3

I have developed a software to meet a specific requirement in a research field. There is only one such software developed so far, mine is second.

In the thesis, I explain the implementation of the software. However, I feel like I also need a "Results" chapter. Most of the master theseses I have gone through from various disciplines have such a chapter. Looks like it is a must.

My software generates tangible models, storing a specific information, to be manufactured with various methods, such as CNC or 3D printing.

Output of the software should be considered as results. However, I don't want to conduct a cheesy case study with a bunch of people and publish it as a scientific result. An extensive, thorough study is takes more than 6 months and neither I have that much time nor the the knowledge/experience to conduct such a case study.

For example a friend of my implemented a software for twitter clustering and analyzing. The results are quite straightforward with his software. Run the tool, get the results and interpret them. Done.

I am not sure how to evaluate the results. Models are manufactured successfully, whereas that does not look like a result to me.

How would you approach this question for a possible solution? What my alternatives can be?

  • 1
    In computer science, it is indeed typical to have an evaluation/validation section in a master's thesis, but there is no need for this section to be named "Results". You can choose a name that fits best. – lighthouse keeper Jan 12 '17 at 11:30
  • 3
    The "cheesy case study with a bunch of people", as you call it, is often a valuable part of the thesis in that it shows that you are not only capable of coding, but also of designing the material and setting for such a study. The point is not to generate a full-fledged "scientific result", but to show that you could generate such a result if given sufficient funding and time. Furthermore, while a thorough, statistical analysis is beyond what you can do in a Master thesis, do not underestimate the impact of listing at least some "first impressions" uttered by users other than yourself. – O. R. Mapper Jan 12 '17 at 11:33
4

The goal of your thesis was to create a tool that meets certain requirements. Consequently, your evaluation should show that your tool satisfies these requirements.

One way to organize your thesis is to have a section called "Requirements", where you walk through all requirements detailedly. For example, you could have following requirement with sub-requirements:

  • R1: Generate 3D cat models: the tool can generate 3D models from 2D cat images
    • R1.1: Brown cats: the model generation works for brown cats
    • R1.2: Black cats: the model generation works for black cats
    • R1.3: Checkered cats: the model generation works for checkered cats

Then you have another section called "Validation" where you walk through the requirements a second time, this time discussing for each requirement how you satisfied it.

  • R1: The cat model generation was implemented in the catgen component of the tool that uses the XYZ algorithm to build the cat models.
    • R1.1: for 4 sample cats, the tool was able to produce a model (see Fig. 1)
    • R1.2: for 5 sample cats, the tool was able to produce a model (see Fig. 2)
    • R1.3: for 3 sample cats, the tool was able to produce a model (see Fig. 3)
  • 1
    Thank you so much. That sounds reasonable and clarifies the questions – user1449456 Jan 12 '17 at 11:24
3

If someone asked you: "does your software work"? How would you prove to them that it does? How did you test the software and proved to yourself that the software works as intended? That could be the basis of a results section.

  • So, manufactured tangibles through software can be considered as results? – user1449456 Jan 12 '17 at 11:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.