I think you are thinking more of a tutorial than a research object. You may be interested on this.
Research is about obtaining some new knowledge. This knowledge can be simply a compilation of some things that are already done and existing but were never put together to compare them (a survey) for instance. IMHO survey papers get a good number of citations because they are in fact quite useful.
There are a number of things to do research in that field, for instance you could
- put something new in that small debugger (beyond the state of the art) or
- try a set of different things in that small debugger to compare them and obtain that comparative knowledge (being new) or
- do a survey of the debuggers that exist, or evaluate them wrt something new (e.g. usability)
- find open questions in the field of debuggers, define a problem that was not being considered before
Basically it boils down to
- finding and defining a problem (which is open in the state of the art (SOA))
- finding or creating solutions (and adding them to the SOA)
- evaluating and comparing solutions (in the SOA to know better how is it)
I'd say all of them are quite pragmatical, focusing on solutions. The point is to
- solve something new (it's quite pointless solving something that is already solved, right?).
- evaluate properly (so that we can understand why is it a solution and what properties does it have, when is it useful and when may not be so useful).
- finding problems (because the first step to solve a problem is identifying it)
Depending on the approach taken then the research and the paper will look in different ways but basically you need:
- an introduction
- an explanation of the background that may be required to understand everything else (this is optional)
- a description of the state of the art, what is out there that is relevant for the proposal
- the proposal (for definition, solution or evaluation of the problem at hand)
- the results obtained with that proposal
- the conclusions obtained from the results
PS: If that fails you can simply forget about the state of the art and do research on this as if it was new, without checking properly the literature on the topic before. If you apply the methods correctly and do a good work that could be ok in some contexts (I'm not sure about the context of this question, though). This would not be very practical, because most probably someone did that already decades ago.