Quoting Ruedi Steinmann <ruedisteinmann at gmx.net>:

> On Mon, 2005-02-14 at 21:34, Ruedi Steinmann wrote:
> 
> > dispatchEvent: evt
> > "Place to hook Morph specific event dispatching in"
> > "evt someTest ifTrue: [^ self morphSpecificEventDispatchMethod: evt]."
> > ^self dispatchDefault: evt.
> > 
> > Morph>>dispatchDefault: is the same as
> > MorphicEventDispatcher>>dispatchDefault:with:
> 
> should be: 
> 
> Morph>>dispatchEvent: is the same as
> MorphicEventDispatcher>>dispatchDefault:with:
> 
> To hook Morph specific event dispatching in, Morph>>dispatchEvent:
> should be overridden like:
> 
> dispatchEvent:evt
> event someTest ifTrue: [^ self morphSpecificEventDispatchMethod: evt].
> ^super dispatchEvent: evt.
> 
> Ruedi
> 
> 

This is interesting that you look at improving the Morphic Event Dispatching.
But I assume that most people (including me) do not follow what you are doing.

A textual representation of some method calls does not really shed much light at
a design question.

A object interaction graph would or some other type of graph woul help (use
Connecters ... hint, hint).

And how does your proposal fit in the scheme outlined under
http://minnow.cc.gatech.edu/squeak/118

It seems that the whole issue is complex. So if you could shed some light on how
it is working today (by verifying and updating the swiki), that would be already
a big leap forward.

Hannes


P.S. And thank you for reviving the MCP. I think it could start its life as an
cs and a mcz package that people have to add to their base image. You might get
users for it and others will build on it, it will become a prerequisite for
other packages and finally it may sink into the main release.

Cf. Yoshiki's m17n package which is now in 3.8g.