W3C home > Mailing lists > Public > public-script-coord@w3.org > July to September 2011

Re: Vestiges of [Supplemental]

From: Simon Pieters <simonp@opera.com>
Date: Thu, 11 Aug 2011 16:03:19 +0200
To: "Ian Hickson" <ian@hixie.ch>, "Jonas Sicking" <jonas@sicking.cc>
Cc: public-script-coord@w3.org
Message-ID: <op.vz1yntzridj3kv@simon-pieterss-macbook.local>
On Thu, 21 Jul 2011 23:28:32 +0200, Jonas Sicking <jonas@sicking.cc> wrote:

> On Wed, Jul 20, 2011 at 7:52 PM, Ian Hickson <ian@hixie.ch> wrote:
>> On Wed, 20 Jul 2011, Jonas Sicking wrote:
>>>
>>> No other interface that I can think of has this sort of "dynamic"
>>> behavior depending on which global scope it is available in.
>>
>> The Window global object is different based on whether it was invoked  
>> from
>> showModalDialog() or not.
>
> This seems like an equally bad idea (if not worse since a single
> script can reach two types of Window interface objects). Why couldn't
> the properties that showModalDialog needs simply be properties on the
> instance object directly?
>
>> In any case, the idea is just that workers should always look the same,  
>> so
>> that libraries used in shared workers and dedicated workers work fine
>> either way, with only the bits that actually matter being different.
>
> They should look the same except be different in some ways? ;-)
>
> You'll always be able to check if an object is the global scope by
> checking "x instanceof WorkerGlobalScope". The fact that they'll have
> different classes seems like a good thing since they in fact have
> different APIs.

I tend to agree with Jonas. There's no point in hiding the interface name  
 from scripts, it's just a source of potential bugs and confusion.

As a data point, we have Location and WorkerLocation which also "look the  
same but are different".

-- 
Simon Pieters
Opera Software
Received on Thursday, 11 August 2011 14:03:26 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 8 May 2013 19:30:04 UTC