- From: James Graham <jgraham@opera.com>
- Date: Fri, 16 Jan 2009 14:59:14 +0100
- To: Sam Ruby <rubys@us.ibm.com>
- CC: Jirka Kosek <jirka@kosek.cz>, HTML WG <public-html@w3.org>
Sam Ruby wrote:
The main proposal here seems to be <!DOCTYPE html PUBLIC "">. AFAIK this
doctype not solve the original problem for some tools which are
incapable of generating a doctype with a PUBLIC component but no SYSTEM
component [1]. Therefore this seems like a poor solution, at the very
least <!DOCTYPE html PUBLIC "" ""> should be used (assuming these tools
are OK with a null SYSTEM).
> 1) Single DOCTYPE, with a null quoted string
I am not happy with this ("cannot live with this" if I am forced into a
binary choice). It makes the doctype almost twice as long and carries no
extra information. Authors should not have to pay this tax.
> 2) DOCTYPE with an optional null quoted string
I can live with this but it seems suboptimal since the differences
between this and <!doctype html> and the longer form are not
self-descriptive.
> 3) Two DOCTYPES: one "preferred" with no quoted string, and one
> "pejorative" with the value "legacy-compat".
I can live with this but, given that others cannot live with this, I
would prefer the long form be
<!DOCTYPE html PUBLIC "sgml-compatibility" "">
or possibly some alternative with a URI in the SYSTEM component if there
are significant tools that cannot live with empty system components
(XSLT 2.0?). This describes the underlying reason for the long form and
has a certian pejorative value but less than that of "legacy-compat".
> 4) Two DOCTYPES: one with no quoted string, and one with a value of
> "XSLT-compat" that should not be used unless the document is generated
> from XSLT.
I cannot live with this. XSLT is only one tool amongst many (and I guess
not even an especially popular one); if this problem is worth solving it
is worth solving properly.
[1] http://krijnhoetmer.nl/irc-logs/whatwg/20090116#l-317
Received on Friday, 16 January 2009 13:58:13 UTC