W3C home > Mailing lists > Public > public-html@w3.org > October 2009

Re: typeof document.all

From: Brendan Eich <brendan@mozilla.org>
Date: Thu, 15 Oct 2009 19:26:15 -0700
Cc: Maciej Stachowiak <mjs@apple.com>, HTML WG <public-html@w3.org>
Message-Id: <82F17350-C2E8-4423-8096-069BD2DA85C8@mozilla.org>
To: Ian Hickson <ian@hixie.ch>
On Oct 15, 2009, at 6:41 PM, Ian Hickson wrote:

>> Ok, but not the cycle issues.
>
> Could you elaborate on that? If there are issues not yet addressed,  
> I'd
> love to fix them too.

I send a separate mail when I have more time.


>>> Actually I spent very little time (an afternoon?) specifying
>>> document.all to the point where I was happy that it was fully
>>> specified and compatible with legacy content. It wasn't that
>>> difficult.
>>
>> I didn't think so, but it's not necessary to overspecify for interop
>> today. Let's not go in circles: my point is not to optimize only  
>> *your*
>> time, but everyone currently thrashing document.all here, in
>> public-script-coord, and on es-discuss. We are very far from done  
>> with
>> some "value types" solution in ES-Harmony.
>
> I don't think this need take much of anyone's time -- the model in  
> HTML5
> today could be easily supported by ES3 just by ES3 saying that host
> objects can have special behaviour for 'typeof' and 'ToBoolean()',  
> just
> like they have special behaviour for many other things, like [[Get]].


That's not enough. There would have to be special cases for == and !=  
as well.

But isn't it easy for you to suggest even more host object loopholes,  
when Ecma TC39 is trying to constrain or close the existing ones? It's  
not that simple, and again: the world should not optimize for Hixie's  
convenience, or any one implementor's.

/be
Received on Friday, 16 October 2009 02:27:18 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:39:09 UTC