New Squeak Porting Proposals
Raab, Andreas
Andreas.Raab at disney.com
Tue Aug 17 18:08:00 UTC 1999
Dean,
> The stylus interface works, as you've implemented it. It's just not
> very
> usable. Because WinCE doesn't show the mouse cursor, and there is
> essentially
> only one button with now Cmd or Option keys, a lot of the Squeak interface
> is
> inaccessible (especially in Morphic).
>
Do you have any idea how that could be fixed except from writing a complete
new UI for stylus based computers?!
> A couple of people (myself included) have commented on the apparent
> sluggishness of the UI on the Casio E100/105, but I was pretty sure it
> isn't an
> interpreter problem because a lot of things like sound synthesis and the
> character recognizer work fine and feel fast. Morphic reveals this
> sluggishness
> quite dramatically.
>
Ah ja. There are a zillion possible reasons for this but here are two
suggestions:
* Get rid of PasteUpMorph>>interCyclePause: which will wait for twenty msecs
inbetween two morphic cycles and is totally inappropriate for any slow
machine.
* Try different display depths - if it feels okay at 1 bit depth then it is
likely to be a cache problem. This is a particular problem since Squeak has
to shuffle a huge amount of bits around (and is a major difference from
sound synthesis and character recognition).
> Trying out the SameGame morph, which is quite UI
> intensive, but not all that processor intensive revealed a lot. To make
> it
> work, you have to click and hold for several seconds, and many times need
> to
> repeat this action before the click is acted on.
>
This actually sounds like we're loosing some of the input events - the first
suggestion above *might* help here though it would be interesting how it
feels if you set the display depth to 1 bit first.
> My first thought on this was that perhaps the #idleProcess was giving
> up
> too much time to WinCE, so I tried changing it as follows:
>
> idleProcess
> "A default background process which is invisible."
>
> [true] whileTrue: [true].
> "[self relinquishProcessorForMicroseconds: 1000]."
>
> This resulted, fairly quickly, in a WinCE dialog box popping up that said:
> 'Squeak is going to CRASH right NOW.'
>
Whoops?! Is that repeatable?!
> That's about as far as I've gotten with it so far, since Squeak is
> just an
> evenings and weekends thing for me. I've also thought that perhaps the
> lost
> clicks thing that others have mentioned in desktop Windows VMs might be
> related
> to what's going on.
>
Definitely yes. The combination of the 20 msecs wait and the long display
cycle can easily lead to lost events and then you'll have to click for a
looooong time.
Andreas
--
+===== Andreas Raab ========= (andreasr at wdi.disney.com) ==+
| Walt Disney Imagineering Phone: +1 818 544 5016 I
I Glendale, CA Fax: +1 818 544 4544 I
+======< http://isgwww.cs.uni-magdeburg.de/~raab >========+
More information about the Squeak-dev
mailing list
|