Codes are currently used by a large number of researchers to study a specific research problem or to produce results needed for a given publication. Once the work is published, some researchers decide to keep the developed codes to themselves while others make the codes available so that interested researchers can reproduce the published data.
Based on your experience, what are the benefits of sharing a source code (e.g. Matlab code or a User-Defined material model) by placing the code and the data in a public repository offered by the publisher or through a university website.
Following up on the previous responses, I wish to get the readers opinion on the following code-sharing project.
The geo-group at McGill has developed a coupled finite-discrete element code (based on the Open Source Code YADE) that has been used over the past 10 years to analyze complex soil-structure interaction problems. The code (currently being used by graduate students) has allowed the group to develop solutions to a range of soil-structure interaction problems and understand the micro-mechanical behavior of granular material in contact with flexible and rigid structures.
We are now in the process of adding a user/programmer guide to help researchers outside the group to use the code such that they can either reproduce published solutions or solve other problems that require advanced numerical analysis.
By doing this, it is hoped that the code would help a wide range of researchers (as opposed to only group members) and at the same time give our published work related to code development more visibility.
Your thoughts on this are highly appreciated.
Shared source code lets other researchers replicate your work, find faults, customise it for local use or build it further.
I benefited from meeting a key person at a US conference, and obtaining his source later. I returned the favour to the community when we published our important paper "The recursive evaluation of the M/M/r/K/K queue", available now via RG. Nobody has yet pointed out any defects!
Thank you Dr. Kennedy for your input on this.
I agree that published research should be reproducible. Making a source code available to interested researchers will allow for the extraction of additional data that may have not been available in the published work.
That said, I am trying to get a perspective on the benefits to the researcher who shared the source code? For example, did you cite his work? or clearly acknowledge in your paper that the code was developed by someone else?
If authors know that sharing their source codes would result in more citations and will promote their work, I believe more researchers will be willing to participate.
Yes. He was #4 of my 6 references. We did not use his code, but were inspired by his idea and use of recursion.
Of course there are negative aspects as well. A big Software company may take your source, and incorporateit in their corporate giant software. How can you prove that it was your algorithms they used? How can you sue such a giant even if prior to publishing you had taken the precaution of patenting your intellectual property?
Thank you Dr. Kennedy for pointing this out.
I agree. No guarantee that an algorithm or a source code of any commercial value will not get into the hands of giant software companies who may or may not give any credit to the source.
In this case, do you recommend that a code be given only if requested from the author.
Thank you Hans for providing your thoughts on this.
People are more comfortable sharing their codes with known colleagues who can become potential collaborators.
Public code sharing is now encouraged by many journals and I believe there is some hesitation on the part of researchers to actually upload their codes online. They would rather wait for the code to be requested so they can send it privately. This may result in several missed opportunities to expand the developer network and gain more visibility. However, authors may argue that they would like to be in control and choose who uses the code and clarify their expectations.
I agree codes developed using public money should be made public, however, changing the existing culture on both sides (the developers and the users) may take a little bit more time.
Following up on the previous responses, I wish to get the readers opinion on the following code-sharing project.
The geo-group at McGill has developed a coupled finite-discrete element code (based on the Open Source Code YADE) that has been used over the past 10 years to analyze complex soil-structure interaction problems. The code (currently being used by graduate students) has allowed the group to develop solutions to a range of soil-structure interaction problems and understand the micro-mechanical behavior of granular material in contact with flexible and rigid structures.
We are now in the process of adding a user/programmer guide to help researchers outside the group to use the code such that they can either reproduce published solutions or solve other problems that require advanced numerical analysis.
By doing this, it is hoped that the code would help a wide range of researchers (as opposed to only group members) and at the same time give our published work related to code development more visibility.
Your thoughts on this are highly appreciated.
I think shearing a source code be beneficial. If a new researcher uses your code or program, will acknowledge your code and cite your code in the journal. It can help lot of new researchers and help produce more effective way of analysis.
As I am from Bangladesh and when I was doing my undergrad thesis, I use PIVLab. A open source MATLab program created by a German PhD student. My work was related to soil erosion through the defective pipe. I use their code for analyzing the particle velocity of eroding soil. But that German PhD student use that program for another purpose.
Helping another researcher will not harm you work but will strengthen your work. In my sense, the research is done for the wellbeing of mankind not for personal wisdom.
Hi everyone,
I very interesting discussion. I will link a recent article on data sharing
https://www.bmj.com/content/360/bmj.k510
Hi everyone,
I very interesting discussion. I will link a recent article on data sharing
https://www.bmj.com/content/360/bmj.k510
https://www.bmj.com/content/360/bmj.k510
Hi everyone,
I very interesting discussion. I will link a recent article on data sharing https://www.bmj.com/content/360/bmj.k510
Hi everyone,
I very interesting discussion. I will link a recent article on data sharing
Data sharing in medical research
BMJ 2018; 360 doi: https://doi.org/10.1136/bmj.k510
Hi everyone,
I very interesting discussion. I will link a recent article on data sharing
Data sharing in medical research BMJ 2018; 360 doi: https://doi.org/10.1136/bmj.k510
The user manual for the "Coupled 3D Finite-Discrete Element Analysis Tool" has been made available "for request" on Research Gate. The code and the user manual will be made available for interested researchers on the Geo-Group website next month. http://geogroup.mcgill.ca/
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publisher