What is the Computer Science Capstone like?

[dcan]

Hello all, my long-delayed Comp Sci degree journey is finally nearing an end. I now have two classes remaining, analysis of algorithms and the final capstone.

The problem is that I have essentially *no idea* what to cover in a capstone for a CS degree, so I'm hoping someone here can give some advice, especially since I see a few people have gone through the program since I began it several years back. I've looked through the relevant threads I could find here regarding the capstone but none that I found specifically addressed the CS program.

Specific questions I have:

1. Does the capstone strictly require a thesis, or is there a project or project + thesis hybrid available as well?
   
2. If thesis is the only option, must it be a pure research paper or can it be a literature review that summarizes a topic?
   
3. What kinds of problems are studied in an undergraduate CS thesis at TESU?
   
4. How would a CS capstone paper ask, then address, a research question in a thesis?
   

Writing is not much of a problem for me. Before switching to CS I completed the entire BSBA CIS program including the strategic management capstone which included a 20 page paper, but it was not a traditional thesis, more of a business study. I understand and expect a thesis from this class to be potentially twice that long or more, and I'm fine with the actual act of writing that. What I am having trouble with is understanding what is expected (in terms of topics, approach, etc) in a CS thesis at this level. Any insights on that would be appreciated. Thanks!

[davewill]

CS students take the exact same liberal arts capstone all other BA degrees take. You can choose anything as a subject, it does not have to be CS related. My work is in automotive infotainment, so my paper was on the safety impact of these systems and design approaches to improve safety.

I'm not sure that a programming project would be accepted, but it might well be. Remember you still have to write a significant paper even if you choose a project for your capstone.

You could dive into social impact of computers or social media. How about generational attitudes towards privacy? Is television dying?

What broke the logjam for me was attempting to frame a main research question and a few sub questions for a few ideas. I originally wanted to research brain-computer interfaces, but the questions just didn't flow for me. The ones for auto safety were easy to come up with.

[dcan]

Thanks for the info. I knew it was the same course everyone in the BA program takes but wasn't clear on the specifics.

So if I understand correctly, you have to define a research question that is broad enough to be interesting yet narrow enough to be focused. That's fine. The issue then is coming up with a research question.

Were you asking questions and trying to solve them, or just summarize current research and industry trends on that topic? Looking at other undergrad CS thesis topics online the majority seem to be about developing new algorithms which seems out of scope at TESU. Can you give a brief rundown of how you approached your paper, how long it was, how broad/narrow the questions were, etc?

Also how is the OMSCS going? What's it like? I am interested in that and may start it in 2018. Thanks!

[davewill]

The main thing on the research is that the conclusions have to be your own. In my case, I looked at crash stats which showed that handheld devices are still causing increased numbers of accidents. This established a need to do better. I looked at research on different input modalities (touch, voice, buttons, gaze, gesture) and the research that showed the extent of cognitive and attention impairment. I also looked at the NHTSA guiding for IVI design. Based on that I made specific conclusions about features and design choices that should (or not) be given attention.

The length came in right at 25 pages, and I got a high A.

Here are the questions I ended up with. I originally had a couple more that I cut as my research progressed.

Problem Statement: How do newer and more sophisticated infotainment systems affect driving safety?
Subquestions:
What is the size and nature of the problem?
Are all drivers equally distracted by these features?
How can infotainment developers mitigate the problem by changing interfaces or using technological solutions?

Unfortunately, I simply did not have the time and energy to devote to OMSCS. Since it is not a real career enhancer for me. I had to let it go. I suppose I better take it out of the sig soon.

[dcan]

Ah ok so it was more about conducting a large literature review and drawing conclusions and recommendations from the literature. It sounds more like writing a chapter from a textbook that explains how something works or how it should be applied to solve various problems. That is certainly much more doable. I was concerned that we had to do "original research" i.e. conduct an experiment of some kind, and had no idea how to approach that in this context. Thanks for the insight.

[davewill]

... I was concerned that we had to do "original research" i.e. conduct an experiment of some kind, and had no idea how to approach that in this context. Thanks for the insight.

A number of my classmates took that approach as did the sample paper that was provided. All of those used either surveys or interviews or both to generate original research. My feelings were similar to yours that my subject didn't lend itself to doing original research as it would require more time and resources than I could bring to bear. I instead tried to stick to sources based on primary research so that I was relying less on other people's conclusions.

[TrailRunr]

Since I thought this class was a complete waste of time for my CS degree, I decided to avoid any interviews or surveys and wrote it as a literature review with my own thesis and conclusion. I picked something with lots of sources so I would not have a hard time. If I had aptitude in a completely different area like history that would help me shave time off this paper, I would have written a history paper. I saw some excellent Powerpoint presentations and interviews from classmates who went that route. They had a lot more time and interest than I did.

[dcan]

Sounds good, thanks to both of you for your inputs. Seems like the difficulty then is coming up with a good topic that is broad enough to be interesting but not too broad. That matches what others said about the course as well.

Previously I had blown the course off thinking I could get away with writing something bogus like "history of computer science" or something equally ridiculous. I even have a couple of older math history books that might have helped. Now it looks like I may need to re-evaluate that and consider what, exactly, the thesis question should be.

TrailRunr -- I wanted to ask you about your math courses. How was the TESU modeling course? It looked interesting but I haven't taken Calc 3 or diffy q's. I know the course lists Calc 2 as the requirement but since it involves DEs it seems it might be helpful to learn those first. On that, which two edX courses did you take and were those worthwhile beyond checking the box for the degree? I am trialing MATLAB and considering a license for it especially since I can use it in the algos class, and would like to explore using it beyond that. I might take the edX courses or get the TESU modeling text and go through it on my own later. Thanks!