From d9e50185d9ad792148f98a0b5b3c9eeca36a8315 Mon Sep 17 00:00:00 2001 From: Alan Date: Fri, 22 Dec 2023 22:09:27 +0000 Subject: [PATCH] Initial template of an article discussing my work teaching computer organization. This includes a write-up of the overall article. Still need to add concluding notes and do some proof-reading. --- projects/organization/index.php | 218 ++++++++++++++++++++++++++++++++ 1 file changed, 218 insertions(+) create mode 100644 projects/organization/index.php diff --git a/projects/organization/index.php b/projects/organization/index.php new file mode 100644 index 0000000..e401c09 --- /dev/null +++ b/projects/organization/index.php @@ -0,0 +1,218 @@ + +
+
+
+
+
+

Preface

+
+

Pedagogy: College Education and Computer Science

+

+ Upon the conclusion of my first semester of grad-school, I was asked to take on a sole-responsibility role to teach Computer Organization during the summer semester. I took pride that they offered the position to someone who had only been in grad-school for 5 months; I must have impressed as a TA. I've had prior life experience, being a non-traditional student, which contributed to my confidence as both an individual and an instructor. My diverse background experience also lent well to being able to communicate what was being taught in a multitude of different approaches. +

+ +

+ My brief stint as a PhD student has shown me that the fact the university is an R1 school likely also played a role into my selection - no tenured faculty member wanted to take on the task and a fresh grad student is low in the rung of production for what the school ultimately values - research. +

+ +

+ Regardless of motive of selection, things aligned with my motives as a student. I did not pursue any of my degrees, (Master's included), for the sake of getting my name on a paper. I attended these educational institutions to learn and to satisfy my curiosity of the discipline. This curiosity extended to wanting to learn how to effectively teach the subject. +

+ +

+ Computer Science is a hard discipline to learn. I pose that most have difficulty teaching the discipline. Frustratingly, most professors have difficulty teaching the discipline - Experts of the field! Why is this? I pose that most professors are post-doc researchers who are obligated to teach. As alluded above, their motives don't often align with superb instruction. To teach a subject is an annoyance that comes with the job. Anyone who has gone to college, regardless of discipline, will likely agree with this assertion whilst mentally (re)evaluating majority of their professors. Doubly so for those that are tenured. +

+ +

+ The question needs to be asked of who is at a disadvantage through this environment? What type of student is at risk here? These questions are too general for the focus of this writing, but I invite the reader to linger and muse on them. I will make the following assertions based on my experiences which I believe should also be considered: +

    +
  • + In the domain of computer science, there are three different types of students: +
      +
    • + There are those who are studying a different discipline who are required to take a CS course as a prerequisite. +
      • +
    • + There are those who have heard jobs related to the field pay well, and thus are studying on the prospect of the paycheck. +
      • +
    • + And then there are the individuals who are genuinely curious of the subject. +
      • +
    +
    • +
  • + An individual can initially be one of the former two and eventually be molded to fit the later of the three. +
    • +
  • + Students who are genuinely curious of the subject will succeed. The definition of success is that they will get a degree and they will have a solid and flexible intuition of the mechinations of discipline. +
      +
    • + The students in the other categories will get just a degree. +
      • +
    +
    • +
  • + Students applying for a a job related to the field of computer science will have a technical quiz during an interview. This is rare; I know of no other disciplines that exhibit this. A doctor who is applying for work at a practice does not have to recite information and procedure they learned as a pre-med student. A computer science student needs to memorize what they've learned in data structures, regardless of the fact that as a software developer you have access to a wide range of resources to help solve problems. +
      +
    • + The fact this is how things exist must be due to the fact that students of the above mentioned second category exist and need to be filtered. +
      • +
    +
    • +
  • + No instructor, myself included, can wholly prepare an individual for these technical quizzes. It is up to the individual to take a vested interest in the subject - this involves taking it upon themselves to truly comprehend a lot of the core concepts that are a part of this discipline. +
    • +
+

+ +

+ The fact that technical quizzes exist in the interview stage implies that a lifetime of irreversible damage has been done within the instructional system. Some will say that instructors aren't doing a good enough job of filtering out those who don't do well enough. I will instead assert that instructors aren't doing a good enough job teaching the material to a wider range of individuals. +

+ +
+ +

+ Loops are easy. They're a concept taught early in the stage of programming. They are also super frustrating as a beginner; an esoteric construct that goes against the grain of normal human experience with respect to the level of granularity required to convey that some thing needs to be repeated. I almost dropped out of community college because of loops. My instructor simply told me the surface facts of a loop and did not encourage any form of reflection. I was given a loop of a specific problem, the pieces that describe the condition of execution were pointed out to me, and it was noted that the body acted as the logic that needed to be repeated. There was no exploration of how to determine what belongs in the body and what doesn't; or how to fine-tune the condition of execution. +

+ +

+ Due to the fact a loop is easy, it becomes intuitive for someone who has enough experience to be able to teach. It's difficult to teach something that is intuitive to someone who does not have that intuit. My approach has always been through analog. This begins with discussion of the various iterative happenings that may occur through life and then discussion of how to encode them. After this discussion, examples are given. +

+ +

+ Once examples are given, I've always been keen to sharing alternatives to a given implementation of a loop. A for-loop absolutely should be taught in tandem with a while loop along with a sprinkling of recursion. Various conditions of execution should be here as well. Analog is not only given through context of every day living but also with respect to evaluating the loops themselves! +

+ +

+ The usage of analogous example shows that the same problem can have many solutions. This allows a student to view solutions through different perspectives. This abstracts the study to what it is - how to abstract the algorithm. It also encourages a student to think outside of the box and to be creative in their own solutions. This helps foster one of the most crucial traits of being successful at this study - being curious with critical thought.* +

+ +

+ A student who isn't genuinely interested in the study is less likely to apply critical thought. They may have been boxed into a narrow view of the study to which no interest is curated. This type of student is the type most likely to find solutions on stack overflow and not learn from them. This student is the type most likely to learn what little they need to from generative tools like Chat GPT. +

+ +

+ At the end of day, regardless of which category a student belongs, they are paying good money for the education. They are taking a class to learn the subject involved. The goal of an educator should be to engage with the student's expectations and maximize the value of the experience. "Value of the experience" is subjective to each individual, but given the context of school being a place to learn, my efforts are focused on providing the value of knowing the subject. In the example above, I take extra effort in providing the value of knowing what a loop is and does by inviting a student to be a better critical thinker.** +

+ +

+ What is it that a paying student seeks to learn by taking Computer Organization? Any student who is paying to be in the class is likely a Computer Science major. Many would describe it as a weed-out course that teaches a programming paradigm that is antiquated. They may not be able to justify why it has this perception, but those who have a genuine interest will tell you that it's due to a combination of Moore's law and compiler optimization techniques. +

+ +

+ A post-doc researcher teaching the course will rattle off a set of objectives set in place by the institution, likely falling upon the course description provided by the university. They may say: +

+ By the end of the course, you will be able to explain how a computer works, from “bottom to top”: transistors to digital logic to assembly language (e.g., x86, MIPS) to the programming language (e.g., Javascript, Java, Python). To do so, you will uncover how data is represented and stored in digital computers. You will learn how to write programs in an assembly language made up of basic instructions (e.g., multiply two numbers, write a number to memory) that a computer can understand. You will learn how to construct and use the basic blocks of a computer processor: arithmetic, memories, I/O, and gates built using digital logic (circuits whose wires hold 1’s and 0’s). You will use design tools to build a CPU that runs assembly programs. Finally, you will analyze computer systems with respect to engineering metrics like cost and performance and discuss factors, such as data locality and parallelism, that affect these metrics. - UIowa.edu +
+ This course description was indeed the ultimate appeal for continuing my education from community college to university. Having learned how to program at a higher-level, I was interested in knowing what occurs close to the circuit; I wanted to learn the "1's and 0's," as it were. +

+ +

+ I do not believe the above course description does the class justice. This course brings together and synthesizes Data and Discrete Structures. Data Structures formally introduces the concept of data representation by exploring alternative implementations of algorithms that are commonly abstracted by higher-level programming languages. Discrete Structures provides the mathematical background to describe the actual representation of data as it exists within the logical circuitry. This synthesis is more than the sum of its parts. The propositional logic used to describe the circuitry is represented as a certain set of symbols which is then translated to the symbolic diagrams which compose the logical circuitry. The assembly languages translated to this circuitry alludes to one that there may exist any amount of layers to translate a problem represented in one language to another. This course implicitly informs any curious student that the algorithm is programming-language agnostic. Wholly understanding this course trivializes studying Theory of Computation; this course teaches a student a model of computation! +

+ +

+ The context provided by this preface is exempt from the post-course report I've written for the university. The context provided by this preface hopefully conveys to the reader that I tried to maximize the value of of learning the subject to my students. The extra value was explicitly relayed to my students with an emphasis on the fact that what we are studying and building is a model of computation. This gives extra context to the student by pushing the course beyond the study of some antiquated paradigm. +

+ +

+ The following is the report I submitted to the university. +

+
+
+
+

CS:2630 - Context and Methodology

+
+

+ People tend to have the perception of Computer Organization being a difficult class involving a lot of grit. This perception typically dominates discussion about the topic. Emphasis on key concepts can help alleviate these points. Specifically, directly correlating concepts of this course to concepts of previous courses. +

+ +

+ Personally, it was Computer Organization that cemented the idea that learning a new programming language need not be daunting. The data and functions described throughout this course ultimately represent the same functions and data described in the higher-level languages of courses prior. Knowing this, the primary focus while teaching this course was to emphasize this fact. The means of describing the data in machine language in addition to the propositional logic used to manipulate that data serve as the model of computation in which the higher-level languages are built. This theme was highlighted quite often. Hopefully this made the course less gritty as concepts were built up from propositional logic to logical gates, logical circuits, control units, and assembly. +

+ +

+ Recognizing this scaffolding does require familiarity with some higher-level language. To fully appreciate it, it also requires being comfortable* with some higher-level language. This is based on my experience as an undergraduate student. Stepping into university, my context was being comfortable with a handful of high-level scripting languages for web development. One key observation during the onset of my university experience was noticing that, within various courses that followed the introductory programming course, a lot of students were still gripping with the core concept of learning how to program. This tended to prevent them from understanding and appreciating the primary concepts being taught in said courses. This is especially true in data structures. +

+ +

+ The university of Iowa has an advantage over the university I attended in the fact that Data Structures is a prerequisite to Computer Organization, as opposed to a corequisite. The point made in the previous paragraph still holds true regardless of this fact. The advantage becomes apparent when looking over the lab assignments given during this session. +

+ +

+ As indicated in the topic outline above, the first couple of weeks were spent discussing the representation of data. Something that stuck out to me in the (optional) textbook associated with this session is the following aside: +

+ For converting larger values between decimal and hexadecimal, it is best to let a computer or calculator do the work. There are numerous tools that can do this. One simple way is to use any of the standard search engines, with queries such as
+ Convert 0xabcd to decimal
+ Or
+ 123 in hex
+
+

+ +

+ It is amazing that this was suggested in such a textbook, especially one that is of “a programmer’s perspective.” Knowing this cheeky blurb existed, the first lab was one to practice and test the programming skills of the student. They were to implement the functions which perform the conversions indicated by the aside in some high-level language of their choice. It was made apparent to those given the assignment that not only were they being gauged on the correctness of the functions but their coding styles were also being gleaned. Additionally, it was stated that should they find and use a solution on stack-overflow or an alternative that taking such a measure likely was indicative of a lack of problem-solving skills. It was made apparent that not taking the time to shore up those skills, (through this lab), that this lack of skill will make the course more difficult as time progresses and as more difficult problems are presented; problems that will not be so easy to find using said sources. +

+ +

+ Looking at the subsequent labs, one may surmise that the reason it may be difficult finding solutions would be on account of using both a relatively obscure assembly language, MARIE, and an oversaturation of programs solved in higher-level languages, as per the lab assignments pertaining to MIPS. +

+ +

+ While the MARIE programs acted as a means for one to become comfortable coding assembly, (while also being able to track the control and flow of data within the architecture), the MIPS programs acted as the means to cement the idea that the data and functions discussed throughout this course are the very same as those described in other languages. As the first lab asked to represent the algorithm which was applied by pen and paper in earlier lecture assignments and as it was through lecture that various representations of both data and functions were explored, the MIPS labs asked to represent various algorithms that should have been learned through data structures. Each facet drawing heavily from prior courses, something that was emphasized during instruction. +

+ +

+ The submission policy outlined in the syllabus is an indication that should a student not be solid in any of these fundamental areas they will have the opportunity to make it up. What is not shown in the syllabus is the test policy. The tests given were take-home tests where a few days were given to complete them. They were essentially glorified labs with less focus on a specific topic. The tests were designed in a manner where, should one be up to speed with the course, it should take one and a half to two hours to complete. Should one not be up to speed, then they will need to invest more time to complete the test; a last chance to put in the time and learn the topics covered. +

+ +
+ +

+ The following pages contain the syllabus, lecture assignments, and lab assignments in the order in which they were given, followed by the two tests. This was the first class I have taught. My initial intentions were to give smaller lecture assignments following each lecture. My intentions were to also give a lab assignment every week. It did not turn out this way. To anyone looking this over, think of more lecture assignments to give. If I am to teach this course again, I will be doing that. As for the lab assignments, time was better spent giving lecture, as opposed to dedicating work time for the labs. This course was taught during the summer session, and it is my opinion that prolonged lectures were too dense. I instead opted to give shorter lectures and shoring them up with a supplemental 20–30-minute follow-up video which retouched lecture concepts and provided more examples or contexts. Labs were kicked off during the last class day of the week where the students had the entire weekend to complete. +

+ +

+ If there are any lingering questions, feel free to contact me. +

+ +
+ +

+ * What it means to be comfortable with a programming language deserves its own essay. I will leave it to your own judgement to decide what this means, or I will leave an open invitation to personal conversation on the matter. +

+ +
+
+

Concluding notes

+

+ +

+ +
+
+
+ +
+
+ +