Re: The argument for |bugmode|

At 00:15 +0300 UTC, on 2007-04-19, Henri Sivonen wrote:

> On Apr 18, 2007, at 19:33, Chris Wilson wrote:
>
>> I want to explicitly state - I neither asked for nor want the HTML
>> WG to specify a "bugmode" attribute in HTML.
>
> Would you like a conformance checker to mark whatever IE-version
> switch you are going to use as non-conforming? (You have very clearly
> stated the intention to have an opt-in switch in future IE versions.)
>
> If you would like the switch to be marked non-conforming, do you
> expect Web designers to accept it?

Speaking for the Web Repair Initiative, I think that would be awful. We're
trying to improve authoring tools, which means amongst other things that we
would like to see authoring tools include conformance checkers. IMO the ideal
conformance checker for such environments would be a simple two (maybe three)
state icon, indicating conformance or not. If the authoring tool would be
forced to generate invalid markup just to opt-in to IE's 'super mega really
truly standards mode', that would mean the user would need to look at the
*details* of the error reported for each and every page, because by just
looking at the conformance checker's icon they can't know what it is
flagging, just that it is flagging. (And for such conformance tools to ignore
the IE opt-in switch would be unacceptable to require of them, because if the
opt-in switch isn't specced, such tools have no reliable source to base on
what the opt-in switch actually looks like. In fact, the authoring tool would
need to allow the user to explicitly opt-in to IE's 'standards mode', so it
does in fact need to know the correct syntax of the switch to be able to
check its conformance.)

Compare to checking pages in iCab, or in Safari or Firefox with Tidy plug-in.
Easy to quickly click through the site's pages to see they're all OK. When
every page contains a non-conforming IE switch, the this method instantly
becomes a drag, which can only result in fewer authors bothering to check
conformance.

If we want the Web's interoperability to improve, we should make it easy for
authors to produce quality code, not make it harder. A non-conforming IE
opt-in switch is unacceptable.

> If you would not like the switch to be marked non-conforming, I see
> three options:
>   1) Catering for it in the spec.
>   2) Tying switching to a syntactic feature that officially isn't an
> IE bug mode switch.
>   3) Putting the switch into comments or baking it into the choice of
> syntactic sugar to cloak it from conformance checkers.
>
> #2 and #3 are bad for intermediate tool vendors who build on XML
> toolchains.

Right. If we truly must have an IE opt-in switch, #1 would seem to be the
only reasonable option.


-- 
Sander Tekelenburg
The Web Repair Initiative: <http://webrepair.org/>

Received on Thursday, 19 April 2007 07:35:44 UTC