Pendulum

[Alan Kay]

Hi Bert --

That is a very good point! I'd forgotten that embedding makes an 
object live in the coordinate system of its owner -- and that is what 
it is supposed to do. However, it's clear from this and other 
examples that I can think of that this perhaps should be the default, 
but it might be a good idea to put a flag on a player to "use the 
coordinate system that its owner uses".

Let me think about this over the weekend ...

Cheers,

Alan

--------

At 1:40 PM +0200 4/19/03, Bert Freudenberg wrote:
>Am Freitag, 18.04.03 um 22:51 Uhr schrieb Alan Kay:
>
>>Hi Phil --
>>
>>At 5:38 PM -0400 4/17/03, Phil Firsenbaum wrote:
>>>Interesting...I actually saw this message on the Squeak archive. i 
>>>didn't receive it on the mailing list, though.
>>>
>>>Anyhow, even if the pen were down and the pendulum simulated 
>>>reality i think it would draw lines on top of lines  unless the 
>>>area under the pen was scrolling. Do you see what I mean?
>>
>>Well, here's a little exercise in relativized thinking ...
>>
>>Embed a round little object like an ellipse to be the bob of the 
>>pendulum, call it "bob". Make another little object called 
>>"plotter", put its pen down. See what happens when you do:
>>
>>                 plotter's y increase by 1
>>                 plotter's x <- bob's x
>
>"Relativized thinking" is fine, but the implementation is itself 
>relative, so this does not work.
>
>After embedding, the bob lives in the coordinate system of its 
>parent, the pendulum. Rotating the pendulum does nothing to the bob 
>- just look at its geometry category, it's static. I then tried 
>using a PolygonMorph as pendulum because it does not use a 
>TransformMorph, but then, the transform of the embedded bob goes 
>wild (try it).
>
>The latter behavior looks like a bug, while the former is probably 
>supposed to behave like this. We would have to introduce a "global" 
>or "absolute" position to make this work.
>
>-- Bert


--