Previous Message
Next Message

Cross: [css-d] doctype declarations

Sent by Jan Roland Eriksson on 25 January 2002 15:03


[Ok, in lack of a response (so far:), let's try it]
[message crossed to [EMAIL-REMOVED]]

>Andrew Clover wrote:

>>Also, the DOCTYPE isn't quite right. You've probably copied it from the
>>W3C site, but unfortunately as it contains a relative URL it can only

>Just as I hit send, the question ocurred to me: what's the
>recommended method for including DOCTYPEs?

Trust me, you don't want to go there :)

>Should we as developers download the specifications and host them
>locally? It seems like it would put a lot of load on the W3 servers
>if browsers continually hit them for the information.

That is the first level of total misunderstanding of what a
<!DOCTYPE... sniffer is doing.

It does not go HTTP to download anything from anywhere, to the best of
my knowledge no one of the popular user agents has a catalog to help
resolve PUBLIC identifiers, and they don't use the SYSTEM identifier
as a URL from where to get an external subset either.

<!DOCTYPE... sniffing as implemented is _only_ about pattern
recognition of some text (the doctype lines) against a "hard coded"
table in the ua, and depending on what kind of "match" that comes up,
the ua will switch between its "standard" or "quirk" modes when it
goes on to render pages.

>What about intranets, or closed networks? What's the behavior of
>the browser when a DOCTYPE is inaccessible?

YM "when the external declaration subset is inaccessible"?

Nothing special because just in this case it's not needed for the
purpose it was designed initially. Doctype sniffers are not looking at
the external subset, they are looking at the <!DOCTYPE... declaration
text itself that maybe, in one form or another, is available at the
top of your document instance.

These two are probably the best ones to move you "forward"...

  http://www.hut.fi/~hsivonen/doctype.html
  http://gutfeldt.ch/matthias/articles/doctypeswitch.html

...and you just do not want to know about the list of various
<!DOCTYPE's/reactions that David Baron has collected :-D

The whole subject is...
...well, let's leave that for some other post maybe :)

-- 
Rex .. [EMAIL-REMOVED]> .. <http://css.nu/>
Previous Message
Next Message