W3C home > Mailing lists > Public > public-qa-dev@w3.org > May 2004

Re: v0.6.6 Release? (was: 0.6.6 bug roundup)

From: Terje Bless <link@pobox.com>
Date: Wed, 19 May 2004 06:03:37 +0200
To: QA Dev <public-qa-dev@w3.org>
Message-ID: <b02010203-1033-81C07C4FA94911D8B4B40030657B83E8@[]>

Hash: SHA1

Ville Skyttä <ville.skytta@iki.fi> wrote:

>If "friendliness" and the core task (or reason for existence, if you
>prefer) of the Validator conflict, to me the solution is clear: be
>honest, sacrifice friendliness, not validation, to the extent necessary.

I agree, completely, but "the extent necessary" is what's at issue here.

>And advertising should not get into the way of fixing known major bugs
>or backing out (if that's the best we can do) new known buggy features.

I disagree that it's a "major bug" (and with the "advertising" term in this
context, BTW). The result it produces is certainly catastrophically wrong, but
only in a limited set of circumstances; this is only triggered by pages that
begin with «<!DOCTYPE html» but have some kind of garbage for the FPI/SI.

In particular, it will _not_ manifest for the most common of these — the
lower-case «<!doctype» produced by Netscape's 4.x editor — nor for any other
case with a missing DOCTYPE Declaration.

The number of documents that specify a Document Type Name of «html» without
giving a proper Public or System Identifier is relatively minor; especially
when compared to the number of pages that lack any form of DOCTYPE Declaration
(which is our biggest problem, and the one this behaviour was designed to deal

Note also that «<!DOCTYPE html>» is perfectly acceptable from an SGML point of
view; it's an explicit part of the SGML design that the “receiving” SGML
Application's Parser be able to resolve the External Subset from the Document
Type Name alone[0]. The current issue is one created solely by the odd special
case of the Web where certain FPIs — and even some SIs — are mandated by the
prose of the relevant specifications.

Not to imply that this behaviour isn't wrong and misleading — we will have to
figure out some way to resolve it eventually — but looking at the
circumstances that trigger it and the consequences of the failure, when
compared to anything we might conceivably do to counteract it in the near
term, the former weighs inconsequentially little compared to the latter.

[0] - If “The Web” had been conceived with acknowledgement of its SGML
      roots, you could even conceivably have implied both SGML Declaration
      and Document Type Definition from the MIME Content-Type — while still
      beeing in perfect alignment with SGML in general — and avoided this
      whole problem alltogether.

- -- 
If you believe that will stop spammers, you're sadly misled. Rusty hooks,
rectally administered fuel oil enemas, and the gutting of their machines,
*that* stops spammers!                                         -- Saundo

Version: PGP SDK 3.0.3

Received on Wednesday, 19 May 2004 00:03:55 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:54:46 UTC