[whatwg] Using the HTML5 DOCTYPE as a new quirksmode switch

----- Original Message ----- 
From: "Benjamin Hawkes-Lewis" <bhawkeslewis@googlemail.com>
To: <matt at builtfromsource.com>
Cc: <whatwg at whatwg.org>
Sent: Saturday, March 10, 2007 4:53 PM
Subject: Re: [whatwg] Using the HTML5 DOCTYPE as a new quirksmode switch


> Matthew Ratzloff wrote:
>> Relying on headers is a good way to get people to ignore that part of the
>> specification.  Web designers don't want to worry about headers and
>> .htaccess files.  It has to be syntactic.
>
> I think expecting the mass of web designers to worry about doctypes
> isn't much less optimistic than expecting them to worry about headers.
> Web designs frequently break quite seriously because web authors think
> that documents sent over the wire are defined entirely within the
> document text itself. e.g. Web authors often seem to imagine a charset
> defined in <meta> is determinative, and end up serving gibberish. Like
> it or not, effective web delivery depends on correct HTTP headers.

Opposite statement is correct too:

"Web authors often seem to imagine a charset
defined in HTTP headers is determinative, and end up serving gibberish"

There are many cases when sequence of bytes with some markup is
the only thing that is available.

Think about partial content updates, "ajax" (or "ahjacs"?) kind of 
scenarios.

And yet: web server configuration of headers is not always available.
Public virtual site hosts is a good example.

And more:
Server adminestering and content creation are different roles/activities.
As a rule different people handle these tasks. Requiring both of them
to be involved in proces of creation of valid content will decrease
probability that result will be valid.

It is better if markup itself will be descriptive enough.

BTW: why not to use <html5>...</html5> envelope
instead of <html>?
<html> is an optional element in HTML so "old"
browsers will switch to quirks mode for the <html5> content.
And html5 UAs will know what to do with it.
So this <html5> element is pretty backward compatible solution.

Andrew Fedoniouk.
http://terrainformatica.com

Received on Saturday, 10 March 2007 17:24:37 UTC