Jecel Assumpcao Jr wrote:

My dearest Jecel:

I am afraid I won't be able to give you more than a very short response
and I am not sure most members of this list are interested in this
discussion.

> Sorry, since it is so few bytes I though t there would be no problem
> at all!

Your email was 81 KBytes.

For me this is too little!!!
 
> > You presented a very large number of analogies, but no concrete
> > advice at all.
>
> First of all it is a position paper!!!! Soon there will be available
> a Scientific Report with 200 pages!!!

Ok.

> > That might be appropriate if the presentation will be in a
> > mostly philosophical forum.
>
> Come on!
> I really do not understand you....What do you want to convey?
> All the references I gave in the paper come from active
> software developers!!!

All I was saying is that some conferences, like POPL (Principals or
Progamming Languages) are more philosophical while others, like OOPSLA,
are more practical. I don't know what the workshop you sent this paper
to is like, so I was saying that it was probably ok as it is.
 


OK!!In this site you can find the paper in the conttext of
the Object Oriented Architectural Evolution Workshop. They are
as practical as any other computer scientists.

You will find a link to your IWPSE paper on
http://prog.vub.ac.be/OOAE/ECOOP2001/ECOOP2001-OOAE.HTML
 
 
 
 
 

 
> Yes, they do not try because they cannot even understand it!!

Check out these student's experiences with Self:

   http://www.daimi.aau.dk/dAOP/report-group4.pdf

There were some things they liked, a lot that they didn't. But they
didn't have the kinds of problems you seem to imply the "most important
researchers" are having with the language.

I will check this later on! However I am not inventing anything. Because
I am sure Self is even more important to implement my ideas than for
you !!!Not only me but the whole mankind is being harmed by the status
quo of computer science!!! In the Communications of the ACM April
2001 vol 44 number 4 there is a comparison of object- and
agent-based approaches.

Page 39 An agent-based approach for building complex software systems.

For the author complexity frequently takes the form of a hierachy.
Well this says all!!! That is, a system composed of interrelated subsystems
each of  which is in turn hierarchic in structure, until the lowest level
of elementary subsystem is reached.
Well complexity in architecture does not have this nature!!!Hence what
computer scientists want to implement even if it is complex is yet
very simple compared to architectural needs. Well sustainable cities
must be built immediately and all computer scientists live in houses!!!
So would you tell me why computer scientists do not care at all for
systems of this complexity?For me it is obvious!!!!

So the author concludes that object oriented paradigm is not so advanced
as agent-based approaches based on his knowledge of class-based
languages!!!!!!!!!!??????????????????

Prototype based languages are so easy and simple that 99% of the
community does not know it!!!!!!!!!!!!?????????

The the author argues We find that objects, classes and modules
provide an essential yet insufficient means of abstraction.
Individual objects represent too fine a granularity of behavior and method
invocation is too primitive a mechanism for describing the
types of interactions that take place.
Recognition of these facts led to the development of more  powerful
abstraction mechanisms such as design patterns, application frame
works and componentware.
 
 

Although these are undoubtedly a step forward, they fall short
of the desired characteristics for complex system development.

By their very nature, they focus on generic system functions and
the mandated patterns of interaction are rigid and predetermined.
Basically relationships are defined by static inheritance hierarchies!!!!!
 

Well I am not a computer scientists but I know very that  prototype
based languages are a far cry superior to agent-based approaches!!!

I abandoned a BlackBoard architecture because there was Beta and
later on much better Self on the horizon!!!!!!!

I am also aware that computer scientists that deal with agents are
also hermeneutic computer scientists and hence the most advanced
computer scientists!!!Can you give me a good explanation why
this so advanced computer scientist does not know prototype
based languages????

Moreover why does he ignore current approaches in OO paradigm
such as that everybody is trying to mimic delegation, separation of
concerns, composition of concerns, the introduction of layers and
views and so on?
Why he has never heard of Self ?US, Kansas and so on?
Simply because the domain model (it it exists!!!) is simple
enough to be implemented in agent based approach ....while
my domain model requires  not only all the best that Self has
but also layers, views, crosscutting  concerns and so on!!!!
Human interaction programming alone or in group is fundamental
to keep the hermeneutic structure of my ecodesign model.
That software is a cooperative game is being emphasized by
those that bet on X-programming!!!
I wonder how they manage to do so in Smalltalk, C++, Java and
so on!!!!They compare in the report from our Workshop OOAE
that X-programming is similar to jazz!!!!!

Well, I spent the whole Sunday afternooon trying to understand
how can one implement design patterns as layers (Jan Bosch's
LayOM)  and if this concept maps onto Us's layers.

Moreover I tried  to understand how they are implementing
layers and views in C++!!! Well this is implemented only
at compiler time or at run time. What does this mean?
In LayOM the layers are explicit, in C++ you play as if
there are layers in your  classes!!!And it is so confusing
so nasty to read all this...!!!!!
For me SELF/US are very nice reading!!!And what I have
implemented in Self is also fun!!!
 
 
 

 

> For
> example AspectJ I have been trying to read something about it and
> attended an excellent demonstration at ECOOP'01 and I cannot
> understand it!!! I need to know much more.

You are an architect, not a computer scientist. So it is no big deal
you didn't understand it. But if someone claiming to be a computer
scientist took more than five minutes to grasp what AspectJ is all
about then I would suggest they consider trying a different profession.

Now *understand* can mean several things. The first large Pascal
program I ever saw was a chess program published in an old Byte
magazine. The author had filled it with label declarations and gotos so
that it was no better than any unstructured Basic code. It was obvious
that he didn't understand what Pascal was all about. But he did
undestand Pascal in the sense that he wrote a large working program in
the language. I will not be at all amazed if you tell me many people
don't "get" Self, but can't believe you if you say famous OO
researchers are unable to program in it.
 

Well I just gave you an example above!!!I am absolutely sure
applications that can be implemented in agent-based approaches
and reflective architectures can be much simpler implemented
in  SELF/US with reflection!!!!

Why they disregard SELF/US????
I think the need for an architecture as I am trying to build one is
indeed necessary for complex applications.

Anyhow I continue getting inspiration on SELF/US and I wish
the best success for the language and its community. I am also
part of it of course.\

Good evening,
Albertina

 
-- Jecel
 
 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/

-- 
.----------------------------------------------------------.
| Albertina Lourenci                                       |
| PhD  in Architecture and Urbanism                        |
| post-doctorate researcher                                |
| Laboratory of Integrated Systems University of Sao Paulo |
| Avenida Professor Luciano Gualberto, 158 Travessa 3      |
| CEP: 05508-900                                           |
| Sao Paulo Sao Paulo State Brazil                         |
| Voice: +55 011 818 5254                                  |
| Fax: +55 11 211 4574                                     |
.----------------------------------------------------------.