DOCTYPE and validation

[Mary Ellen Curtin]

I was reading the New York Times
http://www.nytimes.com

and I took a peek at the code. It's got a DOCTYPE, rather to my
surprise,
but my gut reaction to the scanning the code is, "I pity whoever
maintains this puppy." So for giggles I asked

http://validator.w3.org

what it thought of the NYT, and it said, "I can't find the encoding! I
looked
everywhere!"

But looking at the code I see:

<meta http-equiv="Content-Type" content="text/html">
<meta http-equiv="charset" content="iso-8859-1">

-- so why didn't the Validator spot it?

Not surprisingly, the page doesn't validate.

Another question: Why bother setting a DOCTYPE if your page won't
validate? Why set a DOCTYPE if you're not (as far as I can tell in the
morass of JavaScript and table-formatting) going to use CSS?

Mary Ellen
Doctor Science, MA

[Alan J. Flavell]

On Sat, Jul 26, Mary Ellen Curtin inscribed on the eternal scroll:

I was reading the New York Times
http://www.nytimes.com
[INLINE]
[INLINE] [LINK] [INLINE] [INLINE] The New York Times On The Web
[INLINE] [LINK] [INLINE]
[INLINE] [INLINE] [INLINE]
[INLINE] UPDATED SATURDAY, JULY 26, 2003 3:12 PM ET [INLINE] |
[INLINE] Personalize Your Weather [INLINE]
[INLINE]
[INLINE]

Par for the course, it seems...
what it thought of the NYT, and it said, "I can't find the encoding!
I looked everywhere!"
Oh, it looked everywhere that the specifications said it should look!
But looking at the code I see:

<meta http-equiv="Content-Type" content="text/html">
<meta http-equiv="charset" content="iso-8859-1">

-- so why didn't the Validator spot it?
Do you have a published interworking specification that codifies such
behaviour? I don't know of one. The key applicable specification
would have to be RFC2616 - look at the header field definitions in
section 14. Do you see a "Charset" header there? I don't. So - that
META is claiming to be equivalent to an HTTP header that doesn't
exist.

Netscape did manage to wangle that once, with their "Refresh"
pseudo-header, but as a general principle it's not on.
Another question: Why bother setting a DOCTYPE if your page won't
validate?
Superstition? You've already guessed that it might also be done to
fool a browser into interpreting CSS in standards-mode...
Why set a DOCTYPE if you're not (as far as I can tell in the
morass of JavaScript and table-formatting) going to use CSS?

Some kind of magic chicken waving ceremony, perhaps. Who knows
what motivates such folks, and where they get their documentation
from?

[Mary Ellen Curtin]

Alan Flavell wrote:

Why set a DOCTYPE if you're not (as far as I can tell in the
morass of JavaScript and table-formatting) going to use CSS?

Some kind of magic chicken waving ceremony, perhaps. Who knows
what motivates such folks, and where they get their documentation
from?

Thinking it over, I believe it probably represents one geological
stratum among many in the code for the NYT site. I imagine that
"what motivates such folks" is that they can't afford to redo the whole
enormous site from scratch, so they're just sticking patches on top
of rebuilds on top of patches. Pray it never happens to you.

Mary Ellen
Doctor Science, MA