Mark Guzdial wrote:
> 
> At 10:36 PM -0400 8/19/99, Bob Arning wrote:
> >On Thu, 19 Aug 1999 19:02:16 -0500 Daniel Allan Joyce
> ><daniel.a.joyce at worldnet.att.net> wrote:
> >>Why don't all morphs know how to transform themselves?
> >
> >Probably because it would add a lot of overhead that would be useful in
> >only rare cases.
> >
> 
> Ivan Brusic actually did much of this work in creating his Zoomable User
> Interface (ZUImorph) that allowed any morph to be dropped in and infinitely
> zoomed in or out.  He had a really hard time making it work smoothly across
> all morphs, but he got some really good stuff working.

The problem of having not all morphs knowing how to transform themselves
was one of the major obstacles I encountered.  I basically divided
Morphs into two groups: those that know how to transform themselves and
those that can't.  Andreas' idea of keeping a local copy of the
appropriate transformation inside a Morph's otherProperties field is a
good way to do things, much better my approach which was to use a
Decorator/Wrapper class for every Morph.

Still, this does not solve the problem of the Transformation morph not
deleting itself.  One could check what class its UI parent is before
deletion, but that will not be good OO code.  Does this problem only
occur with Alignment Morphs?  (You will have to excuse my temporary
ignorance, but I do not have Squeak in front of me).

Ivan

-- 

Ivan Brusic                       |  "Computers are nothing but a
perfect
E-mail: ibrusic [at] wsicorp.com  |   illusion of order" - Iggy Pop
Phone: (978) 262-0758             |