An other vierd question :-)

[user]

Find a page some table cells was made like this:
<td width="253" height="10" bgcolor="98a7b2">Some text in white color</td>

Since the page background was white I could't see the text inside the
cell, with Mozilla. But in IE it shows. The difference is that the color
code don't have the # sign included.

Since that affected the rendering different in browsers, I was very
supriced that the W3C validator don't notice that as an error. Or is it
just Mozilla who required the # sign in the color code?

I have a test on my own here, if interested:
http://w1.978.telia.com/~u97802964/tests/test3.html

--
/Arne

[Leif K-Brooks]

us**@domain.invalid wrote:

Find a page some table cells was made like this:
<td width="253" height="10" bgcolor="98a7b2">Some text in white color</td>

Since the page background was white I could't see the text inside the
cell, with Mozilla. But in IE it shows. The difference is that the color
code don't have the # sign included.
IE is forgiving enough (a bad thing) to allow invalid colors. Better
browsers aren't.
Since that affected the rendering different in browsers, I was very
supriced that the W3C validator don't notice that as an error. Or is it
just Mozilla who required the # sign in the color code?

The validator can only validate to DTDs, which aren't very flexible. The
HTML transitional DTD (http://www.w3.org/TR/REC-html40/loose.dtd) only
defines colors as CDATA, which puts very little restriction on their
content.

[user]

Once upon a time *Leif K-Brooks* wrote:

us**@domain.invalid wrote:

Find a page some table cells was made like this:
<td width="253" height="10" bgcolor="98a7b2">Some text in white color</td>

Since the page background was white I could't see the text inside the
cell, with Mozilla. But in IE it shows. The difference is that the color
code don't have the # sign included.

IE is forgiving enough (a bad thing) to allow invalid colors. Better
browsers aren't.

Since that affected the rendering different in browsers, I was very
supriced that the W3C validator don't notice that as an error. Or is it
just Mozilla who required the # sign in the color code?

The validator can only validate to DTDs, which aren't very flexible. The
HTML transitional DTD (http://www.w3.org/TR/REC-html40/loose.dtd) only
defines colors as CDATA, which puts very little restriction on their
content.

When I tried the Strict DTD on my test page, I get an error for the
attribute "bgcolor" as I should. But that was only on the line where
color was set with the # sign. The same line with color set without #
passed.

So as it seems, the validator don't even consider the string "bgcolor
color='000000'" as a code. Is it because # is missing? If so, then the
validator must have some "value" for the sign? Am I confused, or am I
really confused :-)

--
/Arne

[Neal]

On Sat, 16 Oct 2004 17:32:00 GMT, <us**@domain.invalid> wrote:

Once upon a time *Leif K-Brooks* wrote:

The validator can only validate to DTDs, which aren't very flexible. The
HTML transitional DTD (http://www.w3.org/TR/REC-html40/loose.dtd) only
defines colors as CDATA, which puts very little restriction on their
content.

When I tried the Strict DTD on my test page, I get an error for the
attribute "bgcolor" as I should. But that was only on the line where
color was set with the # sign. The same line with color set without #
passed.

Yes, because the validator can't tell #XXXXXX is any more right than
XXXXXX. They're both legal values for CDATA, and that's as much as the
validator is able to check. It remains, however, that compliant UAs should
not interpret XXXXXX and should interpret #XXXXXX.

[Toby Inkster]

user wrote:

So as it seems, the validator don't even consider the string "bgcolor
color='000000'" as a code. Is it because # is missing? If so, then the
validator must have some "value" for the sign? Am I confused, or am I
really confused :-)

It is *valid* (in an SGML sense) but *wrong*. There is a difference.

Just another example that validating isn't the be all and end all.

--
Toby A Inkster BSc (Hons) ARCS
Contact Me ~ http://tobyinkster.co.uk/contact
Now Playing ~ ./ben_folds/sunny_16/05_songs_of_love.ogg

[user]

Once upon a time *Toby Inkster* wrote:

user wrote:

So as it seems, the validator don't even consider the string "bgcolor
color='000000'" as a code. Is it because # is missing? If so, then the
validator must have some "value" for the sign? Am I confused, or am I
really confused :-)

It is *valid* (in an SGML sense) but *wrong*. There is a difference.

Just another example that validating isn't the be all and end all.

Well, I guess the only problem here is that some so called webmasters
will maintain they are right only because their pages validate even if
they don't use # :-P

--
/Arne

[Steve Pugh]

On Sat, 16 Oct 2004 16:25:40 GMT, us**@domain.invalid wrote:

Find a page some table cells was made like this:
<td width="253" height="10" bgcolor="98a7b2">Some text in white color</td>

Since the page background was white I could't see the text inside the
cell, with Mozilla. But in IE it shows. The difference is that the color
code don't have the # sign included.

Since that affected the rendering different in browsers, I was very
supriced that the W3C validator don't notice that as an error. Or is it
just Mozilla who required the # sign in the color code?

It's a limitation of the validator - the bgcolor attribute is defined
as taking values from the set %color and %color is defined as CDATA
which more or less means anything. This is because in HTML colours can
be either #RRGGBB or one of 16 keywords. The limitations on what's an
acceptable value are in the prose of the specification rather than in
the DTD.

Hence, Mozilla is correct in ingoring the incorrect value and the
validator is correct in not reporting it as a validity error. IE, of
course, is correcting the error behind the scenes.

Steve

[Steve Pugh]

On Sat, 16 Oct 2004 17:32:00 GMT, us**@domain.invalid wrote:

Once upon a time *Leif K-Brooks* wrote:

us**@domain.invalid wrote:

<td width="253" height="10" bgcolor="98a7b2">Some text in white color</td>

Since that affected the rendering different in browsers, I was very
supriced that the W3C validator don't notice that as an error. Or is it
just Mozilla who required the # sign in the color code?

The validator can only validate to DTDs, which aren't very flexible. The
HTML transitional DTD (http://www.w3.org/TR/REC-html40/loose.dtd) only
defines colors as CDATA, which puts very little restriction on their
content.

When I tried the Strict DTD on my test page, I get an error for the
attribute "bgcolor" as I should. But that was only on the line where
color was set with the # sign. The same line with color set without #
passed.

Under ths Strict DTD bgcolor isn't a permitted attribute at all. Use
CSS instead. The validator only reports the first instance of a
non-valid attribute so that's why the second (and subsequent) one was
ignored.

Steve

[Brian]

us**@domain.invalid wrote:

some so called webmasters will maintain they are right only because
their pages validate

They are unquestionably wrong. A document can be obviously in violation
of html recommendations and still be valid. Valid in SGML terms only
means that the document adheres to the formal syntax of a dtd.
--
Brian (remove "invalid" to email me)

[user]

Once upon a time *Brian* wrote:

us**@domain.invalid wrote:

some so called webmasters will maintain they are right only because
their pages validate

They are unquestionably wrong. A document can be obviously in violation
of html recommendations and still be valid. Valid in SGML terms only
means that the document adheres to the formal syntax of a dtd.

Just found out that IE does the right thing if CSS is used for the color
settings, and don't try to correct "xxxxxx" as it does with the
depricated "bgcolor". Let's hope the webmasters use more CSS :-)

--
/Arne

VARNING: Konsumtion av alkohol kan få dig att tro att du kan ha föra ett
vettigt samtal med någon av det andra könet, utan att spotta.

[Toby Inkster]

Brian wrote:

They are unquestionably wrong. A document can be obviously in violation
of html recommendations and still be valid. Valid in SGML terms only
means that the document adheres to the formal syntax of a dtd.

The simple example is:

<p>The cat sat on teh mat.</p>

Obvious mistake, but it's not invalid.

--
Toby A Inkster BSc (Hons) ARCS
Contact Me ~ http://tobyinkster.co.uk/contact

[Brian]

Toby Inkster wrote:

Brian wrote:

They are unquestionably wrong. A document can be obviously in violation
of html recommendations and still be valid. Valid in SGML terms only
means that the document adheres to the formal syntax of a dtd.

The simple example is:

<p>The cat sat on teh mat.</p>

Obvious mistake, but it's not invalid.

:-D Humerous, but not on point. There's no markup error in your
example. But you can have clear markup errors without having invalid
markup:

<P lang="thisIsAnExampleOfIncorrectMarkup">There is no validation error
here.</P>.

--
Brian (remove "invalid" to email me)

[Harlan Messinger]

"Toby Inkster" <us**********@tobyinkster.co.uk> wrote in message
news:pa****************************@tobyinkster.co .uk...

Brian wrote:

They are unquestionably wrong. A document can be obviously in violation
of html recommendations and still be valid. Valid in SGML terms only
means that the document adheres to the formal syntax of a dtd.

The simple example is:

<p>The cat sat on teh mat.</p>

Obvious mistake, but it's not invalid.

Or "Colorless green dreams sleep furiously". Syntactically and
orthographically correct, semantically flawed. (In case you want to take
issue with the orthographic part, let me rewrite it as "Colourless green
dreams sleep furiously.")