Provide more context to the work, i.e. why we did this

[moorepants]

Some of the reviewers felt that an explanation of why this data was created is needed beyond "we created the data because we thought it would be generally useful to the world."

The editor says:

Please give a context and objective to the work, as noted by Reviewer 1 and 2. This could stem from the funder driven requirement to share data openly (as suggested by R.3), but I am sure the data was generated and analysed for one or more specific research questions or hypotheses.

Reviewer #1 (Sanguex):

However, I do not think the proposed manuscript represent a 'unit of publication'. I suspect these data were collected for a purpose but this purpose is absent from the manuscript.

As mentioned in basic reporting, there is no research question and therefore I do not think it can join the scholarly literature as is. I would strongly encourage the authors to re-submit their work/data as part of a paper WITH a research question.

Reviewer #2 (Lee):

"The submission should be ‘self-contained,’ should represent an appropriate ‘unit of publication’, and should include all results relevant to the hypothesis. Coherent bodies of work should not be inappropriately subdivided merely to increase publication count." (I am concern about this point, as this manuscript described an open-source data of gait movement, without any data analysis or hypothesis testing. Although I agree that the data are useful and worth publication, the authors could divide the results of their study for few more papers.)

"The submission must describe original primary research within the Scope of the journal." (The authors mentioned that the objective of this study was to provide a rich gait movement data with fluctuations in speed, and I don't believe this was a primary research.) The submission should clearly define the research question, which must be relevant and meaningful. (Please see response above)

Reviewer #3 (Srinivasan)

The authors could refer to new/emerging guidelines by some funding agencies (NSF, etc.) and some journals (Royal Society journals) that insist on making available all data funded by them or published by them. And your example could be a good model for such ‘required’ publication of data.

[moorepants]

I would like clarity from the editor on this issue. The first reviewer is obviously rejecting the paper but the editor doesn't seem to mind.

I am specifically trying to buck the norm here and help data papers become an "appropriate unit of publication". One reason we don't have access to good data is because collecting data, sharing it, and reporting on it is not valued as a scientific contribution. Wikipedia has a small article on the purpose of the data paper: http://en.wikipedia.org/wiki/Data_paper.

I am not explicitly trying to spread the work into more publications, at least not more publications that are frivolous. But I do think that every study can support a number of valid publications and I believe that every dataset can support multiple independent research questions especially for meta-studies.

@tvdbogert This work fell under the current NSF grant. I'm curious what, if anything you all wrote in the data management section of that proposal. If you said anything about the data maybe this can help justify its publication.

[moorepants]

This is a new paper on why people share data that may be relevant to cite or simply provide in the reviewer responses: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0118053

[moorepants]

I emailed Dan to see about the data management plan they wrote.

[moorepants]

Dan shared the data management plan with me. You can read it here: http://embeddedlab.csuohio.edu/BBO/proposals/Simon%20-%20NSF%20SCH/

Here is one relevant quote:

After patient data are obtained for the proposed research, we obtain further IRB approval to release de-identified test data for the use of other researchers in the area of human motion and prostheses. Test data will be freely available on a web site maintained by the research team. Data such as these are not currently available in readily-accessible form, so although this is not our primary goal, it will still be a significant contribution of the proposed research to the prosthetics community.

and this:

Work performed under the proposed research will be made available to other researchers and students on a project web site maintained by the PI, which will be available to the general public. The proposed research will produce at least two general-use software packages, including source code, which will be available on the research project web site. One software package will include prosthesis simulation software. Another package will include simultaneous subsystem optimization software. This software will allow other researchers to replicate research results and to experiment with their own optimization and control algorithms. The research project web site will also include papers and reports that result from the research. Dr. Simon has emphasized broad access to the results of his other NSF research at his biogeography-based optimization (BBO) web page, http://embeddedlab.csuohio.edu/BBO. The web page includes interactive software and source code. The availability of the code has encouraged many researchers and students in the USA and around the world to further investigate and extend his NSF-funded research.

So it seems we have agreed to share the data and software code in the grant. I will clarify this for the reviewers and add a sentence referencing this in the paper.