Open bluerasberry opened 9 months ago
Some answers.
1. In our Resource Hub, what link can there be between these resources and anything else, like an organization, that we want to model?
My first reaction is, Do we credit organizations, or do we credit persons who are part of organizations?
As you well point out, it doesn't seem easy to establish univocal and well-documented relations between resources and organizations. At least in terms of inputs, hence to respond to the question, "Which organizations supported this resource?". I envision this support happening through funding (financial) or partnerships (non-financial, in the shape of laboratories, staff, networking, etc.). Funding information will be stored separately in each project profile. Non-financial support could be attached to the project by stating, for instance, partnerships within the project.
Specific resources, such as databases, publications, open tools, etc., tend to be linked to the project first, then to institutions that are part of the project. Following this reasoning, a resource could be related to more than one project as it gets reused in another context (let's take the ELEMENT's dataset, for instance). It can be the case that a specific institution supports a specific resource, but I don't see this kind of information out there very often.
Back to resources, it makes more sense that resources are linked to those persons who develop/maintain them (authors), those who use them (users), the context where these resources are used/produced (projects, publications), and where they are stored (repositories). Usage won't be part of the project profiled but is something that will be possible to map as credit has been established and resources become findable.
2. Is there more published information about a link between these resources and an organization, but which is not in the text of these resources themselves? Normally, this would be either on published outcomes (white papers, research papers, reports, etc.) or web pages. However, these attributions are normally not explained in a steady or straightforward way. This is not usually the case in the resources I have mapped so far. Resources we are mapping –i.e., data, tools, publications— have different credit and archiving practices even within each resource. Among all, publications and data tend to be the most steady and transparent. Open software and hardware tools are more loosely credited. When a paper is published, links between resources and authors are established in the paper. Hence, the link with an organization tends to happen through the author(s). I can envision why an organization could support a specific resource within a project in specific cases and ask authors to credit their involvement, but I don't recall this as much in the project I have profiled.
3. We already have in draft a project profiling dataset in which we can assert citable authority to have confirmed links between organizations and resources. If we somehow have knowledge that there is a link between a resource and an organization, but there is no published media which verifies this, then under what circumstances will we be readable to make such claims and verify them with our own fact-checked dataset?
What do we do in a situation where don't have (public) evidence of the association of a Research Project with a Research Organization? Cone of uncertainty? Problem of aliasing?
Solutions to how to account for the Research Project - Research Organization association/link.
Lack of public (structured) evidence:
Project Alias
Cone of uncertainty We are facing a problem that will keep arising as we encounter other multi-layered items to catalog (e.g., a project with different names depending on where and to whom it is mentioned, a resource poorly documented). The problem, as I see it, points to something crucial: how do we deal with rather "unstable" or "low-intensity" entities (researchers, resources, etc.)?
[Note resulting from the March 13 meeting; due lack of time to expand on this]
To the problem of lacking public structured evidence linking resources to projects and organizations, if we move forward with option 1, and this dataset asserts citable authority, we will need to standardize the dataset to provide the evidence that is needed to support the Wikidata project profile. We will especially need to focus on standardized information in specific rows (i.e., publications, software repositories, etc.). My question then is (and I am sorry if I am missing something here): Could we persistently identify resources that may not have a PID, such as GitHub projects, using URLs? If this is the case, our authoritative list would end up looking a lot like the second option (producing provenance through links).
This relates to Don's point about inconsistency in data in rows from the project list (https://docs.google.com/spreadsheets/d/1IjcN4BHbFOuHR5A0d5RWvS56cvt8AyYH2pr-yJc9jso/edit?usp=sharing).
Summary: When we have information that an organization has produced a set of research resources, such as papers and datasets, then under what circumstances do we credit that organization with producing them? The general challenge which may surprise some people is that many publications fail to credit organizational support or affiliation, so verifying project or institutional relationships is non-trivial.
Consider the organization Consortium for Resilient Gulf Communities https://www.wikidata.org/wiki/Q124386729
It has received this grant https://www.wikidata.org/wiki/Q124386785 and likely others.
We have unverified but plausible information that the organization put out some research resources. Of those resources, some report a connection.
"the authors wish to thank our collaborators at Consortium for Resilient Gulf Communities ..." https://doi.org/10.1016/j.pdisas.2019.100042 "Our in-person survey was part of a larger project, the Consortium for Resilient Gulf Communities" https://doi.org/10.3390/socsci8010008 "The data for this study is derived from a survey conducted through the Consortium for Resilient Gulf Communities, established in 2015 with a grant from the Gulf of Mexico Research Initiative..." https://doi.org/10.3390/socsci7100203
The other sources do not credit Consortium for Resilient Gulf Communities, but for various reasons, seem related. https://doi.org/10.5670/oceanog.2021.125 https://doi.org/10.1016/j.dib.2020.105390 https://doi.org/10.1016/j.jrurstud.2019.12.019 https://doi.org/10.1016/j.ijdrr.2020.101716
Question: What is the relationship between these resources and any organization?
The reason we want to know this is for our data modeling. More specifically