Hi Maciej,
On Apr 04, 2006, at 12:30, Web APIs Issue Tracker wrote:
What to do about timers? Some rough options:
A.1) Define in an ECMAScript-only way, and assume other languages
all have their own built-in timer
facility. But is that true? In particular, do they have one that is
usable in a browser scripting context? I
am not sure, since timer facilities are generally based on some
sort of event loop, and the browser
controls the event loop in this case.
I think we have to standardise the ES timers as they are currently
available in browsers. That's the only way in which they can ever get
reused by other specs. If we don't do that they'll have to invent new
stuff, which it is our task to make sure they have no reason to do.
B.1) Invent a new timer API that's more like the SVG uDOM one (but
simpler to use, it shouldn't take 4
lines of code to set up a timer callback in ES). Define the
existing implemented Window timer methods
in terms of this for ECMAScript only.
That's already part of our deliverables, so I would expect that we
stick to the plan. We have to stick to what exists because it's what
most people will be used to having and will want (and what will get
implemented) but we need something simple and new that works for
other languages, and that may have the extra advantage here or there
so that ES folks may perhaps actually want to use it when available.
For extra points we should define the relationship between the two
facilities.
There is also the option of making it just like an EventListener,
but I don't think events are a good
model for timers firing, most toolkit APIs I know of make timers
and events separate concepts.
I quite like the idea of an event based model because one could then
wire in timer events into the rest of the system, such that it would
work with SMIL, XML Events, etc.. I'm not married to it, but I like it.
The plan I had in mind was that I would dump what the SVG WG came up
with into a spec, and put it up so that folks could start tearing it
to pieces or falling in love with it (or both, if you're into that).
I didn't have it up in high priority though, the FPWD is on the
timeline in September (but we can be flexible about that).
This is the one time in writing this spec I have felt the
temptation to invent a new feature. Someone
smack me back to reality.
C'mon, hang in there, the better we stick to the documentation part
of our jobs, the sooner we'll be able to get creative where needed ;)
--
Robin Berjon
Senior Research Scientist
Expway, http://expway.com/
