Hi Andreas --

Also take a look at what and how Bill Atkinson did for Quickdraw on 
the Mac. He used pixel boundaries and many other really interesting 
techniques when he did this work. A funny story is that he did all 
this because he thought that we had done the Smalltalk graphics 
"perfectly" so he wound up getting an interesting and useful result 
by trying to reinvent it.

Cheers,

Alan

-----

At 9:27 PM +0200 4/6/04, Andreas Raab wrote:
>Thanks Dan!
>
>>  I have always strongly disliked this particular convention.
>>
>>  Someone else was in charge of arithmetic (mainly LargeInts)
>>  when we settled on the ST-80 standard and I withdrew from the
>>  fray after expressing my opposition.  I was told that a lot of
>>  mathematicians (which I am not, in the formal sense) expect
>>  integer division to truncate toward negative infinity, and I
>>  had to trust that this was a good decision for Smalltalk.
>
>Do we have any mathematicians around here who could shed light on the issue?
>
>>  I seriously considered "fixing" this when we did Squeak,
>>  but it would have been a disastrous incompatibility with
>>  the rest of the community.
>
>Oh, really - is this in the blue book? Actually, I wouldn't worry that much
>but I just had this interesting experience when I tried to change WarpBlt
>and suddenly, when I rotated the guy everything was totally off ;-)
>
>BTW, actually you might be able to help me here - the thing I was trying to
>do is to change warpblt so that instead of operating on "pixel centers" it
>works on "pixel boundaries" as input (e.g., "corners" instead of
>"innerCorners"). The problem with pixel centers is that when you throw in a
>generic transform and try to figure out "what the pixel centers are" there
>isn't really nice and effective solution (at least none that I could find)
>whereas when you operate on pixel boundaries you can easily adjust for this
>in the sampling of WarpBlt itself (it actually simplifies the code ;-)
>
>But if you have any idea how to map the "outer corners" via a full (say
>matrix) transform to pixel centers so that WarpBlt creates pixel-precise
>results I'd be really glad to hear about it.
>
>Cheers,
>   - Andreas


--