What's Left for KCP?
ducasse at iam.unibe.ch
Fri Feb 20 22:39:56 UTC 2004
> I'm really curious how "close" Squeak3.7a-5707 is to meeting the needs
> of the Traits and ClassBoxes teams.
The canUnderstand: problem that nathanael mentioned is still pending and
seriously I think that he was right. (the argument that we should not
have abstract class instantiated
is true in theory but in Squeak even Object is Abstract so imagine how
many abstract objects we create).
Then the systemOrganization has to be fixed. Then adrian and nathanael
has some problems with the system
notification and some changes in Behavior right now.
> I may not remember correctly, but I
> *thought* the main reason for the KCP effort was to support project
> Traits, and ClassBoxes which require a much cleaner/more
> flexible/better-all-around ClassBuilder, etc.
Yes but not only, the put was that Squeak also deserves a cleaner
In fact, this is true that cleaning slow us down and that not giving
would have speed us dramatically but I think that this is important.
> Also of course new
> innovative tools like a Traits-aware Browser, that kind of thing.
> I have been busy with other stuff and haven't kept track of the KCP
> work, unfortunately. How close is Squeak 3.7a? Very close? Half-way
> there? Barely started? Just curious. :)
For traits quite close :)
Adrian is working hard on that. We now have a clean design and
integration of traits in Squeak.
This is nearly working. The refactoring of the compiling protocol was
also a shock for adrian.
More information about the Squeak-dev