Technical recommendations and requiremens

Technical requirements

Data durability

The most important technical requirement imposed on the system is the durability of the stored data. This involves the need to assign persistent global identifiers (DOI^[1]) (cf. Organizational requirements and recommendations ). An important consequence of this is the condition of non-removability of metadata records, along with the deposited data. Moreover, published records cannot be changed, even if they are no longer up-to-date.

Multiversioning

The appearance of newer (better, more relevant) research data in the author's activity results in the need to create a new version. This implies the need to implement multi-version ^[2]. In addition, to remove the danger that someone might still unknowingly try to use a previous version, the system always shows all versions of the research data in a single view and explains what the changes in each successive version are and what the differences are between them.

Data deposition

The responsibility of researchers for the reliability and quality of research data forces the system to allow the researcher to enter the description of his research data himself. This entails the following solutions that have been introduced:

1. A new user-friendly data entry interface has been created, such that the researcher can easily enter a metadata record and deposit research data in the repository.

2. The data entry process may take longer than the typical data entry process in transactional databases, and it is organized in such a way that

   1. it can be interrupted at any time to return to it later, without loss of data;

   2. it is possible to ask other researchers (co-authors) to co-edit the research data record

   3. it is possible to consult a data steward - communication between researchers and teams of editors/stewards was implemented, the subject of consultation can be technical issues related to data entry, as well as substantive, e.g. consultation on the manner of sharing, licensing;

   4. a collaborative process (workflow) between researchers and data-stewards has been implemented;

Given the above assumptions, the system was designed and implemented to support the processes:

a. Data entry (data editing phase)

b. Verification of data by a team of editors/stewards

c. Assigning a DOI identifier

d. Creation of a new version

Recommendations on the IT infrastructure

It is recommended to verify:

1. System security procedures and resistance to attacks

2. Safeguards for the sustainability of information resources and their resilience to hardware failure

[1] Assigning a DOI to research data is understood, as in publications, to be like "publishing" the research data, so just like you cannot change an article once published, you cannot change research data once published. One can only "publish" a newer version

[2] The multi-version feature is a mandatory criterion for research data repositories in DataCite requirements.