open-discussion
open-discussion > RE: AAAS: Your Paper MUST include Data and Code
Mar 10, 2011 07:03 PM | Cinly Ooi
RE: AAAS: Your Paper MUST include Data and Code
Originally posted by Cinly Ooi:
Big oops.... Rereading the editorial, an argument can be made that they wanted the data to be made available for futher research.
In this case, reasonable controls, e.g. data is only handed over once a REPUTABLE ethics committee approve the new project, will be entertained by Science.
My reading of the editorial is computer code is included for the purpose of reproduction and not 'extension', but I mgiht be wrong. If so, I am not really sure how this can be enforced. Matlab won't be a problem as it is naturally on the exclusion list because it is too generic, far away from the processing chain and in theory, can be reaplced. SPM, however, is integral to the analysis and must be on the inclusion list. But suppose SPM is proprietary, then we have a problem. The way out is for the proprietary vendor to outline precisely the steps taken in a way that is reproducable and to permit the steps to be reproduced. Still very difficult.
Originally posted by Torsten Rohlfing:
The journal in question is interested in the reproducability of the data, so, the condition can be as stringent as reproducing the results using the same method described by the paper. No other use permitted. I don't see this as controverial.
Fair
enough - I guess even if you hand-edited images you could provide a
sort of "diff" to "patch" the original to form your modified
version.
Back to the other aspect of my concern though - under what conditions are data made available in the first place? If I use data from someone who wants, say, the first-born child of everyone who downloads it, most of us would probably agree that such data are not really "available" (even though the first-born son clause is probably not enforcable, but I am not a lawyer either). But what about data use conditions, for example, where any data user who writes a paper is required to list the data provider as a co-author? Similarly for software. Again, as a random, fictional but prominent example, if the SPM developers had chosen to require co-authorship for themselves on every paper that uses SPM, would the community have accepted that? And if some researchers did accept this condition whereas other find it unacceptable, would the community be willing to accept SPM as generally available (albeit not under reasonable terms) and would Science under their new policy publish such papers?
Back to the other aspect of my concern though - under what conditions are data made available in the first place? If I use data from someone who wants, say, the first-born child of everyone who downloads it, most of us would probably agree that such data are not really "available" (even though the first-born son clause is probably not enforcable, but I am not a lawyer either). But what about data use conditions, for example, where any data user who writes a paper is required to list the data provider as a co-author? Similarly for software. Again, as a random, fictional but prominent example, if the SPM developers had chosen to require co-authorship for themselves on every paper that uses SPM, would the community have accepted that? And if some researchers did accept this condition whereas other find it unacceptable, would the community be willing to accept SPM as generally available (albeit not under reasonable terms) and would Science under their new policy publish such papers?
The journal in question is interested in the reproducability of the data, so, the condition can be as stringent as reproducing the results using the same method described by the paper. No other use permitted. I don't see this as controverial.
Big oops.... Rereading the editorial, an argument can be made that they wanted the data to be made available for futher research.
In this case, reasonable controls, e.g. data is only handed over once a REPUTABLE ethics committee approve the new project, will be entertained by Science.
My reading of the editorial is computer code is included for the purpose of reproduction and not 'extension', but I mgiht be wrong. If so, I am not really sure how this can be enforced. Matlab won't be a problem as it is naturally on the exclusion list because it is too generic, far away from the processing chain and in theory, can be reaplced. SPM, however, is integral to the analysis and must be on the inclusion list. But suppose SPM is proprietary, then we have a problem. The way out is for the proprietary vendor to outline precisely the steps taken in a way that is reproducable and to permit the steps to be reproduced. Still very difficult.
Threaded View
| Title | Author | Date |
|---|---|---|
| Luis Ibanez | Mar 10, 2011 | |
| hongtu zhu | Mar 13, 2011 | |
| Luis Ibanez | Mar 13, 2011 | |
| Matthew Brett | Mar 13, 2011 | |
| Isaiah Norton | Mar 13, 2011 | |
| Torsten Rohlfing | Mar 10, 2011 | |
| Luis Ibanez | Mar 11, 2011 | |
| Daniel Kimberg | Mar 10, 2011 | |
| Cinly Ooi | Mar 10, 2011 | |
| Torsten Rohlfing | Mar 10, 2011 | |
| Cinly Ooi | Mar 10, 2011 | |
| Torsten Rohlfing | Mar 10, 2011 | |
| Cinly Ooi | Mar 10, 2011 | |
| Torsten Rohlfing | Mar 10, 2011 | |
| Cinly Ooi | Mar 10, 2011 | |
| Matthew Brett | Mar 10, 2011 | |
| Pierre Bellec | Mar 10, 2011 | |
| Luis Ibanez | Mar 11, 2011 | |
| Matthew Brett | Mar 10, 2011 | |
| Cinly Ooi | Mar 10, 2011 | |
| Cinly Ooi | Mar 10, 2011 | |
| Torsten Rohlfing | Mar 10, 2011 | |
| Daniel Kimberg | Mar 10, 2011 | |
| Cinly Ooi | Mar 10, 2011 | |