Guest Author’s Profile:-

Dr. Anu Vaidyanathan is a visiting faculty at the Department of Computer Science and Engineering at IIT Ropar. Prior to joining IIT Ropar, she was the CEO of PatNMarks, the first Indian company to offer Value Added Services in the area of Intellectual Property. She holds a PhD in Electrical and Computer Engineering from the University of Canterbury. She had a brief stint at IBM Austin after obtaining a Masters degree in Computer Engineering from North Carolina in 2002. She has many publications to her credit. An avid athlete, she has founded two student groups, iWiSE (Indian Women in Science and Engineering) and the Ropar Runners (a running club) which in conjunction hopes to change the nature of the canal close to campus and learn more about the place of education in rural India.

Article:-

In the 2010 July Semester, I had the opportunity to teach a class on Computer Architecture to a set of second year Computer Science and Engineering students at the Indian Institute of Technology, Ropar. In watching my own professors teach and discussing their methods, choice of books and assignment of grades based on various components in a course including projects, exams and assignments, I have come to understand that there are many paths to choose from. The nature of the disciplines of Computer Architecture and Systems is closely tied to design followed by effective implementation and robust testing. Additionally, students should ideally see a holistic view of how a high-level program goes through the motions of being compiled/interpreted before being executed. In teaching Computer Architecture, I was faced with the thought-provoking task of communicating the basics, including digital logic design, while challenging the class to think about the specifics. I went about designing the class content using the following general outline:

a.  First, I decided to gauge the interest and final degree goals of my students by having them indicate this in their first assignment. In the same assignment, I also asked them to do some independent research on the ioctl() system call, to gauge how they gathered information. Thirty Six out of the Thirty Seven indicated a Masters Degree to be their ultimate degree goal. There was one student who wanted to get a PhD. Using this as input, I came up with a course-outline from Henessey and Patterson’s 4th edition (teaching the ARM ISA) on computer architecture with a significant project component. I picked this book over Hamacher and Zachy, which has a more gradual introduction to the concepts, given that my students are just being introduced to Digital Logic Design, this semester because if they intend to go get Masters degrees then this is the best resource to challenge them with.

b. The SimpleScalar toolset is usually introduced in graduate school, to familiarize computer architecture researchers with the ins and outs of processor design. This closely follows the MIPS instruction set and runs with a variety of SPEC benchmarks to gauge performance. I introduced the second year students to this toolset for their project, thinking that if they struggled with installation for more than 10 days, I would revert to setting an easier project. My class surprised me with some great work on utilizing the TAs help to set up SimpleScalar quickly and even came back to me with some insightful questions.

c.  I decided to further provide a core assignment where the students have to read a paper on a new Instruction Set (neither ARM nor MIPS but other ISAs such as the IBM/360, VAX, DEC-Alpha, etc.) and come up with a one-page summary of the ISA, addressing modes and register conventions. I was excited about this project as some students are very motivated to research new material and associated papers online.

In doing a-c, I was hoping to establish three metrics to assign a final grade to the students, which includes, the normal exams/homeworks/quiz component, a significant project component with an ambitious attempt at hacking SimpleScalar and running SPEC or other test benchmarks and third, a research component, where they are to perform a literature survey and come up with a succinct description of a brand-new architecture, one that they have not been exposed to via lectures or via the project. This way, I thought I would get to evaluate their test-taking skills, their coding skills and their analytical skills.

Half-way through the semester, this plan ran into a very serious problem, the English language as the medium of communication. There were students whose performance on in-class questions was not in line with their performance on exams. The project was only to be evaluated at the end of the semester, due to which I had no feedback on whether the section of students whose performances were mismatched were struggling because of how the class was being taught or something else. This led me to switch gears mid-semester and create a small portion of credit for students to present a section of the lecture in the class. I chose Computer Arithmetic, one of the more detail-oriented sections of the lectures. My intuition was correct about the mismatched performances occurring due to English (some of our students go through English lessons in their first years, to be cognizant of the talent that comes in from so many different parts of the country). All the students had to do was present three slides and some of them did exceptionally well, even in poor English, by honest measure. If I had left the assessment upon such factors as exams, projects or even in-class answers, I would not have had enough to go on.

Formative Assessment has been defined as a bi-directional process between the student and teacher to create a responsive learning atmosphere. In formal formative assessment, facilitators assess or observe students and their responses to various questions and the general procedure of the class. In our case, I was the only facilitator including a Graduate Teaching Assistant who pitched in occasionally and played a major role in the assessment of projects. The terms formative and summative evaluation were first coined in 1967 by Michael Scriven. The Handbook of Formative and Summative Evaluation is considered one of the older works in this area co-authored by Scriven with Thomas Hasting and George Madaus. Embedding tests during instructions is a practise that was introduced and has found wide acceptance in K12 classrooms in the United States.

The goals of this assessment are stated as producing non-threatening results, providing direct and immediate feedback, utilizing structured information to gauge whether the lesson plan needs to be changed and ultimately provides means to improve the overall learning quality. While formative assessment is widely talked about in the context of Elementary Education, with my experience in July 2010, I think it is high time that this debate be extended to Colleges and higher education as well, in a seamless manner. In terms of assessing students when the medium of instruction is no longer their mother tongue or even what they have been exposed to up to high-school, the field is rife with opportunities to improve the classroom atmosphere.