How to differentiate between the concept and the implementation in a computer science thesis?

Question

A common structure of a thesis in computer science has the following structure:

1. Introduction
2. Related Work
3. (Requirement) Analysis
4. Concept
5. Implementation
6. Evaluation
7. Conclusion

Considering such a thesis, I would like to ask how can I differentiate which aspects are part of the concept and which is considered to be a part of the implementation?

My advisor's advice was: specific things and technical details are written in the implementation section, more general ones are in the concept. There are no strict rules but he can see the difference but not explain.

The problem is that I do not have a very generic concept, since my research question is very strict and the solution is not a generic one, it is for this specific purpose and doing it more generic, it makes no sense since it does not work or has any purpose.

So my question is: are there any other hints or question how I am able to differentiate what are parts of a concept and what the implementation is?

Thanks in advance!

Answer 1

Suppose that, five years from now, someone produces some "new" (open-source) software to solve the same problem that you're solving in your thesis. Of course, he might have used a totally new approach, but suppose that, after you study the new software, you see that it just uses your ideas, changing perhaps the programming language, the specific data structures used, and similar details. So you decide (perhaps indignantly) that this isn't really new at all.

I'd say that those ideas of yours that were copied are your concepts. The aspects that were changed without making an "essential" difference are the implementation.

(Of course, if you had invented an entirely new data structure, then that would become a concept rather than part of the implementation. But that seems to be a rare situation.)

Answer 2

In many other fields, what you have called "concept" is, roughly speaking, called "methodology. You have a problem. How to you propose to attack that problem.

Likely you have some insight (concept) into the inner workings of things that might be tested (verified, debunked) with an implementation, usually code, less frequently a logical analysis.

Back in the day people were trying to create effective and efficient garbage collectors for dynamic memory. There were many approaches. See: https://en.wikipedia.org/wiki/Garbage_collection_(computer_science).

But each them has a high level description (concept) and an implementation (or several) to "prove" the concept. At a certain point in time a small percentage increase in efficiency was good for a dissertation. But without that implementation, it was just a dream.