If you haven't registered for the course, or have registered but still don't have access to LATTE for some reason, you may submit your assignment via email by sending it to both the professor and the TA (Prof. Shrira and Ross). Please register for the course if you haven't yet so that you can submit your assignments via LATTE. Please don't email your assignments if you do have access to LATTE, submit them via the assignment page on LATTE.
If you have registered but can't access LATTE, the first step is to stop by the LTS help desk in the library and report the problem to them, and see what they advise. Please email Ross with any questions you might have concerning submitting assignments.
You can log into LATTE at latte.brandeis.edu. Once you are logged in, you should see COSI-146A listed in your courses in the Fall 2011 section. Click on that link, and you'll see the links for submitting your reports.
Each week you can expect to find an assignment like this one on the cs146a home web page, telling you what you should read over the course of the following week. Occasionally we will be organized enough to predict the readings a few days beyond the next Friday. Numbered readings, such as "reading #2," refer to materials we are distributing.
Before each class you are expected to submit a report. The report is due by 5pm on the day before the class. The report is typically a very short presentation in ppt or a short essay (no more then one page) addressing a topic given on the previous week's assignment sheet. In class you should be prepared to talk about the whole paper, not just the particular point of the report assignment.
Please check the cs146a FAQ for report submitting instructions.
Read the following sections from Simon, "The Architecture of Complexity" .
* Introduction through page 471.
You may skip the sections on "Problem Solving as Natural Selection" and "The Sources of Selectivity" on pages 472 and 473.
* Read from "On Empires and Empire Building" (page 473) to "Conclusion: The Evolutionary Explanation of Hierarchy" (page 473)
The section titled "Nearly Decomposable Systems" (end of page 473 to beginning of 477) is optional.
* Skim from "The Description of Complexity" to the end of the paper, trying discover Simon's point, without necessarily reading every paragraph in detail. Skimming to find the main ideas (and deciding when skimming is appropriate) is a talent worth acquiring!
At first read, this paper appears to have very little to do with computer systems. As you are reading it, try to figure out why it was assigned. It may help to notice that it was written in 1962 and to know that Herbert Simon is one of the founding figures in artificial intelligence, a winner of the Turing Award, a recipient of the Nobel Prize in Economics, and a co-founder of the School of Computer Science at Carnegie Mellon University. Simon contends that hierarchy is a principal organizing force in social, biological, and physical systems.
For your report, we ask you to identify and show in ppt or latex three examples of the use of hierarchy in any of the computer systems you are familiar with, and explain how hierarchy provides significant leverage, advantage or a disadvantage. We ask you to submit your report as a pdf file.
Read from S&K, Chapter 1 "Systems", chapters 1.1, 1.2 and 1.3.
Read the Leveson paper (a shorter version of the original publication ) and a recent NYT article on radiation accidents and write a one-page reading report that addresses the question about "The causes and the remedies of the accidents in critical computer systems":
The Therac incidents took place between 1985-1987. Since then many changes have taken place. For example:
- User friendly GUIs have replaced dumb terminals.
- Better debugging tools and QA strategies exist.
- Government regulations have improved.
- Well defined libraries facilitate code reuse.
- etc.
Now, 25 or so years later, even more life critical systems are interfaced with computers. Yet, as the NYT article shows, it is still possible for such disasters to occur. Discuss the fundamental causes that plagued the Therac-25 and identify the causes that are still relevant today, judging from the NYT article. Many are wondering whether new regulations could help prevent future disasters. Consider regulations in the following areas:
Choose one area and propose one regulation that in your opinion would be most effective in preventing the problems you have identified as common to Therac and the accidents described in the NYT article. Regulation proposals face opposition. Defend succinctly your position.
- Operator training.
- Programming practice.
- Reporting of accidents.
- Device testing.
After reading Leveson's paper you may be flabbergasted by how badly the Therac-25 systems were designed. Our very last reading this year (Lampson's "Hints for computer system design") has a number of suggestions that often lead to better systems. If you have time and energy left, browse through this thoughtful paper. You might want to ask yourself which hints the designers of Therac-25 violated. Don't worry if you do not fully understand this paper yet (or if you only have time to skim it); we will read it at the end of the term.
Read from S&K, Chapter 1 "Systems", chapters 1.4, and 1.5.
System aphorism of the week
There is no such thing as a small change in a large system.
cs146a Handout 1, issued 09/08/2011