Subjective Squeak
Stephane Ducasse
ducasse at iam.unibe.ch
Tue Dec 10 20:58:39 UTC 2002
Hi anthony
If you want by any chance that the blockclosure gets accepted in the
Squeak kernel
as I would like. PLEASE do not mix research ideas with that. Fix all
the problems you know
or others know (decompiler....) first
I think that the validation is really important. May be PIE is cool may
be not.
I should say that I red the papers and find that like a cool idea of
the 70 ;).
Do you really think that we will be programming today in a system like
PIE?
Please do not misinterpret me, I think that experience is ***really***
important, so
what they did is great because they were really in advance. Still I
think that all
these subject-oriented programming papers that were at OOPSLA few years
ago before AOP
were cool but I would not like to program with that because it was
complex.
With Traits we are really convinced that they are worth
because of the properties the model has
- simplicity
- the fact that you can flatten the traits
- discussion with mixin experts such as Gilad that agree about the
lack of scalability
of the mixins
- a kind of consensus around them. We discussed with some other
Smalltalk implementors
and they really like the idea to the point to see how this could fit
in their Smalltalk.
- the experimental validation we are doing. We plan to have a
mini-image
fully traits aware.
Still with all that I want to have a mini-image fully traits based to
feel it.
By the way I tried to find the Us paper but could not find it on the
web, does somebody has an electronic version?
Stef
On mardi, décembre 10, 2002, at 09:26 pm, Anthony Hannan wrote:
> Hello Squeakers,
>
> What do you guys think of adding subjective layers a la PerspecitveS,
> Us, PIE, to the Squeak kernel (part of method lookup)? Beside adding
> the power of context sensitivity, layers would also serve as packages
> in
> the image. I really think layers will improve our package model in the
> image, particularly with respect to dependencies, versions, and
> unloading capabilities.
I think that we do not need that for that. Basic software engineering
principle
and namespaces are enough for that. May be Layers could solve that but
may be
they are introducing another layer of complexity just at the cognitive
level.
(This what was worrying me the most with traits, the model is simple
still
you have one extra concept)
> Dependents would be layers underneath; versions
> would be single layers on top; and unloading/loading would be automatic
> just by switching layers/perspectives.
>
> I prefer Us's extra lookup dimension over PerspectiveS's method wrapper
> dimension, because I think it's a more directly implementation of the
> concept. With the method cache (and eventually the inlining JIT
> compiler), the extra lookup dimension should not slow things down
> significantly.
>
> The layers I would like to implement would be different from Us's in
> the
> following ways:
> (1) Layers would be multiply inherited instead of singly inherited.
> (2) Layer lookup points would be explicit named variables, not implicit
> in every slot. For example, in 'World color': #World would be looked
> up
> in the layer hierachy, as well as #PasteUpMorph and its superclasses,
> but the 'color' slot would not be looked up in the layer hierachy; it
> would just be gotten from the object like it is today. So, unless
> #World is overridden in a layer, or #PasteUpMorph or #Morph is
> overriden
> in a layer with a different behavior for #color, you will get the same
> original color value. PerspectiveS also takes this approach of only
> layering behaviors not slots.
>
> For more info on my design, please check out my Projects section on my
> webpage at http://minnow.cc.gatech.edu/squeak/AnthonyHannan. Also,
> check out the Us paper at
> http://citeseer.nj.nec.com/smith96simple.html,
> and the PerspectiveS paper at
> http://www.prakinf.tu-ilmenau.de/~hirsch/Projects/Squeak/PerspectiveS.
>
> I you guys like this, I will start working on it for the next version
> of
> VI4. Of course, I would appreciate any volunteers to help.
Arghhhhhhhhh
Please separate the two. Do not be naive to the point that people will
all embrace
block-closure and LayersUSAOP......
Stef
Sorry to be a little rude but I really appreciate your effort to add
block-closure
to squeak don't mess it up. But still have fun so build your stuff on a
separate
VM.
>
> Thanks,
> Anthony
>
>
Dr. Stéphane DUCASSE (ducasse at iam.unibe.ch)
http://www.iam.unibe.ch/~ducasse/
"if you knew today was your last day on earth, what would you do
different? ... especially if, by doing something different, today
might not be your last day on earth" Calvin&Hobbes
More information about the Squeak-dev
mailing list
|