Name:
One identifier to rule them all- Or not
Description:
One identifier to rule them all- Or not
Thumbnail URL:
https://cadmoremediastorage.blob.core.windows.net/8dfa446e-2c78-43ef-92aa-f12bc79ed027/videoscrubberimages/Scrubber_1.jpg
Duration:
T00H33M42S
Embed URL:
https://stream.cadmore.media/player/8dfa446e-2c78-43ef-92aa-f12bc79ed027
Content URL:
https://cadmoreoriginalmedia.blob.core.windows.net/8dfa446e-2c78-43ef-92aa-f12bc79ed027/One identifier to rule them all- Or not.mp4?sv=2019-02-02&sr=c&sig=oyzRfPiodfbDyfio5ECPSW%2BP7V%2Blmi0tIaJ92AlH6Wc%3D&st=2024-09-08T00%3A50%3A21Z&se=2024-09-08T02%3A55%3A21Z&sp=r
Upload Date:
2024-03-06T00:00:00.0000000
Transcript:
Language: EN.
Segment:0 .
Well, everyone was letting people join the Zoom link as we can get the discussion started. The numbers are going up quite quickly. Great thank you, everyone, for joining our session. Thank you, Jonathan, Gail and Robert for filling in for Beth.
I really if you haven't if anyone has any questions, please feel free to introduce yourself in the chat and ask them or raise your hand or turn on your camera. The more questions, the better for this discussion, this to get this started. It really is a human aspect. I'm interested in kids. So and in the whole identifiers, what the three of you talk about the human aspect and technology and how that could with so many identifiers growing all the time, how can we work in interoperability?
So if I can go first, maybe. I think there is also a question in the chat, so I'm happy to answer yours. So what I told during the presentation was the human aspect is also important because of course there is some specialization, some knowledge of the PID and its context.
So I think what, what we, we, we see and there was something which was also evoked in the previous session about this. With that we need maybe a forum or a platform to be able to exchange. And of course, there are some technical issues. So as far as the human aspect is concerned, I would say there are not so many places where we can work together.
I think that everybody is very much involved in the daily activities of its own organization is or our own organization. We are for example, at the I mean is a network is comprises 93 national centers and we have regular meetings. We have regular exchanges with our colleagues. So I think the idea is when do we have some spare time to actually try to work with other identifiers, although we have some exchanges, you know, because for example, we use information data from crossref because we use that data to tell which journals are indexed in crossref.
So that's part of the I suppose also what we do at the moment. I mean, as far as cooperation is concerned, is more retrieving data from our partners and exposing it or displaying it in, in the eyes and polo and I know that, for example, we may have some questions from colleagues and, you know, there was one in the chat when can we assign analysis? And, you know, when is the good time to do so? And sometimes we have some queries coming from colleagues saying, hey, this journal cannot have the SSN or could you just, you know, get in touch with the National Center and see where the problem lies.
So I think for the time being, this is more, you know, exchanging data, retrieving data and cooperating, you know, not on specific questions rather than planning projects together. So that's what I feel at the moment. Yes and I will build on that. And thank you. And there are a couple of questions in the chat, which we can perhaps put into this.
I think it is a question of everyone's busy. Time is money and there has to be a reason for it. And that's partly a problem because it stops you thinking about future scenarios. But if you do have a strong use case, as some of the examples that I gave in my piece, I tried to focus on ones that when they came up for us for discussion, they had a very strong use case enough for people to stop the day to day work and say, OK, this is something that's worth working on.
And I think at least that's a way forward. But I totally acknowledge that it stops the more speculative or future orientated possibilities. Oh Yeah. So I see Adam has sort of answered Howard's question. I was going to say raid there's a identifier out of, out of Australia. That's just been ISO approved or is in the processes of it and that's identifying research projects which in theory should be really upstream of everything, maybe not upstream of funding requests.
That would be a good point. And what else was I going to say? Yeah, I think it really is about I mean, certainly we get together a lot because we're a community and we have a lot of wouldn't it be good if and there are an awful lot of things like that, only a few of which really get taken forward. And I think that that's something maybe we can address.
But in the end, it comes down to, you know, we've got a job to do to keep the up and running and that's sort of our primary focus. And I think these are more facilitators. So there are just it's part of it's like a linchpin or something that you can use in within a system. And I guess, you know, there are systems for like trees or other systems that make use of various kids. And so they interoperate with them.
They integrate these ideas into their own system. I mean, plans or projects. And I won't add a whole lot to what Jonathan and Gail said, but I think that there may be a Gail mentioned a forum to draw the free body together to talk about this. There are actually maybe too many forums and that you can't get the right people to go to the right places at the right time, to all have these discussions about the interoperability, the true interoperability.
I know I see different people in different forums, but it's rare to get or specialize in having these ID issuing organizations all in one place to discuss the real interoperability and the technical and social technical work behind making interoperability actually happen. All right. Thank you, everyone.
Regina, would you like to ask your question since I haven't seen it in the chat yet? Jonathan, can I be heard? Yeah oh, wonderful. Jonathan, I was interested to see that great variety of agencies that assign assigned you assign dois. I am from the icis network. And and I'm wondering if they have different fees and if you occasionally find people who rather than going to the community, that seems more natural for them there in Japan, they realize that this other community charges less and they'll go there.
Do you see any of this fee shopping behavior or are all the fees the same? Are they standardized? No, no, that's a great question. Thanks, Regina. Regina no, they're not standardized. They're very much community focused. So I'd say it also depends a little on the way that the registration ANSI is funded.
So crossref, for instance, or data site are member driven member organizations driven by member fees? Kristie, which is the one in Korea is actually government funded and therefore all the organizations who are part of Kristie, it costs nothing for them. The Japanese a bit of a hybrid entity as a private company. And so each of the arrays actually and in fact, we think that's quite a strength because we don't think it's a good idea to be funded as an organization entirely through one business model.
I think there's some sustainability risk if that's the case. So we quite like that. And the same goes for the community organizing their own fees. We tend it's a really good question. I'm not sure I know the answer. I know that what we do have perhaps not that if someone comes with a very large number.
So we have a legacy system that they're looking to move to the Doi. So the most recent one is adeosun, but GMF has been another one where they have a very large number. Then you do get some negotiation because obviously if there's a fee associated with some sense of scale, then coming with 50 million identifiers or 50 million objects that need to be identified needs some negotiation.
But I think that I'm not sure if there's an answer to your question, but in essence, these things are quite. Negotiated there's room for negotiation, let me put it that way. I want a free drive from the government funded ANSI. Do I have to be in the country that government belongs to, or people refused if they go to the wrong community just to get a lower fee?
Gosh, that's a good question. I guess what happens if you speak korean? I mean, do you have to speak korean? In theory. I mean, if it was a scale, then I think someone would say, is this really such a good idea? I mean, if you want if you're an individual and you want one, you just go to the Au or any other. I don't want to give preference to one, but something like xanadu, you'll get a deal with Europe for your object if people do.
I think what we think what we really focus on are people that have a structural interest in not just one identifier or half a dozen, but a large number. And then I think you have some enter into a long term business relationship. I think those are the things that filter out the dealers, if you like. OK thank you so much.
But those identified wants to add something to that I just saw in this video. I had. Hi Hi. Thanks Ed Pence from crossref. And just to say that, yeah, it's some really interesting topics coming up.
So Thanks to all the speakers. And just to say, I think we're talking about persistent identifiers and the different registration agencies in the Doi Foundation case, but it's really about use cases and services. So it's not about getting an identifier and getting a number. It's what does that enable, what services are available. So there are many routes to get a Doi.
The models that each Doi has are different. But really, you have to whoever's looking to use it, I think, as with anything, is what are they trying to achieve? What's what's their use case? What what services do they want available with it? So I think it's a much broader question than just like, oh, it's this idea or that ID. And I think that's the case with the D Foundation and the different eras.
They're all very different. They all have different, different business models. So in that sense, it's not centralized in the way, say, ISP or et cetera. Thank you. Edward has a question. This session was titled one persistent identifier to rule them all.
Should there be a persistent identifier that is established much earlier in the research process that tied together the research publications, data sets, institutions, etc.? Or does one already exist? I think we sort of answered that in the chat. There is there's actually a really good, interesting study, a case study of what I said, the yet another position identifier.
Think there's it was out of Australia an Australian ANSI thought you know what would be really useful is if scientific or projects of any sort they need to have an identifier under which all of the work that's associated with the project, perhaps the funding, the art, the research, the data, everything associated with it would kind of be assembled in one place. And they got together and came up with a proposal and actually they went an interesting route, which which not everyone does.
They went straight for an standard, and that does take a number of years, but that's what they went for. And it's just been approved, I believe, or is in the process of being approved and they're rolling it out. I think we are challenged to us and we're talking with them is how we cover that. I think we cover it as an identifier like we do other identifiers, but it's something that we'll need to look at when it starts to scale.
I'll see if I can find the someone has posted and it has been published. Thank you, Matthew. The chat is alive. Thank you. There's anyone else have any question. So I was going to say to that we did some work here, some test bed type work with pids that assign pids at the point of data creation off of the sensor.
At the point of workflows, we set pids to the computational environment, to the intermediate data products and showed that workflows can be tracked and then grouped within pids. However, it ran into the problem of it being a testbed, so we had a very limited number of use cases who could use it. But really in the end, as far as reproducibility goes, you not only have to track the data and the publications, you have to track pretty much each individual aspect that the research goes through from collection of data to the publication, not necessarily just the end products, which are what are getting a lot of pids now.
You know, it's interesting. We looked I think this was Korea. They came up, there was a request to look at internet of things applications where we would have censors in the environment picking up and identifying, and those censors would have identifiers. And then the data coming from them would also and there was another use case in, I believe, marine biology, where the censors they put down in the ocean would also have identifiers.
And again, the data stream would be and there's a bit of work of the Ida on that. I think most of it, as Rob says, is sort of test bed implementations, but it has to come quite soon because there's a lot of data that's challenging because of the volume of data. But on the other hand, it's quite organized because it's a good structure, but those are things that we see occasionally coming up, although I haven't seen a solid use case yet.
Not not for us from our point of view. The other thing to mention, I think that Beth touched upon. There's a project at the moment called digital objects FDO, which is looking really into the, I think probably the not distant future, but middle future. Middle distant future is if objects could if you come across an object and it can self describe itself, so it has an identifier within the identifier around it is something that says, hey, I'm one of these, I'm a PDF and you need an Acrobat to read me or whatever.
That's a simple example, but it would be more complicated for data. But that's in theory a future proof implementation where you wouldn't necessarily need to know. I would say that and that potentially wouldn't involve humans. It would just involve their machine. But I do think it requires an awful lot of human thinking to make to set it up and set it going.
But if it ever happens when it happens, I'm sure it will. That will be a really interesting development to track. I don't know if you haven't seen the site, then I definitely recommend taking a look. Thanks, Howard. I just.
Oh, that was it. Joe had a joke. Sorry goes. Clarify, clarify. Were you saying that the manifestation and the format of the thing being identified would be encoded in the identifier itself? Is that what you were saying? Not encoded, but accessible.
So if you like, you could see it as there's an identifier at the center and around. It is not just metadata. Well, there's metadata, but that data would in fact self describe. So it would give you information as a future machine on how to use the object. And what the use and how to use it. It's brilliant in its.
Sorry go ahead. Sorry but that's bundled along with the identify rather than being part of it, so to speak. Well, it depends what you mean by part of it. So if you can imagine, everything. So if you have, let's say, an identifier point to, let's say a handle record, it sets the handle system, it points to a handle record. That record could contain either information itself or links to other information.
But for the machine, it doesn't really matter as long as it can follow and find it. So it's a little bit open as to whether the object, the record sitting on a system has to contain the information. But it could just be a link off to something where it had that information. Both those the specs are coming out right now, but both of those options would work. In essence, the machine could follow a trail to discover not only what the object is, but how to use it.
That robs the expert here. Am I doing it justice, rob? Yes, yes, absolutely. The identifier and would point to a kernel of descriptive information. Right and from that descriptive information, you can find the actual data, which is described within that descriptive information. So you get a resolution of your PID with some kernel of information that describes what the object actually is that you would eventually point to.
But as Jonathan said, a machine would only need the description, and then they could take actions based on that, including doing something with the actual. Data within the digital object. It's it's what is described in the principle of for identification of identification that you need a minimal I mean, some minimal data to be able to actually identify uniquely the reference.
So that's very similar or this is in line with this principle. I think the main difference is that that. A lot of their current best practices are focused on humans and human readability and human understanding. For instance, we get a lot of requests about the semantics in the, in the Doi and the identifier, which is kind of funny really, because they only really matter to a, to a human.
The computer doesn't care what's there as long as there's something there. And I think this is really looking forward towards a time where machines are doing a lot more. So it's not just having it readable, but it's machine readable. That's the key. And also, to some extent, if it's possible, machine readable for machines that don't exist yet.
So it sounds a bit like a Hitchhiker's Guide to the galaxy like, but the machines that are to come and Ed said, what other examples. Only testbed examples. And that's one of the things I think there's that part of the reason why it's still. And I start up in the but it's work in progress let's say that. I think the most prominent example is the National Library.
Our national history museums in Europe are looking at providing digital twins of objects that are located in various different museums. So that's kind of the go to case would be this digital object would be a representation of a actual physical specimen located in a vault within a Museum or library or some such. That work's being done in the Netherlands.
And I forget the organization. Disco disco? Yes, disco. I'll take up the link. I think. Charles sorry. Charles, do you have a question? Yeah I was going to come back to one ID versus proliferation of peds.
My main concern, which I haven't heard discussed here, is if there's lots of organizations issuing peds or even if there's one organization which I don't think we'll ever be feasible. Are they going to resolve in 100 years? Because that's my use case. That goes back to the business model that was being discussed and what is the business model for these organizations who are issuing bids?
And is there some succession plan? Because these organizations and Beth made the point pretty clearly that the financial backing of these organizations is not always well determined. So I think that's a great question, and one that has to be resolved. I know the Donna Foundation has some scope that is approximately 50 years that they will have things available.
But, yes, how these will be persistent into the future is a question to be answered and really focuses on the business aspect of providing these services. So so we at the archive tend to be around and intend to be around in 100 years. But I'm suspicious of using pigs issued by authorities who might not be. Well, I don't know.
Well, I can only speak for the diocese and network and for the diocese. And I see and maybe also for other pids, which are under the umbrella of ISO that actually we have a kind of contract, we are acknowledged by ISO as an authority, a registration authority, and we have to comply with the series of duties, obligations.
And one of these obligations is actually to be able to deliver the data to another ANSI which could actually carry on the work. And so this obligation of course, we have some obligation regarding the security of data itself. So ISO what we had to do, I mean, if this contract is not renewed by ISO, we have to deliver all the data to ISO and then there will be actually a call for interest for other organization to actually manage the system.
I can to say that would be awkward for us because I'm not sure who wants our 10 terabytes of data and metadata. Well, I don't know. I mean, maybe that's not such AI mean, it's not only the data. It's well, we're not we're not in these figures at the icis and within the icis network.
It's more about the organization itself and, you know, just working with. No, I'm just saying. I'm just saying. Yeah it would be technically challenging for us to deliver as an example for us to deliver our data to another organization. In a manner in which that organization could maintain it in a useful form.
Well, we're lucky enough to work with many national libraries, and we work there. We have the same formats. We we work really close together, closely together. So they could actually retrieve the data and there could be a national library, could step in, for example, and start the business again. And just to be clear, the persistence here is about the ids, right?
It's not necessarily about the data. The data. No, no, no, no. But but but my main point and my main concern is, of course, the IDs will persist, but the ID will be useless. I would say useless will be of considerably less value in the future if there isn't a service that resolves it.
So this is something that if I look at how much time we spend in the daily foundation, this is probably 80% of our discussions and thought go into this question. And our approach really is to not just have a single back up solutions, but have many. So we also took the effort of becoming an ISO standard for exactly the reasons that Gail mentioned. So there is some legal obligation and there is a backstop.
We also consider the dollar Foundation a backstop, and they have the itu behind them. So that's another one. We also we try every year to realize a small. So we're a not for profit, but we realize a small surplus each year, which we put into a fund. And that helps our sustainability. But we're also very careful that there are many registration agencies and we have a very strict vetting procedure, not because we want to keep people out, but because we want people to stay and contribute and to be around for a long time.
So we look very carefully at the business models of people that come to us. And the third the other thing is the other thing we do is to commit to ourselves. So we've had a number of raise failover the years, but the DLA is that they registered have been taken over by the other IRAs. It's one of the commitments that we have. And so there's this whole sort of scheme of things that we do to try and the other one is technology independence.
So we try and make things as independent of technology. We're not web based for a reason, although people give us a hard time about that. So I mean, it's the thing that keeps us awake at. What's going to happen. We have this whole aspects to it. We could have a wonderful session just on persistence and sustainability. Actually, I think that I think that would be a great session.
Let's do it. We'll take that. Hey, Russell, write it down. Are you for good ideas? Are you representing doi.org? Yes yes, I'm the managing agent. So, I mean, you're online. You said you would not. I mean, you resolve things all the time.
Yes, quite a lot. About 427 every second. So that's several billion. Exactly quite a lot. Yeah we actually have quite, quite a lot of effort to do that. And I I'm archive. I understand your effort, but but that's precisely what I'm referring to is we trust you guys, but not everyone who comes along and creates a pid, even some of the National agencies.
You wonder if the political support will be there, you know, year after year and. Right Yeah. Actually, funnily enough, it's politically for government funded agencies that we are most careful about because governments change decisions. Change suddenly it's out of favor. I mean, it sounds weird to think. You think government funding might be something solid, but I would worry if all our funding came from that source.
If I'm really blunt. Sorry, Rob. Jenny just added a great point about Fort Campbell. Yeah, this is actually something you said, Charles. I forgot to respond to that. It is. It's the technical side of things that are tricky, too. So because there's a lot of know how.
I mean, we could certainly make the code available to people. We could put all the data, but that's only part of how we do things and why and what decisions have been made. It's quite hard to document that in a way that's to be taken over by someone completely new to it. So that there is not as simple as it sounds. That was something that's the same point that I made. If archives are going to go away tomorrow.
And we were just going to deliver the bits to somebody, that's not the same as continuing to have a functioning service. The only thing, the only answer we have that at the moment is unlikely we would go off a cliff. So if we were struggling, we'd probably know about it for several years up ahead. So we've got time to do something right. But that is that's not a huge comfort, but it's at least it's something we're looking at.
Is it possible to escrow things? But how do escrow know how? It's a good, good. This goes back to those humans again. I think the advantage is more for you, but even for us, that what we do is not that technically complex. So and I've only been at the job for seven months and I figured out how we work. But therefore, it's.
It's quite reasonable, as long as you deliver the data for someone else to implement the technical basis underneath it. But it would be bad for that to be, as you say, an unplanned event where someone suddenly gets data dumped on them and then has to figure out how to make it useful again. Thank you. Chris, do you want to you have your hand up?
We have one time for you don't have time for anything. So thank you, everyone, for coming. But if you want to stay longer so Chris can ask us questions that more than welcome to you. Chris, I was just going to make a very brief comment and that was to commend folks who weren't in the previous earlier session that Gayle actually moderated about talking about a report done by knowledge exchange group.
There's a lot of work being done on preservation and sustainability, and I think I'd strongly recommend people who are interested in that topic to watch that session and read the report. Is that the link that Adam posted in the chat? I don't know. Possibly well, this the report is called building the plane as we fly it. So I'm sure you can find it in the knowledge.
Adam? yeah, Adam posted it in the chat. Yeah, I saw it. Yeah, he's done it again. Brilliant thanks, Jason. Thank you, everyone, for the lively discussion. Obviously, this is a very important topic and it can go on forever. So I hope you will all have a great day. A friendly reminder that there is the jeopardy!
Social event tonight at 6 PM Eastern. If you are around, you're more than welcome to come. Thank you, everyone, for coming. Have a great day. Will there be food served? Sure thank you. Bye bye. Thanks for us. Thanks a lot.
Writing thank you, everyone.