How to develop experiment to justify construction of a prototype system?

Question

I have worked hard for my master thesis project to build a system that uses an already existing algorithm, just in a mobile setting, which hasn't been done before. Because the algorithm is complex, it required a client-server infrastructure. Now, as I said most, of my efforts until now went into building the system of already existing components rather than contributing to knowledge.

Now my question is, how can I use my system and efforts to go more in the direction of a scientific contribution? The problem is if I, for example, simulate data loss over the network, I could have done that without even building the system. So I somehow want to include my real system in the experiment rather than doing a simulation. What would be a good approach that justifies also the construction of my real system?

Answer 1

You might think of it in terms of the research question(s), which might look something like this:

1. Is it possible/feasible to use algorithm X in a mobile setting? (Presumably the answer is yes, because you've already built it.)

2. What challenges are there in applying algorithm X in a mobile setting? How can those challenges be met?

3. How does the efficiency/accuracy/data loss compare between the "regular" implementation of algorithm X and the implementation in a mobile setting?

---

Edited to add:

4. What are the advantages & disadvantages of using your implementation over a simulation? I'm guessing that your implementation gives more realistic results than a simulation would.

Obviously I don't know your field, or how novel a contribution your implementation is. But in principle, I think that answering the kinds of questions I've listed above could lead to publishable papers, and could justify your having built this system.

Answer 2

What would be a good approach that justifies also the construction of my real system?

I don't have a crystal ball, so I can't say anything other than: in some fields, in addition to showing that your approach works well in a simulated environment, it is a really good idea to also show that your approach works well in practice.

If your peers value such "real world" results, then you don't need to justify it any further than that.