EE Student, Edit, Proposal Masters, Help (concepts of functional programming, symbolic programming and MATLAB)

EE Student, Edit, Proposal Masters, Help (concepts of functional programming, symbolic programming and MATLAB)

Post by JohnCreigh » Sat, 12 Mar 2005 06:35:11


erzy Karczmarczuk < XXXX@XXXXX.COM > wrote in message news:<d0oueo$f3f$ XXXX@XXXXX.COM >...
"I share your opinion, Sir.
I would nevertheless suggest that you be a little more concrete. There
was a dozen or so postings under this subject. Raising doubts about
the
competence of people who want to help someone should be addressed more
precisely, pointing out, if possible, *what* was erroneous. In such
a way those apparently uninformed people have some chance to learn
something from a renown specialist.

Excuse me for expressing this opinion in 'an authoritative voice',
I hope you understand my feelings."

Well said. I would like to further evaluate this and say that how when
I first started reading the news group sci.math.symbolic, I was
disipointed to see how most of the posts dealt the syntax of a
particular symbolic packages, the advantages or disadvantages of a
particular symbolic package or how to use a symbolic package to solve
a particular mathematics problem. There was virtually no discussion
about the theory of symbolic packages, how they fit into the current
programming paradigms, history or implementation.

To me this is a shame because what makes someone a good programmer is
not a study of the sytax but a study of the concepts used to solve
programming problems. It is my opinion that outside of computer
science departments there is a considerable lack of attention played
to programming theory. Although this lack of attention may help speed
up implementation it results in inferior products.

I will use MATLAB as an example. MATLAB dominate software package used
in engineering applications because of the speed it can be used to
solve engineering problems. It has a large collection of packages
including from, image processing, neural networks, real-time
application, controls systems, fuzzy logic; which are put together by
experts in the area. However, it doesn't support the most basic
computer science concept of pass by reference which precludes many
data structures and can considerably slow down image processing
applications. This lack of foresight may of lead to a symbolic package
that stores symbolic variables in a string field instead of as a tree.

Considering MATLAB has the top people at numeric algorithms has
implemented the symbolic package in this backward way, I am left to
wonder: Did the creators not thoroughly understand the implementation
details of symbolic packages or was MATLAB not well suited from the
beginning to implement symbolic packages. I presume it is the later.
However, if discussion about computer science theory was more
prominent I presume MATLAB would of long ago modified there software
to allow pass by reference.

My next comment is with regards to the scientific process. I
understand the value of making sure every statement is 100% correct
prior to presentation. However, I think at times this attitude is
taken to far by people who are slightly annual and arrogant. Everyone,
must start learning somewheres and a health discussion of the ideas
one absorbs is an important part of the learning processes whether or
not those ideas are completely accurate. It is not possible for a
newcomer to a field to always understand ideas in the most precise of
terms. Thus they must be presented in a language that my sacrifice
preciseness for simplicity. Thus a slightly confused initial
understanding is an important part of the learning processes and