Topic # 1: What Should the Agenda Be for Computational Science Education?
----------------------------------------------------------------------------------------------------------------------
COMPUTATIONAL SCIENCE EDUCATION eMAIL NEWSLETTER
Sponsored by:
Institute for Modeling and Simulation Applications, Clemson University
Shodor Education Foundation, Inc., Durham NC
Supercomputing 2008 Education Program
----------------------------------------------------------------------------------
This seems like a good way to start a discussion going: "What is a
reasonable agenda for us?" This question is probably as old as
computational science itself. On the one hand, domain science is
important and should be fit into any program. There are some issues
that need resolution.
1. We all know that STEM enrollments are in decline. How can
computational science help reverse this decline?
2. There is a deep gulf between secondary science and introductory
post-secondary science. How do we bridge that gap?
3. How much architecture, programming, and systems systems education
is needed? [Not necessarily taught by computer science].
4. It has been very difficult to get engineering faculty involved in
computational science education. Steve Gordon is leading the effort
to change this. Please forward ideas on this issue.
5. What training do graduate students really need to have to
develop reliable, verified, and validated models and simulation?
6. I'm running into a large number of engineering grad students who
have only Matlab experience. When their theses/dissertations are
more than what Matlab can handle, how do we bring them to minimum
proficiency? What is minimum proficiency?
* * * * * * *
NOTE: Anyone can SUBSCRIBE to the Imsa-Ed Mailing List by going to:
http://www.ces.clemson.edu/mailman/listinfo/imsa-ed-list
Subscribe to:
Post Comments (Atom)
.jpg)
3 comments:
#3
I think as we look at what should be taught regarding architecture, programming, and systems education, that this is something that needs to be part of the canon for modern science majors. At a minimum, science students need to be able to understand issues with their computer in terms of storage, RAM, CPU, and cache. Additionally, students should be able to express mathematical logic in a computer language in order to solve problems.
For students who intend to specialize in computation, I thin they also need data structures and software engineering.
I think that this is appropriate for a CS department to teach, but I would like to see CS departments offering special sections for STEM majors that use science problems to motivate the class.
Rubin always presents the survey data for Physics, showing that so few departments require any CS courses. This might have been manageable 10 years ago when high schools still taught programming in their computer skills courses, but I almost never get any students who have seen computer programming when entering their freshman year--the high schools teach computer skills now instead of computer programming.
So much to comment on, so little time! So just one point:
joinerdav writes, "the high schools teach computer skills now instead of computer programming."
What is "computer programming"? Is it a year of Java, with emphasis on the abstract notions of OO and classes?
"Careful and methodical development of Java applications and applets from specifications; documentation and style; appropriate use of control structures; classes and methods; data types and data abstraction; object-oriented programming and design; graphical user interface design." Copied from an intro course description, I won't tell you where. :-)
Would this turn on high school students to computational science?
Or would it send them streaming (screaming?) in the opposite direction?
There are a couple issues that I think should be addressed.
1. "Careful and methodical development of Java applications and applets from specifications"
In my opinion, this would be fine if the writing of specifications came first. There's not a book out there that honors that commitment.
2. My (Clemson's School of Computing) real experience is that the majority come in with the wrong attitude: correctness is nothing and lines of code is everything.
3. Computing is not being used in secondary science --- it's all being taught by business faculty.
4. For computational science, "correctness is the only thing." It's called "verification and validation" or V&V, sometimes VV&A, and sometimes VVR. Regardless, "how do you know the simulation is correct" is the overriding issue.
5. How would I know that a student has learned? They can explain what they did in clear, technical English. Good luck on that.
6. Intro text books all have "problem solving" in the title but they never teach design / problem solving. The computer science community just keeps lying to itself.
Post a Comment