W3C home > Mailing lists > Public > public-webapi@w3.org > January 2007

Re: Timing Interface

From: Maciej Stachowiak <mjs@apple.com>
Date: Sat, 27 Jan 2007 15:47:41 -0800
Message-Id: <497EF61F-9890-46B0-812F-2287D9CF87EB@apple.com>
Cc: Charles McCathieNevile <chaals@opera.com>, Web API public <public-webapi@w3.org>, SVG working group <w3c-svg-wg@w3.org>, Anne van Kesteren <annevk@opera.com>
To: Jim Ley <jim@jibbering.com>


On Jan 27, 2007, at 2:57 PM, Jim Ley wrote:

>
> "Anne van Kesteren" <annevk@opera.com>
>> On Sat, 27 Jan 2007 12:59:00 -0500, Charles McCathieNevile  
>> <chaals@opera.com> wrote:
>>> This is because we are allegedly going to get around to it, if I  
>>> recall
>>> correctly. If you're keen, I would suggest you further track  
>>> exactly where that got up to. It being Saturday morning, I don't  
>>> recall which spec
>>> it falls in...
>>
>> Window will specify it for ECMAScript. Most other languages  
>> already have a library for doing this.
>
> Has there been any attempt to punt timers to the ECMA 2.0 people,  
> as it really is not a WEB API, and it's only being done as a band- 
> aid to the ES limited people.  It might be nice to then just write  
> up setTimeout on a if you're going to do it, do it like this.   
> Rather than make the effort to design a suitable WEB-API for timers?

Timers in the Window spec are needed not just to paper over  
ECMAScript limitations, but specifically to define a timer that plays  
well with the user agent's event/paint loop. In particular, for  
safety and interoperability, you need to ensure that timers fire at a  
time when the DOM is in a consistent state, and you have to describe  
how they interact with things like location changes, style updates, etc.

So even languages like Java need a defined timer API. Java has its  
own timers, but because each runs on a separate thread, they can give  
no guarantees about running at an appropriate time. In fact they  
can't even give very good guarantees about firing order.

Regards,
Maciej
Received on Saturday, 27 January 2007 23:48:40 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:18:56 GMT