Does anyone have any advice or experience implementing FAIR principles in the context of closed systems i.e secure platforms, or secure research environments. To reach the highest levels of FAIR, it appears one must make data and metadata publicly available (for the most part).
In closed systems, it is not feasible nor practical to adopt many of the FAIR facets, as they are defined in much of the existing literature. Instead, can we think about interpreting the ‘world’ of a secure system, as the ‘world in which the system lives’, and ‘FAIR usage’ as FAIR usage within the bounds of the system?
Thank you for opening this interesting topic!
One question for clarification, the secure research environment you mention, is this an environment used during research only or also for archiving and long term preservation?
Accessibility is indeed an important part of FAIR. In case your data cannot be publicly available, maybe you could look into publishing only the metadata?
@millie See Maaike’s comment. Thank you!
Thank you for your response.
It’s a secure environment for archiving and preservation too.
We have considered publishing only the metadata, however, in some instances, this isn’t feasible. My question therefore pertains to a situation where publishing in the public domain is not an option.
Hoping this clarifies the points you raised.
Thank you for the clarification @millie !
I think it’s important to keep in mind here that FAIR does not prescribe openness. The extent of openness of any dataset is to be determined by the data owner. To satisfy Accessibility in FAIR, it is important that it’s clearly communicated how certain data can be accessed, even if the answer to that is ‘It cannot’ or ‘Only registered users within this system can access’. Such information would ideally be presented as transparently as possible.
If your system is flexible enough, it would be best to decide on possible FAIRness and Openness for each dataset separately. For each dataset that can have metadata or data publicly available is a valuable addition to public knowledge.
For the datasets for which this isn’t possible, it would be good to think about a different way to clearly and transparently communicate accessibility. If the whole system is inaccessible, a blanket statement on the system’s or organisation’s landing page could be a way to do this. Alternatively, maybe it’s possible to only make public the metadata elements that should be publicly available, while keeping other metadata elements closed?
I don’t have enough details about the system to make more educated suggestions, but in general I would recommend to think about a tailored solution for your system together with the technical staff. Again, the data does not need to be open, but you could still increase your FAIRness by documenting the relevant information regarding access.
I hope this helps somewhat!
Six Recommendations for Implementation of FAIR Practices
FAIRsFAIR Data Object Assessment Metrics | Zenodo
Cloudy, increasingly FAIR; revisiting the FAIR Data guiding principles for the European Open Science Cloud - section 4, subsection “FAIR is not equal to Open”
FAIR-Aware Q4 and Q2
Thank you for the comprehensive response @maaike.verburg.
At present, the entire system is inaccessible to the general public. One has to be granted access to the system and applications held within, in order to access the catalogue and data. Within the system, access levels differ - so I suppose my question is - can we interpret the users of a proprietary system as a sort of ‘system’s public’ in this case? And therefore, when we’re thinking about access, could we look to transparently communicate access to users of the platform.
Similarly with openness, levels of data openness will differ for different users; we already make the metadata available to everyone on the system; but users will only have access to the data, if their role permits it.
Thank you for sharing the resources.
In my personal opinion, this could indeed be a good way to approach your situation. This does not facilitate true accessibility as prescribed by the original FAIR principles, but being as FAIR as you can be sometimes calls for these interpretations and choices. The question of communicating outside of the system could maybe be a future endeavour (e.g., making a public page about the system without granting access to the system), if this is something that you feel has added benefits.
Working on transparency will always reap benefits, this will make your system more easy to use for your users. Good luck and thank you for posing this interesting question!
FAIR and Open are orthogonal concerns. It may help to view FAIR as “Open-able” — the FAIR principles’ emphasis on clarity in authentication and authorization protocols, and in licensing, should help to make “opening up” a FAIR data system straightforward if one so chooses.
Here is a nice page from a commercial materials informatics platform provider on the value and interpretation of FAIR for closed research data environments: < Why Data FAIRness is important in the corporate world - BLOG - Citrine Informatics>.