A321: Computing Tools for Scientific Research

Syllabus for CS A321: Computing Tools for Scientific Research

Purpose:

This course provides some basic programming capabilities for students in the natural sciences. It is tool and skill oriented, and intended to be sufficient for creating short computer programs and using software produced by other scientists. The primary language is Matlab, a "fifth generation" programming framework that supports scripting, graphics and plotting, and provides a "glue" for accessing and using software written in other languages (C, C++, Java, Fortran). Matlab is valued because it provides a rapid prototying environment for quickly and easily testing out new ideas.

C/C++ is also covered, but primarily how to connect from Matlab to an existing program written in C/C++. The same ideas and general methods apply to Fortran and Java for this linkage.

The course does not emphasize object oriented programming, software design methodology, or higher level design patterns and data structures. Students needing those capabilities are probably better off taking some of the A-courses in the Computer Science Department.

Grading:

The necessary skills define the grading schema:

F: no skills needed
D: The ability to access (read and/or write) data in files into/from Matlab and appropriately plot and graph results. Basic knowledge and use of data represented as vectors and matrices.
C: All in D, plus writing and documenting functions in Matlab. Basic programming control structures (for, while, case), use of simple data structures, and combining multi-file programs to create an executable program. The key distinction between a computer address and the data stored at that address.
B: All in C, plus basic data analysis and statistics on data. This is not a course on the mathematical foundations of stats, and will just cover how to computationally perform common statistical functions to some given data. Use of toolboxes (Matlab) and other libraries. Character and string manipulation. Debugging and advanced use of files.
A: All in B, plus performance engineering, object-oriented ideas, and interfacing Matlab with C or C++. Advanced plotting and graphics, and use of API's for accessing instruments and other data sources.

These skills need to be demonstrated via a sequence of coding and analysis projects.

Textbooks

No textbook is required. If you want supplementary material most of it is online. But if you intend to keep using Matlab it would be worth the money to get the first book listed. Everything needed will be in online or through Matlab's help, lookfor, and other commands.

The books listed are intended to be long-term reference works that can be useful even after the course is over. The course will not cover all of the material in them. You can get along just dandy without any of them, and then when you graduate make your employer buy them for you.

Mastering Matlab, Hanselman and Littlefield. A comprehensive reference work for Matlab, and even those from versions 5 and 6 are still useful (Matlab is up to ver. 7 now).
Matlab Guide, Desmond Higham and Nicholas Higham. This is more of a follow-along text, but assumes more math knowledge than A321/A521 does.
Practical C Programming, Steve Oualline. An easily read introduction to C.
C++ How to Program , Dietel and Dietel. This is more of a reference work - nobody could get through this in a single semester.

Outline of Material

Some flexibility will be exercised depending on the students taking it. So the following outline is not contractual, but instead meant to give a rough idea of what will be covered. The timing of when material is presented may also be shifted depending on questions/needs from the students.

Week 1
1. Matlab basics: Starting and stopping matlab, the windows it provides; Interactive use as a calculator; Variables; Scripting and help facilities (especially whos, what, which, lookfor); ; Case sensitivity, save, load, diary.
2. Motivating example(s) (elementary least squares, fitting a line to a set of experimental data)
Week 2
1. Vectors and matrices: Ideas behind vectors (geometric and from examples) and matrices; Range notation (colon operator); Basic plotting, graphing, visualization.
2. Control structures and ideas: If tests and relational operators, for and while loops, empty ranges.
Week 3
1. Using vectors and matrices in control and graphing
2. More/better plotting, graphing, visualization; Documenting code.
Week 4
1. Functions (versus scripts) and design issues.
Week 5
1. Basic data analysis: mean, std deviation, variances. Plotting error bars and using legends.
2. Built-in functions and toolboxes.
Week 6
1. More detailed Data input and output to/from files; Formatting output.
2. Matlab general catch-up and review
Week 7
1. Timings on a computer;
2. Performance analysis, bottleneck identification. Some Matlab-specific tools for performance engineering.
Week 8
1. Data types and memory allocation in Matlab with foreshadowing of of linking it to C or Fortran later on.
2. Weird cases: NaN, empty matrices, complex numbers.
Week 9
1. Character and string manipulations; Output formatting.
Week 10
1. Critical distinction of call by address or reference, call by copying.
2. Passing functions as arguments in Matlab
Week 11
1. Survey of existing standard libraries; Matlab toolboxes compared to other language libraries.
2. Simple data structures (linked list, queue, maybe a stack).
Week 12
1. Intermediate debugging. Pointers and memory allocation follies in C/C++ and Fortran.
2. Advanced debugging,
Week 13
1. Simple data structures
2. Combining the two: the mex interface and Matlab engine.

Changes to this page:

Initialized, Thu Nov 2 08:54:44 EST 2006
Updated: Mon Jan 7 11:33:32 EST 2008