Dear Trygve,

I've played with your star field .image, looks good :)

Also, browsing senders of #Role: and of #<role-name> shows what's going on  
and where things belong to. This is one of the advantages of declaring  
pragmas in methods.

I'm looking forward to your next steps, perhaps with role-aware Traits?  
(which then could be attached everywhere :)

Cheers
Klaus

On Mon, 17 Dec 2007 17:41:08 +0100, Trygve Reenskaug wrote:

> Dear Klaus,
> Thank you for pointing me in the right direction. It proved fairly  
> straight forward to open for programming with dynamic roles by hacking  
> the Parser/Encoder. I also added some other extensions to make the roles  
> visible in the debugger and to show them with the right value.
>
> I used this extension as the foundation for a demo and talk at the ETH,  
> Zürich, Switzerland. The complete program can be loaded from two  
> packages:
>        http://folk.uio.no/trygver/babymc/BabyPragmaroles-TRee.2.mcz
>               that shows what I have done to the Parser etc.
>        http://folk.uio.no/trygver/babymc/BabyETH-TRee.11.mcz
>               that contains the demo itself and illustrates my first  
> excursion into programming with roles.
> If anybody should want to load these packages, BabyPragmaroles must be  
> loaded first, otherwise the BabyETH methods will not be properly  
> compiled.
>
> But loading is not necessary because a fully loaded image is stored in  
> the following zip file that contains image and changes files (note:  
> Squeak 3.10.6 which is Squeak3.10.gamma.7159):
>        http://folk.uio.no/trygver/2007/BabyETHImage.ZIP
>
> I post this to the squeak-dev list in case anybody would like to see a  
> snapshot of what I'm up to these days.
> I've added a lot of comments that may explain more than they confuse.
>
> My next steps: Larger examples and badly needed programming tools.
>
> Enjoy
> -- Trygve
>