the doctype sniffing mess is getting messier still

1 view
Skip to first unread message

Bertilo Wennergren

unread,
Jul 2, 2002, 10:38:59 AM7/2/02
to
As we all know Doctype sniffing turned out to be a bad idea, or actually
a %&(!#& mess. But now it's getting even worse:

Mozilla is about to introduce a third rendering mode besides quirks mode
and standards mode. It will be called "almost standards mode". A special
select list of Doctypes will trigger "almost standards mode".

This is not fun, but you can read all about it here:

<http://evolt.org/article/New_DOCTYPE_sniffing_in_upcoming_Mozilla_releases/1/32271/index.html>

--
Bertilo Wennergren <http://www.bertilow.com> <bert...@gmx.net>

wizard

unread,
Jul 2, 2002, 9:27:13 AM7/2/02
to

Not quite a new news

Bertilo Wennergren

unread,
Jul 2, 2002, 11:49:42 AM7/2/02
to
wizard:

> On Tue, 02 Jul 2002 14:38:59 +0000, Bertilo Wennergren
> <bert...@gmx.net> wrote:

>>Mozilla is about to introduce a third rendering mode besides quirks mode
>>and standards mode. It will be called "almost standards mode". A special
>>select list of Doctypes will trigger "almost standards mode".

>><http://evolt.org/article/New_DOCTYPE_sniffing_in_upcoming_Mozilla_releases/1/32271/index.html>

> Not quite a new news

So where did you read about it before?

wizard

unread,
Jul 2, 2002, 9:59:05 AM7/2/02
to
On Tue, 02 Jul 2002 15:49:42 +0000, Bertilo Wennergren
<bert...@gmx.net> wrote:

>So where did you read about it before?

mozila's site

http://www.mozilla.org/docs/web-developer/quirks/doctypes.html

Bertilo Wennergren

unread,
Jul 2, 2002, 12:07:44 PM7/2/02
to
wizard:

> mozila's site

> http://www.mozilla.org/docs/web-developer/quirks/doctypes.html

Indeed. There it is. And that piece of information was added to
that page on Jun 28. That was four days ago. So I'd consider it
news.

wizard

unread,
Jul 2, 2002, 10:25:52 AM7/2/02
to

Almost standards mode was created after 1.0 and 1.1alpha, but before
1.0.1 and 1.1beta. Before almost standards mode was created these
doctypes triggered full standards mode..........

So probably we should expect this thing since it is not implemented in
1.0 e.g.......... well it is a news ;)))))

Bertilo Wennergren

unread,
Jul 2, 2002, 12:38:17 PM7/2/02
to
wizard:

> On Tue, 02 Jul 2002 16:07:44 +0000, Bertilo Wennergren
> <bert...@gmx.net> wrote:

>>Indeed. There it is. And that piece of information was added to
>>that page on Jun 28. That was four days ago. So I'd consider it
>>news.

>>Bertilo Wennergren <http://www.bertilow.com> <bert...@gmx.net>

> Almost standards mode was created after 1.0 and 1.1alpha, but before
> 1.0.1 and 1.1beta. Before almost standards mode was created these
> doctypes triggered full standards mode..........

> So probably we should expect this thing since it is not implemented in
> 1.0 e.g.......... well it is a news ;)))))

Doctype sniffing was a bad idea from the start, and it's not getting
any better with more rendering modes added to the list, and the Doctypes
being split up into more groups. It's already complicated and bad
enough as it is now. We even have to know what bugs there are in the
sniffing routines!

There are people out there that have carefully built their pages taking
all the quirks in the sniffing routines into consideration. Some of them
might even have choosen a Doctype that used to give standards mode in
Mozilla, but that will trigger almost standards mode when new Mozilla
builds start to come out.

wizard

unread,
Jul 2, 2002, 10:53:46 AM7/2/02
to
agree, but let us see what the future will show ;))

Bertilo Wennergren

unread,
Jul 2, 2002, 1:18:44 PM7/2/02
to
wizard:

> On Tue, 02 Jul 2002 16:38:17 +0000, Bertilo Wennergren
> <bert...@gmx.net> wrote:

>>There are people out there that have carefully built their pages taking
>>all the quirks in the sniffing routines into consideration. Some of them
>>might even have choosen a Doctype that used to give standards mode in
>>Mozilla, but that will trigger almost standards mode when new Mozilla
>>builds start to come out.

> agree, but let us see what the future will show ;))

You mean when MS adds "Close to Standards Mode" (but only for the
Windows version - the Mac version will have to wait two years for it),
Opera adds "Extremely Standards Compliant Mode", and Mozilla counters
with the additional "Standards minus Box Model Mode", "Extra Quirky
Special Mode", "Tag Soup a la Carte", and "Classic Rendering Mode
Picante"?

Chris Hoess

unread,
Jul 2, 2002, 1:19:07 PM7/2/02
to
In article <ui3nj5a...@corp.supernews.com>, Guy Macon wrote:
>
> Another annoyance:
>
> In http://mozilla.org/docs/web-developer/quirks/doctypes.html
> you find:
>
> "The following trigger quirks mode [...] A DOCTYPE declaration
> that cannot be understood (e.g., no matching quote)."
>
> Which , if I am not mistaken, means that
><!DOCTYPE html PUBLIC "-//W3C//DTD XHTML Basic 1.0//EN"
> "http://www.w3.org/TR/xhtml-basic/xhtml-basic10.dtd">
> triggers Quirks mode.

Wrong, but only because the page is unclear; this refers to DOCTYPEs,
AFAICT, that don't conform with SGML production 110. (I.e., if the parser
can't break it into document type name, public identifier, optional system
identifier.)

Note under "Full Standards Mode" that it's triggered by "Any unknown
doctype."

--
Chris Hoess

Tim

unread,
Jul 2, 2002, 8:43:11 PM7/2/02
to
On Tue, 02 Jul 2002 16:38:17 +0000,
Bertilo Wennergren <bert...@gmx.net> wrote:

> There are people out there that have carefully built their pages taking
> all the quirks in the sniffing routines into consideration. Some of them

> might even have chosen a Doctype that used to give standards mode in

> Mozilla, but that will trigger almost standards mode when new Mozilla
> builds start to come out.

And I cannot resist adding: That those people will get what they
deserve (their broken pages, showing up as broken). ;-\

--
My "from" address is totally fake. (Hint: If I wanted e-mails from
complete strangers, I'd have put a real one, there.) Reply to usenet
postings in the same place as you read the message you're replying to.

Henri Sivonen

unread,
Jul 3, 2002, 1:05:19 AM7/3/02
to
In article <3D21BB03...@gmx.net>,
Bertilo Wennergren <bert...@gmx.net> wrote:

> Mozilla is about to introduce a third rendering mode besides quirks mode
> and standards mode. It will be called "almost standards mode". A special
> select list of Doctypes will trigger "almost standards mode".

Time to add a new column to the table on my doctype page.

> This is not fun, but you can read all about it here:
>
> <http://evolt.org/article/New_DOCTYPE_sniffing_in_upcoming_Mozilla_releases/1/
> 32271/index.html>

The article doesn't mention that doctype sniffing only applies to
text/html, so I think it is worth reiterating that documents served as
application/xhtml+xml are always rendered in the (Full) Standards mode
regardless of doctype. So XHTML 1.0 Transitional documents are rendered
in the Almost Standards mode when served as text/html and in the (Full)
Standards mode when served as application/xhtml+xml (or as text/xml or
as application/xml).

--
Henri Sivonen
hsiv...@niksula.hut.fi
http://www.hut.fi/u/hsivonen/

Bertilo Wennergren

unread,
Jul 3, 2002, 6:21:34 AM7/3/02
to
Tim:

> On Tue, 02 Jul 2002 16:38:17 +0000,
> Bertilo Wennergren <bert...@gmx.net> wrote:

>> There are people out there that have carefully built their pages taking
>> all the quirks in the sniffing routines into consideration. Some of them
>> might even have chosen a Doctype that used to give standards mode in
>> Mozilla, but that will trigger almost standards mode when new Mozilla
>> builds start to come out.

> And I cannot resist adding: That those people will get what they
> deserve (their broken pages, showing up as broken). ;-\

What broken pages? I was reffering to people authoring in (X)HTML
Transitional, according to the specs, expecting the browsers to
render the pages in Standards mode, not in Quirks mode or any
Almost Standards mode.

Some of those pages will get Almost Standards mode in future
Mozillas.

Or do you mean that all pages in (X)HTML Transitional are broken
per definition?

Keith Bowes

unread,
Jul 3, 2002, 4:40:47 AM7/3/02
to
Guy Macon wrote:
> I can't help but wish that the latest version of the most
> popular browser running on the most popular desktop OS
> would display those XHTML 1.0 Transitional documents when
> served as application/xhtml+xml. That limits my choices.

It works for me. I push the "Open" button of the "File Download" dialog
and it opens the document in Mozilla. Voila!


Tim

unread,
Jul 3, 2002, 9:04:05 AM7/3/02
to
On Tue, 02 Jul 2002 16:38:17 +0000,
Bertilo Wennergren <bert...@gmx.net> wrote:

>>> There are people out there that have carefully built their pages taking
>>> all the quirks in the sniffing routines into consideration. Some of them
>>> might even have chosen a Doctype that used to give standards mode in
>>> Mozilla, but that will trigger almost standards mode when new Mozilla
>>> builds start to come out.


Tim said:

>> And I cannot resist adding: That those people will get what they
>> deserve (their broken pages, showing up as broken). ;-\


On Wed, 03 Jul 2002 10:21:34 +0000,
Bertilo Wennergren <bert...@gmx.net> wrote:

> What broken pages? I was referring to people authoring in (X)HTML


> Transitional, according to the specs, expecting the browsers to
> render the pages in Standards mode, not in Quirks mode or any
> Almost Standards mode.

Whatever /else/ you were referring to, you also wrote about people
building pages to take advantage of quirks. You didn't preclude the
other types of quirky coding that people do. "Kludging" things, no
matter what the reason, has always been bad coding. And has always run
the risk of failing, when the author isn't as clever as they thought
they were.

Rijk van Geijtenbeek

unread,
Jul 3, 2002, 9:20:56 AM7/3/02
to
On Wed, 03 Jul 2002 10:21:34 +0000, Bertilo Wennergren

<bert...@gmx.net> wrote:
>Tim:
>> On Tue, 02 Jul 2002 16:38:17 +0000,
>> Bertilo Wennergren <bert...@gmx.net> wrote:

>>> There are people out there that have carefully built their pages taking
>>> all the quirks in the sniffing routines into consideration. Some of them
>>> might even have chosen a Doctype that used to give standards mode in
>>> Mozilla, but that will trigger almost standards mode when new Mozilla
>>> builds start to come out.
>
>> And I cannot resist adding: That those people will get what they
>> deserve (their broken pages, showing up as broken). ;-\
>
>What broken pages? I was reffering to people authoring in (X)HTML
>Transitional, according to the specs, expecting the browsers to
>render the pages in Standards mode, not in Quirks mode or any
>Almost Standards mode.
>
>Some of those pages will get Almost Standards mode in future
>Mozillas.
>
>Or do you mean that all pages in (X)HTML Transitional are broken
>per definition?

People who use a Transitional doctype are not yet ready to trust style
sheets for all their styling, and to use the markup only for marking up
content.

I don't think many designers are currently writing pages that depend on
small gaps showing between their single images in otherwise empty table
cells - those people don't seem to use tables purely for marking up
tabular content.

But I've read the bugzilla reports on this issue, and I sympathize with
the sentiment that Standards mode should be encountered by designers as
much as possible, so they have an incentive to fix their pages. And I
wonder why IBM didn't just put TD>IMG {vertical-align:bottom;} or
something like that in their site wide stylesheets, instead of pestering
Netscape to include one custom Doctype in this new 'Almost Standards
mode'.

--
.. it should look right, but that's no more than | Rijk van Geijtenbeek
finishing the paintwork on that wrecked car. If | mailto:ri...@iname.com
the underlying structure is unsound, no amount of | http://rijk.op.het.net
polishing the paintwork is going to make it fit for the road. (Alan Flavell)

Bertilo Wennergren

unread,
Jul 3, 2002, 1:32:22 PM7/3/02
to
Tim skribis:

> On Tue, 02 Jul 2002 16:38:17 +0000,
> Bertilo Wennergren <bert...@gmx.net> wrote:

>>>> There are people out there that have carefully built their pages taking
>>>> all the quirks in the sniffing routines into consideration. Some of them
>>>> might even have chosen a Doctype that used to give standards mode in
>>>> Mozilla, but that will trigger almost standards mode when new Mozilla
>>>> builds start to come out.

> Tim said:

>>> And I cannot resist adding: That those people will get what they
>>> deserve (their broken pages, showing up as broken). ;-\

> On Wed, 03 Jul 2002 10:21:34 +0000,
> Bertilo Wennergren <bert...@gmx.net> wrote:

>> What broken pages? I was referring to people authoring in (X)HTML
>> Transitional, according to the specs, expecting the browsers to
>> render the pages in Standards mode, not in Quirks mode or any
>> Almost Standards mode.

> Whatever /else/ you were referring to, you also wrote about people
> building pages to take advantage of quirks.

No, no. I was talking about the quirks in the _doctype sniffing
routines_, e.g. the xml declaration bug in MSIE. What I mean is that
doctype sniffing is a mess, but a mess some have learnt to live with.
They know the problems, the bugs and the quirks, and they have built
their pages _to get standards mode_ in all browsers that do such
sniffing. But if the browser builders go on and introduce still more
rendering modes, changing the sniffing routines, some such pages might
break, _although they are done according to specs, and are intended
to be rendered according to the standards_.

Either the browser builders should ditch the doctype sniffing
altogether (using standards mode for everything), or they should
leave it just as it is now (with the possible - but unlikely -
exception that MS ought to fix the stupid xml declaration bug).

Chris Hoess

unread,
Jul 3, 2002, 1:09:26 PM7/3/02
to
In article <3D233526...@gmx.net>, Bertilo Wennergren wrote:

> Either the browser builders should ditch the doctype sniffing
> altogether (using standards mode for everything), or they should
> leave it just as it is now (with the possible - but unlikely -
> exception that MS ought to fix the stupid xml declaration bug).

If you think you can convince our Evangelism team/IBM to fix their pages,
I certainly won't complain :-)

--
Chris Hoess

Alan J. Flavell

unread,
Jul 3, 2002, 1:35:31 PM7/3/02
to
On Jul 3, Tim inscribed on the eternal scroll:

> On Wed, 03 Jul 2002 10:21:34 +0000,
> Bertilo Wennergren <bert...@gmx.net> wrote:
>
> > What broken pages? I was referring to people authoring in (X)HTML
> > Transitional, according to the specs, expecting the browsers to
> > render the pages in Standards mode, not in Quirks mode or any
> > Almost Standards mode.
>
> Whatever /else/ you were referring to, you also wrote about people
> building pages to take advantage of quirks.

I really think you misread what the hon. Usenaut was saying.

> You didn't preclude the
> other types of quirky coding that people do. "Kludging" things, no
> matter what the reason, has always been bad coding.

I didn't for a moment read Bertilo as advocating that.

I interpreted it rather as advocating constructively capitalising on
their quirks to avoid one's standard-conforming content falling foul
of those very quirks.

Problem is, what the Mozzies are asking us to countenance here is that
some existing standards-conforming constructs would suddenly start to
be given a quirky treatment, i.e breaking something that wasn't
broken, in order to help something that _is_ broken to look as if it
wasn't. That is an unfair tradeoff in my book.

And I'm frankly offended by this waste of resources to pander to
broken implementations, when there's a whole list of unfixed bugs that
ought to be fixed, rather than spending effort on designing-in new
bugs - as seems to be happening here. And setting an unfortunate
precedent that no doubt another big player will be only too happy to
escalate with a different set of quirks. Talk about aiming at a
moving target!!

Henri Sivonen

unread,
Jul 4, 2002, 12:26:27 AM7/4/02
to
In article <3D233526...@gmx.net>,
Bertilo Wennergren <bert...@gmx.net> wrote:

> But if the browser builders go on and introduce still more
> rendering modes, changing the sniffing routines, some such pages might
> break, _although they are done according to specs, and are intended
> to be rendered according to the standards_.

(To be fair to Mozilla, the new Almost Standards Mode won't break your
page any further than the Standards modes of Windows IE 6 and Mac IE 5.)

> Either the browser builders should ditch the doctype sniffing
> altogether (using standards mode for everything)

In that case, what would you have done about bug 22274? Would you have
chosen the Opera way or the Mozilla Standards mode way?

Henri Sivonen

unread,
Jul 4, 2002, 12:59:03 AM7/4/02
to
In article <Pine.LNX.4.40.02070...@lxplus035.cern.ch>,

"Alan J. Flavell" <fla...@mail.cern.ch> wrote:

> And I'm frankly offended by this waste of resources to pander to
> broken implementations, when there's a whole list of unfixed bugs that
> ought to be fixed, rather than spending effort on designing-in new
> bugs - as seems to be happening here.

What actually happened is that the tradition and, by extension, author
expectations are incompatible with the CSS2 inline box model.

The other options were:
* Not implementing the inline box model according to the spec. (IE)
* Implementing the inline box model according to the spec except for
the conficting case. (Opera)
* Implementing according to the spec, emailing site authors, dealing
with scores or duplicate bug reports.
* Implementing according to the spec but only in the standards mode and
narrowing the cases that actually trigger the standards mode, that is,
making sure no one uses the standards mode.

Bertilo Wennergren

unread,
Jul 4, 2002, 7:25:53 AM7/4/02
to
Henri Sivonen:

> In article <3D233526...@gmx.net>,
> Bertilo Wennergren <bert...@gmx.net> wrote:
>
>> But if the browser builders go on and introduce still more
>> rendering modes, changing the sniffing routines, some such pages might
>> break, _although they are done according to specs, and are intended
>> to be rendered according to the standards_.

> (To be fair to Mozilla, the new Almost Standards Mode won't break your
> page any further than the Standards modes of Windows IE 6 and Mac IE 5.)

That's right.

>> Either the browser builders should ditch the doctype sniffing
>> altogether (using standards mode for everything)

> In that case, what would you have done about bug 22274? Would you have
> chosen the Opera way or the Mozilla Standards mode way?

I don't know which reading of the specs is the correct one. The Mozilla
gurus know far better than me.

Reply all
Reply to author
Forward
0 new messages