> the Date() object should be fairly accurate since its creation doesn't
> take any mentionable time. I don't really see how you can create
> something that'll work with it though, since the only tools you have
> for forking (which is required to get the screen update) is
> setTimeout() and setInterval(). so regardless of how you do things
> you won't get clock ticks below 55ms on 95/98-machines.
I wasn't being clear enough. I would use different threads based on
settimeout, but the synchronization would be done by the date. Now as you
said, you can't get clock ticks below the minimum slice, but nevertheless
comparing to the date is more accurate than relying on the settimeout or
setinterval (Since they might get pushed forward by the minimum slice at
some point when the browser is not idle). Blaah, This isn't a bit more
clear, is it? :)
Anyway, how I would do my fancy animation on w9x would be relying on a
applet based timer (which uses Thread.sleep) Now, again I don't have w9x so
I can't test it, but at least with my NT I can acheive timed events smaller
than the 10ms by using Thread.sleep.
> I'm not sure I agree with you regarding setTimeout(). if you read
> <URL: http://www.webreference.com/dhtml/column28/part-4.html > and
> look at the test results on the page before it, you'll see that it's
> rather difficult to predict the overall results of a tail-end
> setTimeout() call. setInterval() gave much more consistant results,
> particularly when you check the various systems (Mac vs Linux vs NT,
> for instance).
I've read it already. Thanks by the way for the article, great work. But the
question is, why you insist on using tail-end timeouts? Why not set them in
the beginning of your method?
BTW, have any of you studied a possibility to use some system caused events?
Just out of curiosity, how many times for example a mousemove event occurs
in a second (propably not enough, but anyway...)
I have to take a look of the url you gave a bit later, since it does not
rain in Helsinki :)
--
Tuomas
_______________________________________________
Dynapi-Help mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/dynapi-help
