2019-04-25 Outreach WG Agenda and Notes

MG: Two-part questions should be separated into individual questions. Questions should be clustered by topic; the order right now seems arbitrary. Question 9, “Would you participate in an open software development effort to create the next great ARK resolver?” raises the question of whether such an effort already exists, whether it's meant to ask about participation in ARKs in the Open, etc.

KE: This was meant to be a followup to the question about contributing money or effort. Open source resolver development would be a sort of stretch goal for ARKs in the Open, but whether we pursue that would depend on the answers to the other survey questions, which would gauge interest. We should also add a question about followup and whether respondents are willing to participate in a focus group.

JH: The survey should also have some introductory text which provides enough context for people to respond. Some respondents will have a fairly shallow understanding of the project.

KE: We can describe why we're doing the survey and link out to other webpages on ARKs and ARKs in the Open.

TS: Question 10, “Is your organization a member of Duraspace or Lyrasis?” is demographic. Are there other demographic questions we want to ask about people's institutions, such as type or size of organization?

SP: On the original survey, we asked name of organization, country, number of employees, type of institution, and tax status. Those questions should be the same if we publish multiple surveys, to create consistent data.

KE: We'll go back to the original survey and be consistent with that.

TS: We might have to adjust the survey for different audiences, depending on how much they know already about ARKs. We're asking about potential benefits of three things: the identifier, the resolver, and ARKs in the Open. Do we also want to ask respondents about the risks or concerns they have regarding these three things?

KE: We can take a stab at incorporating prompts for negative feedback. Once we have a good draft, we can prototype it by sending it to this working group and then improve it before sending it to wider lists.

TS: The first question asks about the identifier and resolver, but links only regarding the resolver. Can we disentangle these? Additionally, most online survey tools don't send you out of the survey. Usually, introductory information is in the invitation email text or in the introduction to the survey itself.

KE: We were using SurveyMonkey but it had usability issues, so we might use some other tool. It's certainly difficult to distinguish between identifier, resolver, and service used to manage identifiers, so the survey should be clear about those categories.

SP: For multiple choice questions, we can include an uncertain answer for those who don't have an opinion or conclusion.

KE: Should we ask how much people are paying for ARK resolvers?

TS: It's okay if we give people the choices of "don't know" and "prefer not to answer." Sometimes vendors will ask you not to share costs.

MG: Could we be more explicit about what we mean by "community-run"? Question 8 ("Would you be willing to use a community-run ARK identifier management service (similar to EZID.cdlib.org) instead of your current solution?") may imply that EZID is community-run, rather than implying that EZID is an identifier management service.

JH: That's really a question of whether respondents would trust a service. There's really no basis for trust in something as general as "community-run."

Discussion of other organizations to compare to: Duraspace, Hathitrust, DPN, IIIF, Internet Archive.

PC: Is "consortial" an appropriate replacement for "community-run"? It could more specifically refer to a member organization, nonprofit, etc., which might be what we're going for.

JH: In order to promise service and reliability, the organization can't be too loose. It needs some substantial structure that people can draw up an agreement with.

KE: What we're trying to get at with that question is that some organizations are constitutionally opposed to systems not invented there. The real question is, if someone had an offering that met all your needs, would you use it? If the overwhelming response is that organizations want to run their own, then it doesn't make sense to develop a consortial one.

JH: The question could be, would you prefer to run your own resolver (or identity management service), or acquire the service through a service provider?

TS: And then a followup question could ask what sort of vendor you'd prefer to use: for-profit, non-profit, membership consortium, etc.

KE: My thinking is that it's more efficient to share a single system, but we want to find out whether other people agree. We'll incorporate this feedback and share a new draft with you early next week. We'll also learn more about the survey tools we have available and how that will shape the survey itself. Were there any questions glaringly absent, for the sustainability working group to develop membership models and services?

TS: Do we need to know if people use other identifier systems in addition to ARKs, and if so, why?

KE: I assume they do, because there are so many domain-specific identifiers. It might be interesting to know what organizations use ARKs for.

JK: The original survey did ask such questions, but it might make this survey long to put it here.

PC: I wonder if it might make sense to do a survey first about current use which also educates people about the benefits of ARKs and the range of uses at other institutions. The responses to the sustainability survey might be more positive if there were an outreach survey first which educated as well as solicited information on current practices. We can provide multiple choices for things that we know are in scope, such as types of documents which ARKs resolve to, as well as an option to provide your own answer. If the sustainability survey is urgent, it might need to go out sooner, though.

KE: The two working groups could have complimentary efforts if the outreach working group compiled information complimentary to the survey, rather than another survey.

review draft FAQ

From J Howard email:

• The most common question I encounter about ARKs is, what are they used for? The FAQ doesn't address this or the more general question of why there would exist the various other identifier schemes that are referenced. (At the end, the discussion of ORCIDs and UUIDs addresses purpose more clearly--I think that we need to address the purpose and applications of ARKs with equal clarity.)
• In the question 'How do ARKs differ...", I'm not clear about what "non-siloed" means exactly. Also, particularly in light of the Sustainability FAQ's several questions about costs of ARK services and the existence of paid services such as EZID, is it appropriate to refer to ARKs as being "non-paywalled." Clearly use of ARKs is not "free" for users of services like EZID, and use of ARKs may not be free for many users in the future. If you mean something different from "free of cost" then I think that has to be made explicit. I generally think of a "paywall" as a vendor's mechanism for managing access to paid subscribers; can't ARKs also be potentially used for resources that are commercially access restricted?

I wonder, too, whether it might be useful to note that identifiers refer only to resource location without implying anything about the attributes of whatever item might reside at an address; in other word, perhaps work in a sentence or phrase that simply notes that content represented by a persistent identifier is subject to "content shift"?