Assignment 7

take a look
a) at the website of our guest lecturer Amer Diwan at: http://www-plan.cs.colorado.edu/diwan/

b) articulate one question for the guest lecturer (e.g.: you may want to take a look at one of the numerous interesting papers of Amer)

Michael Otte – You are involved in many different areas of research, what advice can you give about working on multiple projects at the same time?

Jeffrey LaMarche – Overall there has been a decline in the number of CS students since the dot-com bubble burst. How do you think Computer Science education can be improved to entice new students and also retain the students who are already part of the major?

Lee Becker – Your paper describing the conversational classroom strongly reminds me of techniques used in TESOL (Teaching English to Speakers of Other Languages). Aside from the domain knowledge, what aspects of computer science education are truly unique to computer science?

Saroch Panichsakul – You have a number of papers about computer science education. What are your opinions towards computer science education here at CU? Is it okay in your views when compared to other superior institutions, if you have experience with it?

Paul Marshall – You seem to have been very successful at collaborating with researchers at other institutions and companies - do you have any advice on how to cultivate such relationships?

Jinho Choi - Do you think if it is possible to build a compiler with some kind of intelligence? Compiler gives very good ideas about what kind of errors there are for the most of time. However, there are situations that compiler becomes less useful: when there is one semicolon missing, it treats many lines below as errors. Have people tried to develop a compiler such that it gives suggestions for such mistakes?

Dan Knights - The proposed garbage collection scheme described in the dissertation, "Connectivity-based garbage collection," of one of your students, seems to be worth using. Are there plans to try to implement it in a real JVM?

Yifei Jiang - You are maintaining Programming Language Preliminary Exam Syllabus in your website. – Do you have any advice on how to prepare preliminary exam, especially in the area of programming language?

Keith Maull - The work in your paper "Catch Up" (2005 : Henkel, Diwan) was very interesting. What is the status of this work now? Allowing easy ways to synchronize and update client code with changes to dependent libraries is highly valuable. Is there any work underway to couple such concepts with suggestor systems to _improve_ code to that utilizes refactored libraries, for example, to suggest more efficient methods of doing the same work with the refactored library?

Yuli Liang - I'm more interested in the PL-Detective. I learned programming by understanding the examples and I can do it by myself. If PL-Detective provides a great informative environment, what make students want collaborate with each other while they avoid to do so in the first place?

Nwanua Elumeze - There's the argument that "Lisp makes one a better programmer", and I agree with it. Apart from sheer momentum, is there any reason why CS undergrads still have to slog through C as their first programming experience at CU? In other words, am I missing something about C?

Mohammad Al-Mutawa – when we compare the performance of high level languages (say python) with C (a middle level language), C will be faster and more efficient most of the time; why is it difficult to make/have a high level language that can produce applications that are as fast and efficient as applications produced by C?

Ashok Basawapatna- When you researched student culture and uncovered a strong student preference for solo assignment work– It lead me to wonder if you uncovered any specific set of situations within the "status quo" classroom wherein students naturally wanted to collaborate (ie: the collaboration didn't have to be fostered)?

John Michalakes – How can one believe performance measurements in the face of perturbation from the instrumentation itself? (This is a softball question; I heard your talk at the FRACTAL workshop this past Sunday).

Holger Dick – Referring to "Student Culture vs Group Work in Computer Science": Have you any experiences with Extreme Programing? It would be interesting how students think about Pair Programming, where collaboration is "natural" and not something an assignment or a teacher asks them to do. Also, it would be interesting how the "student culture" differs when they do work for fun, thus where they cannot get credit from a teacher or have to take responsibility for bad grades.

Jane Meyers - In all undergraduate computer science curricula I have seen, the introductory classes have primarily consisted of programming intensive courses. Although these are important, I would argue that they present a rather limited introduction to this field. What do you think of this?

Joel Pfeiffer - Using a debugger (such as gdb) is inherently difficult on optimized code. Have any ideas been proposed for this or has an attitude of "get it working, then optimize" been mostly adopted.

Shumin Wu - While taking a formal language class as an undergrad, one of the students said he didn't see the point of the theoretic computer science class as he had no interest in compiler design. How do you convince students the value of such topics when they do intend to pursue a graduate degree in computer science?

Soumya Ghosh - I have always found formal specifications rather cumbersome and difficult. Under what circumstances would you say that the investment in formal specifications well worth the effort.

Caleb Phillips - I see that you have previously taught this course (Ph.D. Intro). From your experience teaching it, was there any activity or assignment that you consider to have been especially useful for (us) fledgeling Ph.D. students?

Dola Saha - Is profiler-based optimization better than compiler-based optimization? Are there any specific cases (benchmarks) where one performs better than the other?

David Gnabasik - Do you think a Theory of Software Extinction, systematically placing lifetime limits on software, would place an upper bound on the unlimited growth of software complexity? Could this ever be something more than a managerial function?

Last modified 18 October 2007 at 9:06 am by dgnabasik