Boosting Code Search with Structural Code Annotation
Round 1
Reviewer 1 Report
The authors propose a code annotation model to extract features from six perspectives, i.e., functionality, usage, structure, development, and exception. Then the authors conduct a code search engine, CodeHunter, based on the structural annotation model. The topic is interesting. The detailed comments are given below:
1. The structure of the whole paper is not so clear. The related work should be given in Section 2.
2. The advantages and disadvantages should be summarized and listed.
3. Why an abstract syntax tree should be used as the code feature extraction. It seems that many techniques can be used to obtain this aim.
4. From related works, we can see some existing works. More compared results should be included in this work.
5. Please check the references. Ref. 22 and Ref. 23 are same.
Author Response
Response to Review Comments
Dear editor and reviewer:
Thank you for taking time out of your busy schedule to review our manuscript entitled “Boosting Code Search with Structural Code Annotation”. We really appreciate your help on making constructive remarks and useful suggestions, which has significantly raised the quality of the manuscript and has enable us to improve it. We have uploaded the revision into the system, all the refinements are marked in blue font, and in the following we will respond to the reviewers' comments point by point.
Comments from Reviewer 1:
Point 1: The structure of the whole paper is not so clear. The related work should be given in Section 2.
Response 1: Thank you for this valuable comment. We have totally refined the structure of this manuscript. In the revision, the related work is presented in Section 2 (Page 2-3).
Point 2: The advantages and disadvantages should be summarized and listed.
Response 2: Thank you for your rigorous consideration. We have carefully summarized the advantages and disadvantages of the proposed technique in Section 1 (Page 1-2). The main disadvantage is that CodeHunter is limited by the quality of code feature and annotation features, the effectiveness may decrease without the well-conducted code base.
Point 3: Why an abstract syntax tree should be used as the code feature extraction. It seems that many techniques can be used to obtain this aim.
Response 3: Thank you for this valuable comment. An abstract syntax tree can comprehensively represent the features contained in the code snippets. We have added the reason to use AST in Section 3.2.1 (Page 5).
Point 4: From related works, we can see some existing works. More compared results should be included in this work.
Response 4: We totally understand the reviewer's concern. The selected techniques in our experiments, i.e., Lucene and DeepCS, are the representative code search methods in software industry and academia. We will extend our experiments on more subject projects and techniques in the future work.
Point 5: Please check the references. Ref. 22 and Ref. 23 are same.
Response 5: Thank you for pointing out this problem in manuscript. We have revised it and other typos in this version.
Once again, we thank you for the time you put in reviewing our manuscript and look forward to meeting your expectations. We hope that the revised manuscript is accepted for publication in Electronics.
Yours sincerely,
Xianglong Kong
Hongyu Chen
Ming Yu
Lixiang Zhang
Reviewer 2 Report
This paper contains well researched topic and is suitable for publications after addressing minor concerns.
The section “4. Related work” may be shifted just after “Introduction” sections.
The paper can be divided into scientific sequence such as… Introduction, Methodology, Results and discussion, and Conclusion. The present form looks difficult to follow.
Kindly, revise conclusion with something different (but linked) from results. Also, please suggest all the possible future directions.
Please add some latest research papers from 2021-2022 related to the code search methods.
Overall, the paper is good. However, the authors’ great work should be appreciated by research community. This is possible only with rearranging and modifying the paper according to reviewers’ suggestions.
Author Response
Response to Review Comments
Dear editor and reviewer:
Thank you for taking time out of your busy schedule to review our manuscript entitled “Boosting Code Search with Structural Code Annotation”. We really appreciate your help on making constructive remarks and useful suggestions, which has significantly raised the quality of the manuscript and has enable us to improve it. We have uploaded the revision into the system, all the refinements are marked in blue font, and in the following we will respond to the reviewers' comments point by point.
Comments from Reviewer 2:
Point 1: The section “4. Related work” may be shifted just after “Introduction” sections.
Response 1: We are grateful for the suggestion. In the revision, the related work is presented in Section 2 (Page 2-3).
Point 2: The paper can be divided into scientific sequence such as… Introduction, Methodology, Results and discussion, and Conclusion. The present form looks difficult to follow.
Response 2: Thank you for this valuable comment. We have totally refined the structure of this manuscript. The resubmitted version is easy to follow.
Point 3: Kindly, revise conclusion with something different (but linked) from results. Also, please suggest all the possible future directions.
Response 3: Thank you for this valuable comment. We have rewritten the Conclusion and listed some possible future directions (Page 14-15).
Point 4: Please add some latest research papers from 2021-2022 related to the code search methods.
Response 4: Thank you for your nice comment. We have removed some irrelevant or early references and added some works that focus on code search.
Once again, we thank you for the time you put in reviewing our manuscript and look forward to meeting your expectations. We hope that the revised manuscript is accepted for publication in Electronics.
Yours sincerely,
Xianglong Kong
Hongyu Chen
Ming Yu
Lixiang Zhang
Reviewer 3 Report
The work proposes the use of code search technology using deep neural networks based on both code and annotations as input to then produce the results.
The results are very relevant to the area of software engineering, especially with regard to the quality and cost of software development.
The text is clear and the work is well organized and presented. However, some aspects, if improved, can facilitate understanding for the reader.
- the contextualization, at the beginning of the abstract, can be clearer, in the same way that it is done at the beginning of the Introduction chapter (the first 5 lines of this chapter).
- the conclusions of the work do not reflect the effort spent and need to be increased
- the authors do not suggest future works.
Author Response
Response to Review Comments
Dear editor and reviewer:
Thank you for taking time out of your busy schedule to review our manuscript entitled “Boosting Code Search with Structural Code Annotation”. We really appreciate your help on making constructive remarks and useful suggestions, which has significantly raised the quality of the manuscript and has enable us to improve it. We have uploaded the revision into the system, all the refinements are marked in blue font, and in the following we will respond to the reviewers' comments point by point.
Comments from Reviewer 3:
Point 1: the contextualization, at the beginning of the abstract, can be clearer, in the same way that it is done at the beginning of the Introduction chapter (the first 5 lines of this chapter).
Response 1: Thank you for this valuable comment. We have rewritten the Abstract and Conclusion to make them clearer. And the beginning of the Introduction is also refined (Page 1).
Point 2: the conclusions of the work do not reflect the effort spent and need to be increased.
Response 2: Thank you for pointing out this problem in manuscript. We have rewritten the Conclusion to present the significant efforts in our work (Page 14-15).
Point 3: the authors do not suggest future works.
Response 3: Thank you for this valuable comment. We have totally rewritten the Conclusion and listed some possible future directions (Page 15).
Once again, we thank you for the time you put in reviewing our manuscript and look forward to meeting your expectations. We hope that the revised manuscript is accepted for publication in Electronics.
Yours sincerely,
Xianglong Kong
Hongyu Chen
Ming Yu
Lixiang Zhang
Round 2
Reviewer 1 Report
The authors have carefully revised. I have no further comments.
