University Postmortem
After closing Blazing Games due mostly to too much competition, I returned to university in an attempt to get my bachelor’s degree to have the minimal degree that most companies hiring programmers were looking for and to make sure that my skills were up to date. While pursuing the degree, I was near the top of my class and thought that perhaps it may be worthwhile to get into the research side of computer science so opted to go for my MSc in Computer Science.
What Went Right
Just completing a master’s degree is an achievement. The number of students allowed into a program, at least at the university I attended, is only a small percentage of those who apply. The dropout rate is surprisingly high as well, especially within the sciences. While the graduate level courses are more work than third/fourth year courses, the difficulty jump is not as high as it was from first/second year courses to third/fourth year courses. Granted, the first/second year courses may appear easier to me as I was already familiar with the materials but tended to have to take the courses anyway as self-taught doesn’t count. While it is possible to challenge some courses, it is still a fair bit of work and you don’t save that much by doing it so may as well take the courses to fill in one or two missing spots in your knowledge.
The other major thing that went right was getting an answer to “is research an appropriate field for me?” While I was okay at research, looking at some of my alumni, it is clear that there are people much more suitable to academia than I am. This means that a future pursuit of a PhD is not likely for me. What I have learned about performing research will be useful for my further endeavors and to my surprise has improved my coding skills. The necessarily precise nature of research can be applied to software development, especially for testing software.
Even though there are no plans to pursue a PhD, I found myself really enjoying my work as a teaching assistant. Working as a teacher at a college or equivalent may be something in my future. Failing that, writing books to teach others may be an option. Videos are something I would like to experiment in, but with how over-crowded YouTube is, gaining an audience large enough to cover the effort is unlikely. Still, a small (few dozen) audience is all I would need for a hobby so will probably give it a shot sometime.
What went wrong
The thesis aspect of my MSc took longer than expected due to a pandemic and the simple fact that I switched what I was working on part way through my research. My supervisor did warn me that the change in research direction would add time but was kind enough to let me go my own direction with research. The area I ended up researching involved multiple areas which did not help matters, especially when I was a newbie at machine learning so was still learning the basics while trying to expand upon existing research.
Covid-19 probably impacted a lot of research. Normally, the effect on computer science would be minimal, but my research required a user study which requires people so having the campus shut down for a lengthy time was not good. This delay was longer than it should have been as my expectation was that classes would resume in the fall. When it became clear that this was not going to happen, the study was redesigned to be online.
As already mentioned, my knowledge of the field of AI and machine learning was sparse. I had thought that graduate work was going to be like other courses where students learn the material. Graduate courses do work that way, but the thesis is really about taking a field further (even if only by a baby-step) so existing knowledge about a field is important. Picking a research area because it is something you want to learn is not a good idea. Learning a field while researching is not a good approach. My recommendation to future students is to make sure you are familiar with the field you are getting into before picking your area of research.
Mixed blessings
I am not as good at math as I should be. When working with 3D graphics, I can muddle through the concepts to get things working but don’t fully understand the math. Having to work more with the underlying math behind procedural content generation via machine learning, it became clear that the problem wasn’t entirely a lack of understanding. I use some of the more advanced concepts behind math, the problem is I do this outside of the official math route so don’t know the symbolism or the tricks and techniques to best utilize the concepts. Improving math skills is definitely in my future, it is just a matter of finding time to do so.
When I entered my master’s program, I had thought that my organizational skills were good. There is software available for helping organize and write a thesis. The need for such software, as my organizing methods aren’t bad, was not evident to me. While manually organizing things works, there is a lot of overhead searching for things. Having software that handles and organizes all the research details would have sped up writing the thesis and would have also made me make more notes instead of relying on my memory. There are simply too many things to remember so no matter how good your memory is, you are going to forget things. What is worse, you will remember a detail you want to reference but not remember what article it was in so will have to spend too much time skimming through older articles you have read.
Final Thoughts
While I am not going to pursue my PhD, teaching in the future is not ruled out. While a MSc is not required for being a developer, it doesn’t hurt and the skills I learned while attending graduate school are useful. My future right now is in a bit of a hiatus as I am taking care of my father, and will be doing so until he is at a stage where he can take care of himself with minimal aid. While that is happening, my spare time is going to be on the development of my own game engine. Why? Partially for learning but mostly because of my NIH complex. More on that in the future!
