Defining and Researching “Dynamic Systems of Systems”

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors  

This paper introduces a deep survey of classification and characterstics for dynamic systems of systems. The result comes from a sufficient interview and literature review. The authors present the six characterstics of dynaSoS and compare them with traditional SoS. The characterstics contains  heterogeneous open systems, continuously improved & innovation-driven, complex systems, distributed systems, data-intensive systems and AI-based autonomy. The research findings can guide the current digital transformation. The idea of this paper is meaningful and the paper is written in a good manner. However, I still have some suggestions to make it better.

1) The current contribution C2 is a little inadequate. I have not found innovative application of the classification and characterstics in this work. And the scope and characterstics can support two contributions independently.

2) It is not clear that why the authors name these techniques as "dynamic" systems of systems. Through the analysis, I find that the main improvements come from the data-intensive and AI-based autonomy. They have enabled our community in many respects, why do the authors emphasize "dynamic" feature?

3) It will be better that if the authors can present the trends of research questions related with characterstics in  the literature review.

Author Response

Dear reviewer,

 

thank you very much for your feedback. Please find our answer to both reviewers attached.

 

Best regards,

Frank Elberzhager

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The topic of the paper is relevant and interesting. Many interesting issues are brought together. New insights are generated. Bu the paper has some deficits.

The scientific approach  is adequate, with lots of efforts taken to collect and evaluate data, which is great!! Nevertheless, its soundness is questionable as it is a discussion of an abstract topic with experts. Such discussions usually deliver random results. Therefore, it is important to explain how you handled that problem! In addition, you should explain how you processed the collected data. Which type of coding did you apply? Etc.

The main problem of the paper is, that things are mixed to a degree that makes it difficult to learn from your findings. In some parts of your paper, it is not clear what is the result of your research and what is already there in the literature. In other parts without references to literature that is rather clear, but the different views are not separated or there discussion is very imbalanced among different cases. Your central figure No 3 is hard to decode and reading the text does not make it much easier: Do arrows indicate one-to-one relationships or cross tier relationships? In general, you frameworks needs more clear explanations. It is well-known that drawings in IT are least precise looked at from the point of view of visual design, but in your case the vagueness limits the usefulness of your publication.

You introduce the concept dynaSoS without clear definition and without references. You only vaguely refer to examples.  You present abstract arguments about its properties. concrete examples are given only later. You should rearrange your text in parts and try to come up with definitions. I assume that you started your research without a clear definition of your topic. That is not to be criticized. It may even be worthwhile to describe this aspect in the paper explicitly. However, in the current version you push the reader through your own experiences, this is not how papers should be written. 

Furthermore, there are some mistakes with respect to terminology. For example, you call geographically distributed system "Distributed Systems". While this only non-standard terminology and as such acceptable, you later talk in this context about "distributed systems engineering", which has nothing to do geographic distribution, but relates to the specific challenges of distributed systems like no total order of events, consensus problems etc.

I understand that in a multidiciplinary context it is pretty challenging to avoid contradictions like the above, nevertheless the one mentioned is very much at the heart of engineering challenges. In practice, the basics of distributed systems design are usually forgotten, but the classical issues persist wit off the shelf solutions - people just don't know that there is a theory around. Indeed, even in political discourse on the good and evil distributed systems theory and IT zeitgeist clash sometimes. 

Summing up, you obviously have a lot of interesting findings to publish, but you should structure them more clearly. Try to tell a story for the readers of the journal that creates curiosity from the beginning. Try to separate aspects clearly, even if this reveals white spots. If you find it difficult to create a text that helps to take disciplinary views apart, then you could add more diagrams.

 

 

 

Author Response

Dear reviewer,

 

thank you very much for your feedback. Please find our answer to both reviewers attached.

 

Best regards,

Frank Elberzhager

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

All positive remarks in the original value remain valid. Concerning the critical remarks, significant progress was made but some concerns are still open.

Although revisions were rather minor, the quality of the paper has been significantly improved. There is more clarity at critical parts of the text and there the visual models are much better now. 

There remains the one main issue that research methods are not discussed in depth. This was not usual in the past, but it is becoming a standard in business informatics literature.  Of course, this change is debatable. One a very fundamental level: While there are important arguments in favor of more transparency, there are also arguments against it:  Sometimes research is done very scientifically and the result nevertheless turns out to be crap, because the researchers have no deep understanding of the domain. And on a pragmatic level: the focus on methods has changed the contents of the output. 

Thus, it is really a question of community standards. I leave it to the editors to decide on acceptance or continued revision.