[ANN] First public release of ContextL

[Pascal Costanza]

Hi everyone,

I have just uploaded a first release of ContextL. It's both a
non-trivial example of using the CLOS MOP and a full-fledged library in
its own right, providing constructs for context-oriented programming.
See http://common-lisp.net/project/closer/contextl.html for more
information.

Cheers,
Pascal

--
OOPSLA'05 tutorial on generic functions & the CLOS Metaobject Protocol
++++ see http://p-cos.net/oopsla05-tutorial.html for more details ++++

[Fernando]

Pascal Costanza wrote:
> Hi everyone,
>
> I have just uploaded a first release of ContextL. It's both a
> non-trivial example of using the CLOS MOP and a full-fledged library in
> its own right, providing constructs for context-oriented programming.

context-oriented programming??

[Pascal Costanza]

Yes. I think we have found a genuinely new approach at changing program
behavior at runtime. There's a paper linked from the ContextL website
that explains the concept. (Of course, there are also previous
approaches that provide similar ideas, but there are also new elements
in ContextL that I am not aware of being available anywhere else.)

[Rainer Joswig]

Am 23.08.2005 12:10 Uhr schrieb "Pascal Costanza" unter <p...@p-cos.net> in
3n0b0cF...@individual.net:

> Fernando wrote:
>> Pascal Costanza wrote:
>>
>>> Hi everyone,
>>>
>>> I have just uploaded a first release of ContextL. It's both a
>>> non-trivial example of using the CLOS MOP and a full-fledged library in
>>> its own right, providing constructs for context-oriented programming.
>>
>> context-oriented programming??
>
> Yes. I think we have found a genuinely new approach at changing program
> behavior at runtime. There's a paper linked from the ContextL website
> that explains the concept. (Of course, there are also previous
> approaches that provide similar ideas, but there are also new elements
> in ContextL that I am not aware of being available anywhere else.)
>
>
> Pascal

Pascal, you might want to look at some stuff that groups in Hamburg
did (Plakon and later Prokon). The domain was 'planning and configuration'.
They used Common Lisp for configuration problems - there might
be some ideas there. They used views and mixins and a lot of other stuff:
http://www.hitec-hh.de/ueberuns/home/aguenter/literatur/konwerk.pdf
http://www.hitec-hh.de/ueberuns/home/aguenter/literatur/prokon-buch-kurz.htm
l
http://www.hitec-hh.de/ueberuns/home/aguenter/literatur/wr-kurz.html
and so on...

[Cunis et al. 87]
Das PLAKON-Buch
R. Cunis, A. Guenter, H. Strecker (Eds.)
Informatik Fachberichte 266, Springer, 1987

'Maybe' also interesting for you:

Jürgen Kopp, Konstruktion von Wissensrepräsentationssprachen durch
Nutzen und Erweitern objektorientierter Sprachmittel, DISKI 13

[Pascal Costanza]

Rainer Joswig wrote:

> Pascal, you might want to look at some stuff that groups in Hamburg
> did (Plakon and later Prokon). The domain was 'planning and configuration'.
> They used Common Lisp for configuration problems - there might
> be some ideas there. They used views and mixins and a lot of other stuff:
> http://www.hitec-hh.de/ueberuns/home/aguenter/literatur/konwerk.pdf
> http://www.hitec-hh.de/ueberuns/home/aguenter/literatur/prokon-buch-kurz.htm
> l
> http://www.hitec-hh.de/ueberuns/home/aguenter/literatur/wr-kurz.html
> and so on...
>
> [Cunis et al. 87]
> Das PLAKON-Buch
> R. Cunis, A. Guenter, H. Strecker (Eds.)
> Informatik Fachberichte 266, Springer, 1987
>
> 'Maybe' also interesting for you:
>
> Jürgen Kopp, Konstruktion von Wissensrepräsentationssprachen durch
> Nutzen und Erweitern objektorientierter Sprachmittel, DISKI 13

Thanks a lot - I will definitely check this out...

[Tayssir John Gabbour]

Fernando wrote:
> Pascal Costanza wrote:

> > I have just uploaded a first release of ContextL. It's both a
> > non-trivial example of using the CLOS MOP and a full-fledged library in
> > its own right, providing constructs for context-oriented programming.
>
> context-oriented programming??

Interesting. From a Lisp perspective, it appears to be the holy
marriage of dynamic scope and CLOS. In addition, you can incrementally
define extensions to classes and methods which are only used in certain
dynamic scopes ("layers") that you set up.

So, stealing the paper's example, you might define a class representing
a person. And then incrementally define the parts of the person
relevant to her employment, which will only come up in an employment
layer (where it's relevant).

There's also the cool part of using SLOT-VALUE-USING-LAYER to update a
GUI only when the layer is active.

I hope I got all that straight... I think it's genuinely interesting. A
well-structured way to use dynamic scope.

Tayssir

[Pascal Costanza]

Tayssir John Gabbour wrote:

> I hope I got all that straight...

I hope I got it straight as well... ;)

Pascal

P.S.: Yes, your summary matches our ideas.