The Australian government, through research management/funding bodies, have published articles in relation to responsible research.
Ownerhip can be a very blurry line current day and age, but all entities/people associated with a research project have a role to play. The list below is just a small sample of guidelines/suggestions:
- Place the primary copy of you research data on a resilient storage environment as soon as possible after collecting/generating it.
- Ensure that metadata exists for all research data and is easily associated with the research data - e.g., store in same location(s).
- Where possible, identify and use resources/services provided by major research groups (national or international).
- Maintain at least one more resilient copy of you data in a location that is geographically diverse from the primary copy.
- Many researchers are collecting data (e.g., photos/video of nature) that can never be recaptured. Above 4 points become critical and deletion of files should be very carefully considered in this case - e.g., you should ensure that data and/or metadata is not lost when modifying/converting the original files before deleting the originals.
- Never delete files (data or metadata) associated with a publication.
- Clean up (delete) files that have no value - e.g., if there is no metadata, the data may/will have no value.
- Associate each of your publications with all research data/metadata them.
- Use server class infrastructure (e.g., a HPC facility or cloud services) to perform all computational research.
- Where possible, keep a record of software/hardware (inc. versions) used in computational research - this is metadata.
- For researchers involved in software development, use version control software (e.g., git) and services (e.g., GitHub or BitBucket).
- Ensure that you have permission (entitled/licensed) to use all software that is involved in your research project(s).
- Ensure that you have permission to share your research before doing so, especially if your research involves sensitive data.
Never forget IT security:
- When sharing data, read-only permissions are safer than read+write. Will you be able to identify and revert all changes people with read+write privileges could make?
- Don't share your account (or password) with anyone. You could be held accountable for any/all actions associated with your account.
- Key and certificate based authentication is more secure than password based authentication, providing you don't share private information associated with the authentication (e.g., passphrase and/or private key).
- Two factor authentication is more secure than a password, key, or certificate authentication. Preference providers that offer two factor authentication.
- Always use and configure secure services/protocols (e.g., VPN). Avoid using insecure/plain-text services protocols (e.g., http, telnet).
- Keep all your computing environments up to date with security patches.
Efficiency & Sustainability:
- Avoid re-inventing the wheel. Search for software/libraries that reduce the burden of your programming efforts.
- Preference higher level languages (e.g., R) for your programming efforts - will usually decrease time you spend in development.
- Search for alternative algorithms/methods to computational research. E.g., preference algorithms/software that can be run on multiple CPU cores in parallel.
- On HPC facilities and cloud resources you have the ability to scale up. Even if your software can only use 1 CPU core, you should be able to modify your workflow to run many jobs in parallel.
- Think about the costs of your research plans. E.g., Do you really need to use and store (long-term) 4K or raw-format video?
Responsible Locations for Research
|Classification||Computational Example(s)||Storage Example(s)|
|Public cloud||AWS, Azure||AWS, Azure, OneDrive|
|Research cloud||ARDC (e.g., QRISCloud)||ARDC (e.g., QRISCloud), AARNet CloudStor|
|HPC environments||NCI, Pawsey, JCU HPC||NCI, Pawsey, JCU HPC|
Many national/international research programmes (e.g., GeoScience Australia) have facilties and, more importantly, services available for use.
PS. Contact eResearch or HPC staff if you need advice, clarification, and/or assistance.