and authoring for a living. If
site appearance is important, and it is for my clients, this goes triple.

I'm just glad I don't have to support NS4 anymore!

Cheers,
Jeff

 

[David Dorward]

Because my clients would go balistic if 18% of their audience had serious
problems with their site and I wasn't even aware of it!

That is a good reason to test pages in IE5.

It leaves two questions. Why do you not test your pages in Internet Explorer
6? Why do you not use IE6 for normal use (or better yet, a decent browser
like FireFox)?

 

[Eric Bohlman]

Because my clients would go balistic if 18% of their audience had
serious problems with their site and I wasn't even aware of it! It's
really the difference between authoring for yourself, and authoring
for a living. If site appearance is important, and it is for my
clients, this goes triple.

And it's important to remember that many viewers may be using IE5 because
it's what's installed on the machine that they're using and because the
machine isn't their *personal* machine, they *can't* upgrade the browser.

 

[Jeff Thies]

So, please, can you tell me why are you still using IE5?

That is a good reason to test pages in IE5.

It leaves two questions. Why do you not test your pages in Internet Explorer
6?

Probably because I do no want to dual boot and I know what it looks like
elsewhere.

Why do you not use IE6 for normal use (or better yet, a decent browser
like FireFox)?

I rather like NS7.1. Not familiar with FireFox.

Cheers,
Jeff

 

[Barry Pearson]

Jeff Thies wrote:
[snip]

I rather like NS7.1. Not familiar with FireFox.

Try it - with the Web Developer Toolbar. The latter is super. (I wouldn't use
Firefox for general browsing).
http://chrispederick.myacen.com/work/firefox/webdeveloper/

But perhaps you already use the Mozilla version of the toolbar, that works
with Netscape 7.1? Also super.

 

[Alan J. Flavell]

And it's important to remember that many viewers may be using IE5 because
it's what's installed on the machine that they're using and because the
machine isn't their *personal* machine, they *can't* upgrade the browser.

Sure. But conversely, it may be important to remember that there will
be viewers using IE6 because _that's_ what's installed on their
machine. So designing a page "to" the bugs of IE5 (which might have
been corrected in later versions, but if the author doesn't have one
then he won't necessarily know that) is not the answer, either.

Unfortunately, it's a lot of fuss to maintain more than one version of
Win IE. In fact, due to security concerns we get into trouble if we
run old versions of IE on our network: updating to the latest MS
fixes isn't an option - it's a requirement, in this context.