[aosd-discuss] Paper: Roles and Aspects:Similarities,Differences, and Synergetic Potential

[Juri Memmert]

--- Albertina Lourenci <lourenci at lsi.usp.br> wrote:

Hi Albertina,


> This chat is very nice! Why is it not exactly enlightening?

Hmmm... Is it not?


> Simply because in my viewpoint software systems are dependent
> on applications. I was taught that by computer scientists (that were

Yes... software systems are dependant on applications, which can not be
realized without an underlying system. It's a circular system and that allows
us to talk about the one or the other and both are important and valid parts of
the whole.


> electronic engineers - that's a great difference! maybe) om 1972!

Yes.


> PHDs from Berkeley, Stanford, Imperial College and so on!
> These people were interested in technological independence!

Meaning what? Independance of what? Independant of what?


<snip: off-topic>

> I believe I will have no difficulties to make architects and urban planners
> realize why the diagram for roles has nothing to do with crosscutting
> concerns at least if one implements them in pure object oriented languages.
If

And roles are not cross-cutting? Since when?
Example:

Class Director
Class Worker
Role Employee

Since when is Employee not cross-cutting Director and Worker?


> you start seeing things from a class viewpoint...everything gets distorted!
> It seems computer scientists prefer this than to face the problem of identity
> of the object as  Daniel Bardou points it out in his paper about split
objects!

*reads the paper* ... which is one way to look at separating the different
concerns and then merging them together.
Doing it in an inheritance (or delegation) hierarchy as proposed in the paper
leads to hierarchy hardening and limited evolavbility... this is why delegation
can be replaced with AO technologies and the result will work better.


> The attached diagram about my ecodesign model is isomorphic to the role
> diagram!  It is easy to see there is nothing crosscutting here because
> each process is completely different for each element! continuity

Cross-cutting means, among other things, that completely different things
happen to the same class, influencing it in different ways. These different
things cross-cut in the class.

So, your waves cross-cut in the object /class they are applied to.


> in thermal function deal with the mesoclimate and continuity for
> structural systems deal with the geology of the region, and so on!
> Then it has nothing to do with quantification!

What has cross-cutting to do with quantification?


> Even if you think about the sites of a city, each site has unique
> microclimate and unique geological conditions. This demands qualitative
> reasoning, and morphodynamical reasoning naturally emerges!

Yes... each site has its own microclimate... and so the climate cross-cuts all
sites. In different sites, the microclimate may be different, but all these
different microclimates are still microclimates and thus can be seen as one
large concern with many different expressions in different sites.


> It is impossible to aspectize these constitutive concerns of the objects!

Wrong. It can be done. I know. I'm the son of an engineer for facade technology
and could read his drawings before I could read the newspaper.

They _do_ have different schematics for the same element, based on the various
different concerns they need to look at.


> However they must be woven together to generate architectural and urban
> design and planning!

Yes. Woven together _after_ they have been defined in smaller units and
compromises between different concerns have been made.


> In two years I hope I will be able to have an international edition of my
> ideas and put forward this clearly!

We'll see, then.


> Only Greimasian semiotics will enable computer scientists to solve this
> conundrum. Applying Greimasian semiotics, I designed the attached
> semiotic square. It clearly shows there are four types of processes dealing
> with concerns:
> separation of concerns, parallelism of concerns or common concerns,
> crosscutting of concerns and composition of concerns.

And? This is nothing new to me. So, what's your point?


> In my ecodesign model there are concerns attached to the architectonic
> object:
> Concerns atttached to the neighborhood, to the ecocity, to the bioregion!
> Concerns are the constituent elements of the object it doesn't matter
> how big or large it is! It is not incompatible with the class concept!

What has size to do with anything here?


> But a class is not a synonym for an object! It would be worth to semiotize
> it, but I will do this when I will be lecturing  in Urban and Regional
> Planning in Porto Alegre next year. This takes time!

A class with all concerns attached is the best abstraction we currently have.
And no, it's not as good as the original... but it is good enough to simulate
an atom bomb, sparing us the horror of more explosions on remote islands. The
abstraction is pretty close to real, I'd say.


> Now separation of concerns, crosscutting concerns, and composition of
> concerns happen at the architectonic object, the neighborhood, the ecocity,
> the whole Mother Earth!

Yes. And?


> To try to erase the concept of object is the same as to try to erase the
> concept of thing (see Heidegger, Kant and so on, this is impossible!) The
thing
> pervades everything!

Who wanted to erase the object?


> I see  there is nothing bad with objects, classes and aspects, and concerns
> and so on!

Neither does anyone of us. But we _do_ see that there is a problem if you deal
with apples in one place and with pears in another place when the things you're
dealing with (apples and pears) are basically the same thing on different
levels of the development process.


> The main hurdle is the inability of many to see things from the viewpoint of
> the thing!

Which is what OO is all about.



Regards,

     Juri

__________________________________________________
Do you Yahoo!?
HotJobs - Search new jobs daily now
http://hotjobs.yahoo.com/