>> From the Library of Congress in Washington DC. ^E00:00:03 ^B00:00:23 >> SALLY MCCALLUM: Okay good morning and thanks to all of you for coming out on this snowy morning. I guess there wasn't much else to do [laughter]. You may as well come to BIBFRAME. Okay, I'm Sally McCallum from the Library of Congress and we have I'm just going to lay out the agenda. These are the people who are going to give you the content this morning. When we finish the presentation we'll have some questions and I think there are microphones that you can come to to ask any questions, but we'll do all the presentations first. So midwinter last year I said that we wanted to take a year to experiment and I said that we would then after the experimentation, adjust the vocabulary that we're using in BIBFRAME and we would start doing some pilot implementations after we adjusted the vocabulary. So we established the testbed group. It was self-selected. If you are doing something you joined it. If you weren't doing something, you weren't going to join it because you weren't going anything. And so, and they have been actively experimenting with last year. They had their own LISTSERV that they could speak among themselves, among people who actually were doing something. But we also kept open a big general LISTSERV; it's been very active this last year, for some of you, with some of you who are in the audience today probably a lot of discussion. Now we are going to begin to model to adjust the model and adjust the vocabulary. Last week I think some of you may have seen that we floated a proposal for adjusting the white annotations in holdings are going to be included in the model. We also have a we are seriously looking at the issues of types which we had a comment last fall and some discussion on the LISTSERV about how we treated types as classes and/or properties. And we will be doing some adjustment there. There are other things also coming up the way we treat titles and some other things like that. We're also engaging RDF Expert to review and make more consistent, and set up some principles for, in fact augmenting the vocabulary. Because we see this next year, the implementations we're going to be doing as creating a need to expand. And we need principles so that it is expanded in a consistent manner. We also needed to in fact get what we had more consistent. And that's what that Expert will be doing. By summer ALA, we hope to have a pilot going and you'll hear more about that in a few minutes from Beacher Wiggins. So what is the program today? It's Beacher Wiggins from the Library of Congress will discuss his plans for a pilot with his staff at the Library of Congress. Paul Frank and Nate Trail will be giving you an early glimpse of the tools that they are working on to support the pilot. Philip Schreure will be, from Stanford will give you a review of the pilot plans for Stanford University with BIBFRAME. And finally, we are aware of work that several of the search engines have carried out to standardize product information, called schema.org. And that's the product information they receive on everything from fingernail polish to car parts. You know it includes everything. And so we, Ted Fons is going to talk about, we've been doing some joint work with OCLC. He's going to talk about their approach with using schema.org. And also the comparison we've been doing of the models that we're using, the model of Schema, and the model that OCLC is using. And we published, I think, a paper this last week. We called it a white paper about that, about that joint work. It was a high-level discussion of the differences and comparison of the model. We're going to follow it up with much more detail. In fact, they had the people who are writing it, Gene Godby from OCLC, and Ray Deneberg from the Library of Congress had much more already analyzed, but it wasn't publishable yet. And so they will be working on that and that will be coming out later. And finally, Eric Miller from Zepheira will report on his experience BIBFRAME and Schema. And that will be it, so I won't have to do any more introductions and we'll just let the program unfold. Thank you. ^M00:05:13 [ Applause ] ^M00:05:19 >> Good morning. ^E00:05:21 ^B00:05:29 >> BEACHER WIGGINS: Oh, 6 to 8 months ago we determined that we needed to do something more concrete so that the community could begin seeing concrete evidence of BIBFRAME in action. As you heard from Sally it looks like it will take a village to make this happen, but we are determined to do that. We're looking at some time in the third quarter of our fiscal year which runs from April through June to kick off our pilot. And what I have in mind is I want all of the cataloging divisions that report to me and some that don't to have staff who will participate in creating data natively using BIBFRAME. And as we see it now. There will also need to be a parallel production of cataloging in MARC format. Because we're mindful of that whenever we create at the library needs to be shared with the community throughout normal distribution channels. Cataloging distribution service. So what we envision is 25, or 25 to 30 staff that will be engaged for some period of time and I'm looking at this as phase 1 or phase 1 a of a continuing phased approach to getting into BIBFRAME. We have selected some 11 to 13 staff who have begun having in depth training in RDF, semantic Web, ontology, so that we will have a cadre of experts assistants who can work with the staff that are chosen for the pilot, to give guidance and input. Like the RDA resource description testing that we did three or four years ago before we launched RDA fully. We want to use this as a way to green information. We want to gather information that will feed into network development and MARC standard. Technical folk who will be refining vocabulary and other technical aspects of the BIBFRAME model. We want to have staff who are engaged and work in all languages. In all scripts. We also will engage some staff from our Culpeper Campus that deal with audiovisual materials so that we can see what the impact on special formats will be. We will likely engage staff from our music division so that we will have some component of music cataloging involved in this. We haven't yet determined if we will actually have a converter or some conversion program that can convert the data that are created in BIBFRAME to MARC, which is why it appears as of now, I will have the staff create their original work both in MARC and in BIBFRAME. We realize that this is dual work but for this first phase. I think that this is going to be the most productive way of assuring that we keep things on track. Obviously, there is going to be some training needed for the staff who will be engaged in producing the BIBFRAME data. So we will have Judith Cannon who leads our cooperative and instruct the program division. Our training unit to develop training plans for the staff who will be engaged. We will use this phase to see just how much training is needed. Just as we did with RDA and just as we did with RD, we'll also make whatever we come up with available freely available, to the community to use. So we'll mount that. You can pull it down; you can use it without even asking for permission. Just take it and run with it. Let us know if there are issues, and problems, or ways that we can fine-tune it. So we are using this pilot to give us some grounding in actual work in BIBFRAME. What the impact is on staff and how we move forward. In some of our discussions, there are all sorts of issues related to authorities. ^M00:10:03 How does that fit in with everything. How do we have the interaction between an existing authorities and authorities in the BIBFRAME and Link Data environment? And we are working on that. How do you even track the materials that are processed? Or how our statistics impacted the number of workflows that we need to change. So we will be looking at all of these things as part of this pilot and we hope that we will have a broad enough array of folk and materials that will be able to inform some decisions going forward. And we will keep the community apprised of the outcome and the issues that are involved in what we come up with. Who's next? >> SALLY MCCALLUM: Paul is. >> BEACHER WIGGINS: Paul is next. ^E00:10:53 ^B00:11:01 >> PAUL FRANK: Good morning everyone. Okay, who's done online training here? Participated or taught? Okay as a trainer online training is wonderful. The only thing you miss is the look of fear in someone's eyes [laughter]. So I want to tell you a story. Recently I was doing a webinar and when you can't see the eyes, the whites of their eyes, you rely on the text chat. Okay that's my cue. Things are going well. I'm getting questions, getting questions, answering questions, hopefully everything is going well. So all the sudden five or ten minutes nothing happened. So it's got to be a technical glitch right. The text isn't coming through. So I say, "Everybody okay?" Another minute passes. The answer comes. Well we were until you took that 90-degree turn about 10 minutes ago and then threw us off the truck and kept on driving. So okay what am I going to learn from this experience? I'm going to take the 90-degree turn right now before you even have a chance to know what I'm doing. And I'm being a little bit facetious, but I'm up here to talk about the BIBFRAME editor, but I really want to use this as an opportunity to talk about how we're dealing with this change at the Library of Congress as we anticipate the pilot. And a lot of it obviously has to do with training and that's my role, not only for the program for cooperative cataloging but also for the Library of Congress. When BIBFRAME has been being developed and coming along, a lot of us in the cataloging world were intimidated by the discussion on many of the discussion list the BIBFRAME list, things we heard on venues such as ALA, local conferences, all addressing really technical issues. So it's very easy to become disenfranchised and think I can't do this. But speaking for catalogers who are going to be doing the work inputting the data that is where we are dealing at LC with this possible discomfort and we're approaching BIBFRAME as something that they can do with their knowledge of MARC, with their past experience working in the editor. Now I'm purposely not even going to call the editor up and I think a lot of you have seen it. I hope so, if you haven't seen and you haven't played with it, you really will enjoy it and I encourage you to do it. But it's futile to demonstrate the editor right now because we don't have that backend; you know there's nothing behind it. So I can create a record or you know we don't want to call it a record. But I can do work in BIBFRAME and see the result of that, but unless I copy that result into a notepad document or a Word document, it's lost. So the whole power of the editor of linking is lost right now. So Nate's going to talk about that, the backend and what's going to happen with authorities, but I just want to talk a little bit more about how I'm working more as a well I call it a sympathetic listener maybe even a therapist. To help catalogers who really I mean and I felt this strongly all along that this cataloging knowledge is what's going to drive BIBFRAME forward. So we as catalogers need to be a part of this. And that's why when we are looking at training at the Library of Congress for this pilot, Beacher mentioned we have a group of 11, 12 people that got the very in depth training. I'm included in that group and I'm struggling with it but learning. It's Semantic Web, RDFA, Ontologies, Link Data, SPARQL, XML, Turtle. All the things that I was really afraid of for years. I'm now feeling a little bit more comfortable with, with an I to assisting the members of the pilot LC to get through this. Now what we purposely are not going to do with the testers in the pilot is teach them all of this. They don't need to know it. And I've said that over and over. You work with the editor. You don't need to care about what comes out. You just need to know how to interact with that editor. And Nate will also talk about profiles. We'll have a nice profile for the pilot at LC. We will have solved the decision of where the authority work is going to be done. You know whether it can be done in the editor or whether our catalogers will just go back to our Voyager ILS to do their authority work all these issues will be ironed out. But it's the area of understanding the labels on the editor and that extends obviously to understanding a little bit more about vocabularies. And I don't see this as being a problem at all with catalogers because they love controlled vocabularies. This is just another set of controlled vocabularies. So the learning curve is not very steep. We have to get over that. We have to understand that is just a pilot, it's going to be no different from RDA. Fortunately the RDA test is still fresh enough in our mind that we understand sort of the methodology of a pilot. Not necessarily the same subject matter, but we will get through this pilot. We will assess how things look. Beacher said that we will be doing duplicate work. Which is a wonderful business model. So you can compare you know what takes more time. Where so things change and things will be able to change you know there are modifications going on based on the outcomes of this pilot, to increase the efficiency when we finally move into full production in BIBFRAME. So that's really all I wanted to talk about and I encourage you to take this back to your institutions. This role of just assisting, of alleviating the discomfort that I know is in there. I know it's in catalogers in your institutions and let them know that their role here is much greater than they may realize. Thank you. ^M00:17:51 [ Applause ] ^E00:17:57 ^B00:18:03 >> NATE TRAIL: It's not quite the same title as you have in your table of contents, but that's okay. So we are having pilot and all problems will be solved before the end of May [laughter]. >> BEACHER WIGGINS: I'm counting on it. >> NATE TRAIL: But we're also in talks with some people at LD4L the because we hope to share some kind of either common code or infrastructure or best practices or things like that, whatever we can. So we're not going to be going it alone at the Library of Congress. There are a number of tools that we have started or are in development and this is a list of most of them. The BIBFRAME editor is there, but it's not polished and finished and will require ongoing maintenance. The profile editor we have contract out and it is now in acceptance testing so that's coming out soon. We also have a contract for search and display system. I'm going to just go through each one individually as well. So here's a picture of the editor in case you haven't seen it. It's the things on the left-hand side represent profiles of templates of elements that go into each those kind of things. So you can set it up based on your favorite type of material or foreign language or you know whatever flavor you want. The profiles are documented on the library's website and we have a profile editor built to that spec. A JSON document can look very intimidating just like showing raw XML. And if you wanted to change how the BIBFRAME editor works, you would have to tinker with one of these things by hand right now. So the contract is to turn that into something a little bit more friendly. So here's a peek at what they've been doing at Smart Software Solutions. ^M00:20:00 It's very flexible. You can choose to add or include a element from whatever languages you have loaded whatever vocabularies, I'm sorry, you have loaded into your system. So this particular system has the MADSRDF and BIBFRAME vocabularies available to it. You can also in a profile change it to your local property. So if there is not a BIBFRAME property or MADSRDF, or remise, or whatever your favorite vocabulary is, you can invent your own and store elements in that, data in those elements. If you're once you've selected a particular property, you then have the ability to tell whether you want it to be repeatable or have a default value. In this case I've chosen to put the default language for this template into [inaudible] just for fun. Two more tools that we have had to do with SRU which is of course the 3950 on the web. Meta-proxy has been updated by index data in the first instance so that any ILS system that has meta-proxy can now have their data returned as BIBFRAME. And once we get data stores of BIBFRAME dated they're working on the next contract which is to do SRU over a triple store. And they have it working but we're working out a lot more details about how you decide whether if you ask for BIBFRAME results. Do you want works only, do you want instances do you want only names, etcetera. And another issue that comes up is when you get results that since we're linking so much in Link Data. You have to decide on the result set how deep do you want to go in your links. Your data store may be so deep but it also reaches out to other systems. So you might want the system to actually go out to those places and retrieve back labels for those values or for those URIs, that we're still working through a lot of those kind of details. The one more important contract that is currently underway is the search and display system. And that's being developed right now. They have two components. One is the data store component and one is the catalog component so that's search and retrieve. They've separated them out purposely so that you can have your own data store, or you can have your own web interface. And it's using Fedora 4 which is now able to store and manipulate triples which is really good. And they have two different ways of searching because it's kind of hard to search RDF for strings. So they have Elasticsearch for keyword kind of searching and then Fuseki for their SPARQL engine. They don't have a frontend yes so I'm just going to show you a few things. Here's their search page right now it's very rudimentary, but you get results back and it tells you the kinds of nodes, like is it monograph, is it text, or is there just a title entity. And then this would be the raw information they had to build a display and we haven't seen anything about their display. They're at the point of designing the frontend. And of course of the Library of Congress on the BIBFRAME site. We are continuing to develop our transformation from MARC to BIBFRAME and one of the things that I've changed is to make the comparison tool show it to you side-by-side. It used to be that you that you would have to tab between one of the other which I found frustrating so I changed it. The other big change in the transformation tool is that now you can send in holdings as well. So any MARC XML collection of as holding interlinked or interleaved or at the end, it matches up the BIBS with the holding as it sends them in. So I think that's going to be very powerful for you to be to take data out of your system and push it into the transformation tool. You've been able to do more than one at a time for a while now, but it's only been accepting bib records. And that's all I have for today. One actually there is a tool that of a set of tools that I have not mentioned. And that is when you transform your data into BIBFRAME you get a pile of works and instances and authorities. But there's no real relationship to the rest of the data. So what we need to do is we need to have our bibs have a tool that will let you ingest it into a triple store and match up the things. So that if you had a whole bunch of 7xx, or 130 cross-references they go find that information that's already in your database and merge the results together. So you end up with works with many more subject headings and many more links to instances. And that's a piece that we're still developing ourselves in the library. Congress think you very much. ^M00:25:35 [ Applause ] ^E00:25:41 ^B00:25:55 >> PHILIP SCHREUR: So I before I bring up my presentation. I have to admit, after Sally asked me to talk and I did have a martini. I changed the title of my talk. From what I originally proposed. It's wrong one. Oh yes, it's so the new title is "Le Boeuf Sur le Toit" a surrealistic exploration of BIBFRAME for production. So while this may mean more to music people that it does to others. But it will make sense in the end. So in the 1920s, a cultural movement began called Surrealism whose goal was to resolve the previously contradictory conditions of dream and reality. So as we approach the surrealism's one hundredth anniversary, this goal seems particularly aft as we move strive to move, strive to move from the dream of linked open data to the reality of production. In music, this movement had its parallel in the compositions of Les Six or the six. One of those composers in that group was Darius Mihaud who was the composer of "Le Boeuf Sur le Toit," or "The Cow on the Roof." So as we explore Surrealism in the context of technical services production I have pulled together another group of intrepid six institutions to explore this, together with Stanford. So the group that will be looking at this over the next year are Stanford, Cornell, Harvard, Princeton, Columbia, and we'll also be working with the Library of Congress and their pilot project as well. So the group met for the first time yesterday. So these are just sort of very preliminary thoughts about the directions that the group are going. I would have to say that the only fixed goal of the group is to move towards production in some way, whatever that means, over the period of the pilot. So I think we tried to leave it very open, we're not talking about converting all of our production. I have to be very careful in saying that because in the past I realize I gave that wrong impression. It may be as little as 10 or 100 items that we will be doing as BIBFRAME, but it will be actual production work and not just a test. I think one of the strengths of this particular project is that it is six different institutions and we have very different workload locally. So we will be working through it all individually at an institutional level, and I think it will give a great variety then of approaches to the problem and so then a variety of solutions as well. Of course, as we get together there's going to be areas of commonality which we're all going need to be able to resolve. And I think those will be the most interesting to the rest of you here. So what, I'd like to mention a few of those sort of the more common issues that we will be dealing with. First is a cloud, a shared cloud space for us to work in. So there's going to be a lot of shared materials and shared just an area for us to go do our own work and that's going to be very important for us. We're going to have to think about shared resource files. So not just authorities but we need to think about the resource files for BIBFRAME works and for BIBFRAME instances. We're also going to need to think about those authorities though, and what do authorities mean in this new context is it the same kind of authority record that we've been dealing with in the past in our NACO. Is it a shift to purely an identifier, or is there some sort of new definition that we need to come up with for what an authority should be in this type of contact. And the last will be some sort of shared tool development for the common set of tools that we will all be working with. ^M00:30:05 Now I would feel badly without giving sort of a shout out to another project going just north of us in the Bay area at UC Davis called the BIBFLOW. Which I think is really going to be a perfect complement to this project. So the differences will be UC Davis is focused on BIBFLOW, so the analysis of work flows and how this could work at a single institution, as opposed to a multi-institutional project. I think they're also focused on the tools that are being developed by Zepheira, and we are interested in the tools being developed of Library of Congress. But I think the point is that it is a wonderful, it is much better to have multiple tools to be able to work with than just a single set of tools. So again, I think that is a very strong part of this. And also UC Davis is looking at some sort of integration with OLE as the ILS to do a lot of the work. So again they're working through this in the context of a particular ILS and we'll be dealing with that in the context of multiple ILS's. I'm not sure how many all of them that we deal with the I'm sure it's at least two or three. So what is Stanford's pathway in this project? So for ourselves. We are thinking about sort of three end-to-end workflows as far as this project goes. So the first one we'll be looking at is sort of the copy-cataloging paradigm. So in our context this will be starting from the point of view of vendor-supplied copy. So what do we need to do starting with vendor-supplied copy coming in acquisition to be able to spit out BIBFRAME at the other end. So as we think about that particular workflow, there's a couple of big issues. We know we're going to have to deal with. The first one is going to be the conversion of MARC to BIBFRAME. But then, what we consider after that, perhaps, which is more difficult is that we know because MARC is so inarticulate. The BIBFRAME will reflect that as well. So to get sort of more quality BIBFRAME, some sort of hand editing is needed to be done. So we would like to be able to define what is that editing that we feel needs to be done and also who is going to do it. As this work used to go through our acquisitions department, they are totally streamlined and they don't do any editing whatsoever. So trying to figure out who is going to do this work is going to be a big issue for us. And another part is going to be getting those authorities right now we automated authority control we send the MARC records out to our vendor and it all comes back. So in this sort of situation how we're going to get that data back from our authority vendor. The second end-to-end pathway we're looking at will be original cataloging. So again, that is a little bit different and that will not be starting from MARC record will be cataloging natively in BIBFRAME. One of the issues we're trying to decide is again like LC do we need a BIBFRAME to Mark converter. Would be like to be of the have that MARC record as some sort of fallback. Could it be useful in our ILS to fulfill some of the other obligations we have? So again, that will be a big issue that we have to decide. And again, I think as Nate mentioned authority work, you know right now, we create a lot of new authorities in our original cataloging. It is something we're going to have to do to the editor. Is there going to be some other model we need to follow. And last, we are sort of analyzing the switch to FAST instead of LCSH for our subject access. Just because of the ability for FAST to work so well in a Link Data environment and currently LCSH not to. The last project which we are considering is a pathway focused on a particular domain, and in this case we're thinking about doing a project with metadata for reported performed music. And we've got a lot of music expertise at Stanford, which makes it attractive, but I think there's also a lot of room for these sort of domain-focused projects, so trying to work with Music Library Association as the group most interested music. Working with the PCC to develop a profile for music that others can follow as well. But just to be able to deal with music from end to end thinking about the model, the profile, the planning out the conversion of all the retrospective data we have that we'd like to convert BIBFRAME. But then also think about discovery of that music material and perhaps in a black light context. To be to take care of all the new innovations, or discovery that BIBFRAME would bring. So again, those are the three project initially after one day we are thinking about pursuing. ^M00:34:50 [ Applause ] ^E00:34:56 ^B00:35:15 >> TED FONS: Good morning. Welcome to the world's first professional conference held inside a snow dome [laughter]. It's really amazing to look at those giant windows out there and see the snowfall. We're at the pretty part of it now when everybody tries to go home tomorrow it's going to be the miserable part, so. But welcome. And thank you to Sally and Beecher for inviting me today to talk about what OCLC is doing in this area and the collaboration between OCLC and the Library of Congress in this area. I'll try to stay to about 10 minutes and then if Eric does as well, we should have some time for questions for the panel I think. So let me start just talking a little bit about strategy. And what OCLC is thinking about in this area. And those of you who have read Gene and Ray's paper already have a glimpse of this. So, you know what we've been thinking about you know as we focus on you know the value, the cooperative value of that data set that is well WorldCat, and the services around it. You know we're thinking about first what is of interest to the web. So in our data modeling, thinking you know kind of starting from up there. And saying, first what's of interest to the web. How do we make sure that library data is understandable by the folks on the web, which has always been a thread of these conversations about replacing MARC and kind of moving to the next generation of library data management. It is, you know, recognizing that these existing formats are not, don't play well on the web are friendly to the web. So really focusing on that. And starting from there. Starting from up there saying. Okay what do we need to do and then make the data available in structures familiar to the web and that means certain things. And then improving library workflows. So let me say a little bit more about what each one of these things mean. So you know first focus on data elements for web discovery and what is really interesting there is all of the folks we have studying that say that actually the web standards do pretty well. They capture you know a lot in what's required to make statements to make disclosure about what's available in the library. So that's really good. That doesn't mean it captures all of the level of granularity and detail that we have that's important for our inventory management systems and kind of you know historical tracking of things. So there's a difference. But actually the web does very well there. And making them available in structures familiar to the web frankly means start with Schema.org. That is today's you know statement from the web of what's important. So I'll be you know, sort of clear and straightforward on that. And then managing entities. So this is one of the things we also learn from you know studying how things are evolving on the web, and how data is manage, and what data science you know, and modern computer science, and graph theory tells us about managing data. It says you know you have to find these entities in your data and express those things because that's how people interact now with knowledge. And that's how you create pathways to exposing your materials. The things that you have that represent that knowledge. And so that's really important and something that we've been really trying hard to remind ourselves of an OCLC as we proceed and kind of make decisions about where we put our focus. And then you know improve library workflows, you know, just to start is this idea, you know, improving discovery, which really means making discovery more weblike. The experience of discovering more weblike and the technology of discovery. So you saw few years ago we did this exposure Schema.org and WorldCat.org to, you know so we could tell the web what's in libraries. And reinvent cataloging comes after that. Which is okay, now that we know what the web requires. And now that we know where the differences are what should we do to fix that. How should we focus the activity of this community of metadata managers, which is librarians? And so these are kind of the focus, or the things that guide us at OCLC as we make decisions and think about what's next and what we're doing. ^M00:40:08 It's all in this context, and we really are grateful and rely on our OCLC research and folks like Lorcan Dempsey, who are you know helping kind of craft the strategy around this and help us make decisions. So I do want to talk about the white paper. We are again grateful to Gene Godby. Gene is a very modest person, but I'm going to force her to raise her hand. There she is. Thanks Gene. The genus here is not really the last 10 weeks probably more than that probably more like 20 weeks with Ray Deneberg at the Library of Congress. I'm not sure how well the two of them enjoyed the holiday break. So Merry Christmas Gene for getting the work done. But they really worked very hard on the in their approach was very interesting. They said okay will take a few cases. Let's take some data expressed in RDF and let's see how we're doing when we start from MARC and describe that in BIBFRAME, and start from Schema. And see how these two things meet. And it's very well described. It is not a long paper. I keep hearing people say, "Oh yes, I'm going to read it. I brought it with me." Or, "I printed it out." I just say, "You know, and the time it took you to explain that he could read half of it." So, but I'm not out shouldn't be scolding people about doing their homework. But anyway, I really encourage you to it's not long and encourage your colleagues as well. And it's you know written in a very approachable style. As Sally mentioned, there's actually a lot more data supporting this. When we were talking between Library of Congress and OCLC about how to make this work available, we made a decision to kind of break it up. To put out this, which is a nice summary of the approach and the main findings really, and then later we'll make available the rest of the data which we will do either as a big document, or a series of documents. We haven't really gotten there yet. But it won't be long because they've done the work already. So very useful work to sort of understand okay so we have this one approach that kind of starts from the web. And this other approach that starts from MARC. And you know do they meet in the middle. And the statement in the document that stands out for me, which is I think very useful is a very high-level conclusion page 8. The alignment is still accurate and even more defensible. Talking about the alignment between these two approaches. Then in 2013 when we first started looking at these things in Gene's original paper because the primary BIBFRAME concepts are now more consistent with the corresponding concepts defined in OCLC Schema model. So what all that means is that from a technical perspective, you know these people who understand this very well, see where things are already aligning and they're also observing that there's movement, and there's evolution and I think Phil talked about the benefits of multiple tools in this environment, which I think was an excellent observation. I think the same is true for you know different approaches to you know to this problem of modernizing library metadata management and exposing it on the web. The benefits of multiple approaches I think are more, a richer, you know, end result I think. And so I think we're seeing that. And I think this research is validating that. So there's a lot more in the paper. There's some very important statements of where the similarities are and where the key differences are. And some of them are technical. Some of them may become very important as we go forward and we'll have to make decisions. We, meaning you know the smart people at the OCLC. The smart people at the Library of Congress, you know with sort of input from the community as well. So there's important detail in the paper about similarities and differences. And some nice recommendations for you know next steps and where we should go. So it's a very well done document, I encourage you to look at it. So let me tell you little bit more about what OCLC's doing. Some of you have heard me speak before have seen this. Kind of this idea of the library knowledge graph, you know, putting our library data together to make it available for the web. You know well what's going to be better Fons you know when we have this graph thingy. We think lots of things will be better. So we think that you know that discovery will be better if we can make it more weblike. We can make the data available so it can be integrated with the web. We think cataloging will be better. But mostly different I think about cataloging and I'll say a little bit more about that in a second. Interlibrary loan analytics. My interlibrary loan colleagues say well all your web stuff sounds interesting, Ted but really when it comes down to the lending workflow, it's about picking things. You know this thing or that thing to match up what the borrower wants. And so they want, what they want to see is that we make that process easier and more informed for the people in that process. So that means exposing in a more useful way, the details of manifestations, and items, and those kinds of things. So they can better match up a request you now from a borrower to a lender. So we think lots of things. So I'll talk just a moment about progress. So I talked about entities creating entities we are going to very much focus here on the data assets, making them suitable for the web, making them available. That means creating these entities. So we have these work entities. We've got just short of 200 million of those out of WorldCat, and that's before we go to the article data. So wait until that happens, and these numbers will get bigger, and also we'll see much more connections. So you know some people write books and articles and some people just write lots of articles. So it will be nice to bring all of the into this the same ecosystem and you can see here all the audio formats and all that. Person entity is our next one. We're building that data now to be released soon. This is the you know, taking bibliographic data and our authority data and making statements about people. And then saying okay world web these are the persons that have a role in creative work creation. You know, these are authors and these are translators, and performers, and all those things, editors. And here's all the references to other identifiers. So all the link in Link Data, it's gathering all the things together. Improving discovery. So you know we see the possibility of knowledge cards and workspace discovery to solve real problems in discovery. And you remember I said make discovery more weblike that experience. So this data lets us do that. Just the last couple things here. Can we measure impact? We can. My sort of lightning bolt here is when we released the works data and immediately we saw our visits, you know, start to increase and you know traffic start to increase on the things that we exposed to the web. So something's working there. And let me just say a couple things as I finish about our approach and timelines. You know ongoing data modeling all these things measure impacts and repeat a kind of an agile process and then you know these activities around the services that are built on top of WorldCat. You know works and persons, but then discovery and changing cataloging. So our cataloging team now is working on you know person editing. You know, correcting things and adding new identifiers for persons you know that sort of cataloging of the future. And then of course we can't, we heard all of Phil's discussion about creation of data in these new formats. We need to recognize that make and sure that that becomes part of our kind of import export ecosystem. And really you know all of this is kind of building this knowledge graph. And then there's a whole activity of connecting systems. We think there will be a huge role for identifiers. I spent a lot of time with people talking about even today, about OCLC numbers and the quality of those OCLC numbers in our traditional bibliographic data and the incredible efficiencies that has brought having those numbers in lots and lots of systems it can't be overstated I think. And we'll see the same thing as we start to coin identifiers. And one of the hardest tasks for my team back in Columbus is trying to build these entity data so they have persistent identifiers. So you'll always get back the response with that same number. So we want those to be in all the system so we can bring things together. Last one is what can you do. Keep cataloging. I mean that's really what it's about, you know, so if you're wondering, well we don't have this or that. You know it's okay, because all the data you add to WorldCat today in the traditional formats is what we're using to mine you know from WorldCat. So you know we're on the, you know sort of focusing on books and a lot of digitized content now. We're going to add articles as well. So we're using all the traditional stuff to create. ^M00:50:02 And now the next step we will be to change cataloging so it's working on the new stuff. Thank you. I think Eric is next. ^M00:50:11 [ Applause ] ^E00:50:16 ^B00:50:21 >> ERIC MILLER: Good morning everyone. This is one of the rare occasions where actually being outside is not as nice as being inside. I am going to talk a little bit about a wide range of additional experimentation, practical exploration, and the sort of continuum between descriptive standards and discovery standards. Particularly in the context of how those can come together to accelerate what I'll describe as the visible web. And, more specifically the visible library in the context of the visible web. So I'm going to start with a little bit of perspective over the past year. We have been fortunate at Zepheira to work in the past on a lot of the enabling standards that in fact we're discussing here. If you would've asked me 10 years ago if I would've been speaking to a library community about Link Data and RDF and the power of these technologies, I probably would have said no. But what's been actually really exciting in the past 10 years is how quickly the library community has been seeing the opportunity of leveraging its collaborative potential around these open standards. And starting to use this to accelerate a lot of the cooperative, collaborative data sharing patterns that are at the heart of this institution and the heart of this community. We've been fortunate to work as technical and strategic advisors to folks here on the panel. Both OCLC in terms of helping them move into the Schema.org strategy and the Library of Congress in terms of starting to expose the national regional public library needs in terms of the web architecture. We've been focused a lot though on large-scale implementation of how to basically make this happen in practice. Including working with various national libraries in terms of managing, not just their contact but starting to expose authority services to connect their data with their consumers. As Phil had mentioned also in the backend systems with projects like BIBFLOW. How you reinvent ILS and how do you start to think about in terms of more of a web context. And all these have been part in parcel to kind of coming together under a new initiative that we are describing as Libhub. Which is about taking a lot of the assets that we have been managing that are hidden behind the web and leveraging a lot of these to new technologies to start to really start to explore the visible implications of what that means in terms of surfacing them. Using these particular standards, but in a very practical and pragmatic way. How do we expose this data in a way without completely undermining our existing systems, completely changing existing infrastructure, and dealing with the reality of change. Which is very challenging one. In the past year we've been asked to step up and try to help on the education, the practical training, the practical understanding, the practical implication of these technologies. We've had the opportunity of now having about 300 individuals from you know literally dozens of institutions go through this sort of training program. From national libraries, to public libraries, to rare book and management, to music. Look at BIBFRAME from their perspective; help shape the recipes and transformations to expose their flavor of MARC into a larger network of information and feedback that information not just within their local community, but across communities as well. Further, we've had the opportunity and privilege of talking to a lot of different directors of various libraries and grappling with their real-world problems are in terms of this change. And I just briefly if I was to summarize some of the shared observation between what we've heard and our experience in terms of implementation in four quick points. The library is far more than its collection. We talk a lot about MARC and our assets and things that. But the notion of a library as a place, its value it serves to the community, its facilities, its knowledge, its capabilities. That notion of what a library is is more than its collection. Second, I'm here hearing a lot about Link Data but frankly it's been more about the data than the linking. If we want to understand the way the web speaks, frankly it's not a vocabulary it's linking. And that's a key part of what I want to sort of focus on in the next few minutes. That linking is what we have in spades. And it is the power that we have in that linking to make the assets that we have far more visible than they've ever been to date. So Link Data, I'm very excited about that. But you'll hear me focusing a lot more on the linking aspect of that then the data aspect of that. Finally, from someone who's been working on the web of data for a tremendous amount of time. I'm extremely excited to say I believe it is here now. But we in the library community are not taking advantage of it to its full potential. And the opportunities of doing that as we're discussing here are huge. But trying to basically think about very practical, very scalable, very approachable ways of moving from point A to point B is a challenge. And I'd like to basically talk about that in very concrete terms. So when I talked about the visible library and some of the projects that we've been doing and the projects we're about to do in the context of Libhub. I think Chuck Gibson in my opinion has sort of summarized it best. He's one of the directors and CEOs of one of the several libraries that are part of this initiative. But I really like the way he summarized it which is, "When my community searches the web for something we have, we better show up as an option." I mean this is in essence the basis for what we're talking about. So being able to search you know inside a community for something that a library has that result that that library has it better show up as an option. We're not showing beating Amazon or you know basically necessarily sort of dominating the front page of any search engines, but just showing up as an option is a key part of this. So the notion of taking our assets, taking our capabilities, taking the fact that we're a library, and showing up on the web is an option. Right now it's easier to find coffee shops and banks on the web, by an order of magnitude, than libraries, right. Being able to basically put ourselves on the map, being able to put our assets on the map, being able to basically be part of these results is what we talk about in terms of a visible library. And the reason for that is as we've discussed here is because in fact, we're working in interchange formats and exchange formats that basically the web doesn't understand. I mean we should be very clear about the power that we've done. I mean 45 years of interchange. We've been exchanging information you now two decades longer than the web has existed. But the notion that we haven't been able to sort of grab that and expose than in the way the web understands. We have created these barriers in which the content and wealth of our credible assets simply just aren't assessable to be people that are finding it. And I don't mean just searching Google. I mean people finding it the web. Being able to easily bookmark something and share it with a friend. Being able to find it on Foursquare or Yelp. You know, being able to find it where our users are looking for it, how they are looking for it. That thing that binds all of these applications and these search engines and the social networking systems in place is the web. Google's not the web. Facebook's not the web. The web is the thing that basically makes all of those applications possible. But we are not surfacing it in a way that those understand. BIBFRAME. So everyone's seen this slide. I like to think of this as MARCatecture. Nothing, really? MARCatecture, nothing. All right, I thought, all right, MARCatecture. It's a high-level overview of a very powerful model. It's a way of defining control points that the web understands. In a simple, replicable, Link Data pattern. And this as a model is a very simplified way, but it in fact can be extended and replicated based on very consistent patterns, to make the simple things simple and complex things possible. But in a way the web understands. Again, not necessarily because it's a vocabulary because of the linking. So we talk about identifiers and that's an important part of it. But the thing that makes it even more important is the relationships that exist among those identifiers. The things that link these identifiers together. And one of the things that I like about BIBFRAME is it is a very simple but replicable in a very consistent way model that allows that to be possible. So in the course of the past year working with a lot of different groups focused on practical understanding of Link Data, BIBFRAME in the context of libraries. This is one of the many tools that we offer in the course. ^M01:00:02 But the notion is ability to sort of submit your own collection of information and understand the implications of how it gets transformed and basically what is mapped to what. And helping people in libraries sort of understand the transition between, in essence, a document oriented to a data oriented migration. So submitting roughly in the upper right-hand corner 200 records, realizing that we're materializing close to 1500 resources. Breaking those resources down by type as at a certain level, Ted was showing. Being able to take those resources and break them out into people, places, organizations, topics, etcetera, etcetera. MARC has all of this amazingly rich information. And helping people understand their data in the context of this transformation has been one of the very helpful tools in helping these people going through this process to understand this. One of the additional products and services that we've just recently launched our around link assessment, helping libraries become more Link Data ready. Working with the vendors, understanding their collections better, understanding the conductivity. These are examples of how connected their data is. How overlapping and linking their data is. How connected people, and places, and organizations, and topics specific to their data is. But the takeaways of all this is that moving from our traditional record to a web resource environment using Link Data and BIBFRAME, we're increasing the number of ways in which people can get access to our information. We're recognizing and identifying how highly connected our assets are. We're seeing the opportunity. In fact, this community loves its identifiers so much we have more authority services in this community than any other community that we've been working with in the past. We can, not only link among ourselves more effective, we can link out to authority services because we like to uniquely identify people, places, organizations, etcetera. And now with BIBFRAME, we have a vocabulary for how to express those resources, express the linking, and more specifically, based on the underlying technologies that BIBFRAME is based on the world web contortions RDF, a language in which the web understands which is all about linking. Okay so rolling back when I talk about what the web understands, before vocabularies, before search engines, before all of this stuff. It is about the power of links. It is the power of identifiers that link to other identifiers. So this is in essence the complex algorithms that exist under the major search engines generally referred to as PageRank. But the notion of a link is worth 1000 words. The more things link to each other the more happy they are. The more things link to each other, the higher they basically show up in terms of general search results. So the value of something on the web is proportionate to the number of things that are linked into it. Not out from it. So the more we can link across each other. And the more we can connect to each other, the more valuable we're perceived in terms of the web and the more we're visible to those basically using search engines. This is an important realization that there's a community and a continuum here between what we can traditionally think of as description and what we can think of as discovery. And if we start combining and crossing these streams, we start to get ourselves into trouble. And I say this at a certain level as one of the a sum of life one of the founders of the Dublin Core Metadata Initiative. Usually I get a chuckle over that. We cross these streams and basically started to become unclear in terms of are we focusing on description or are we focusing on discovery. We have a tremendous amount of resource description standards. MARC, EAD, VRA, CIDOC, [inaudible]. Some of these we're familiar with in the library community. Some of these are nearby in the archival or museum communities. But the ability of basically taking and focusing on common vocabularies that link things together from a description standpoint, that are based on web principles allows us to project these into a wide range of discovery systems. I want to make a very clear point. These are moving targets. The Schema.org now is different than the Schema.org that's going to be next week or next month or next year. But being able to be very clear on how we have concepts and carriers, which are works and instances, how they relate, how these connect to authorities, how these connect to places. Allows us a very rich inner linked descriptive framework to project into any number of discovery opportunities both now and moving forward. So Schema.org is fine but Facebook is gearing up for additional ways of doing this. Being able to project into the open graph and moving forward in these is another important part of making our data visible. You know, there are other opportunities that are coming online. The web is a constantly evolving thing. And presenting this making sure we have our descriptive standards in place so that we can project any number of visible perspectives is an important one. That notion of visibility at a certain level starts with an agreement. It starts with folks coming together and saying, "Look I want to try something. I want to experiment. I want to see the implications in a very practical way. And I believe that everyone benefits from the visibility of libraries and their contents on the web. This is the pledge of the Libhub initiative. As individuals we can raise our hands and say this. As libraries we can raise our hands and be part of it. This is sort of the icon slide, but of 11 libraries around the United States and Canada that have come online in said, "Look I want to be part of this. I have my collection. It is invisible to the web. BIBFRAME is the replacement of MARC. It's the basis for speaking the web of the language. I want to be able to project my data into BIBFRAME; I want to then project that into the way that the web understands. Which is a combination of RDFA marked up in vocabularies such as schema.org. And I want to connect these together so that when somebody searches," back to Chuck Gibson's point, "for something I have my library shows up as a result. Not someone else's service, not someone else's tool, not someone else's product. But my library." And trying to put libraries back in the mindset of those searching as another alternative. In addition to Amazon or anyone else that they can go in get the assets that are looking for. As I mentioned earlier libraries being far more than their collections, description and capabilities, systems and services, content and collections all can be represented in this model. And each library is its own network. And this library as being part of this network of libraries basically connects to each other. And the more that they connect among themselves within themselves, within themselves and then among themselves, the more that link the acceleration occurs in the more we have a way in which we can start to expose our local and shared assets in a way that a wide range of different consumers of this data can benefit from. The summer of 2015 by next ALA, we will introduce the Libhub network. Which is based on this data network designed to surface each of these libraries data in BIBFRAME, projected into schema.org, accelerated by linking. Each library will be its own set of data, optimized, but then linked across each other. And what we're going to do is experiment. What we're going to do is test. What we're going to do is see the needle change on visibility. And we're going to share the lessons learned in terms of moving from MARC to BIBFRAME, what's needed, what's not. What worked, what didn't. And basically start to explore new ways in which folks can basically use BIBFRAME to find and organize and make the assets they have available, including the capacities, capabilities, services, and assets they contain. And we're doing this in a way that is incremental. In shows value along each of the particular paths. So if we're looking to replace a particular standard and we're looking at it from a feature comparison of A to B so that every one of these things, you know from the old to the new are in place. We're not going to be happy until the very end. On the other hand, if we on the bottom, we're looking to basically you know move from a point. You know, certainly moving from a skateboard, to a bicycle, to a motorcycle, to a car. Every step of the way the notion is we want to see value. And that incremental approach based on a really powerful foundation we have in place is what we believe is a way of basically making this happen. And the way of weaving the best of libraries and the web together. This isn't about gaming any system. This isn't about optimizing to anyone search engine. This is about speaking the way the web understands and benefiting from the collaborative power that it is this community. We in the library community are more collaborative than any other market in any other vertical that I've had the fortune of working in. We're not the most efficient in it though. We're not the most effective in it. So part of this is basically learning about how to accelerate those links and how to accelerate you know that cooperation. ^M01:10:01 And benefit from weaving the best of what the libraries have to offer and the credible assets that we have to the very power of the web. That's what Libhub is about. That's what our focus in terms of practical exploration of BIBFRAME utilization in the web are focused on and that's where we see this kind of continuum between helping libraries move into the very powerful and important aspects of BIBFRAME as a description standard, but then leveraging that potential in terms of the discovery capabilities. The web has to offer. Thank you very much. ^M01:10:36 [ Applause ] ^M01:10:42 >> Okay, now we are open for questions to any of the panelist presenters. Everyone is happy and content [laughter]? >> I'm happy and content. What is the plan for when the Library of Congress may go to creating works in, or creating cataloging in BIBFRAME without a parallel market for supporting those libraries aren't yet ready to consume BIBFRAME? >> BEACHER WIGGINS: Somewhere down the road. Part of why I've been stressing is we are very mindful on the legacy MARC data that exists. We have MARC data that we need to deal with. So this first phase, first pilot is to get us started down that path for us to put our feet in the water, so to speak in terms of creating the data for the materials that we received in BIBFRAME. Until we know we have a sustainable conversion application from BIBFRAME to MARC, we will do the necessary and create MARC data even as we create BIBFRAME data. This is going to be an evolutionary process depending on whom you talk to they'll give you different timelines. But it's not overnight. It's not this year. It's not next year. And as you've heard from all of the presenters, there are various, excuse me, pathways that are being taken to make that happen. Do our technical colleagues have more to say to this? Okay. Anybody else? >> I have, as long as Beacher's got the mic. I have another may be unanswerable timeline question, but you're talking about the BIBFRAME pilot at the library of Congress. Are there any plans, or any foreseen timeline, for national testing efforts? Or is LC going to coordinate any national testing efforts for like you did for RDA, or national training efforts like a CCTP? >> BEACHER WIGGINS: Well there's a combination of things going. On what Phil was describing and he can weigh in further on that, is to my mind a collaborative approach to having real pilots and tests. Until we at LC have carried out a pilot and have enough plans, structure in place that we can support the internal test, then we could think about something that is more general and that is more national. That's one of the reasons I was stressing that whatever we create, whatever our training plan is we will mount, it will be freely available, you will know what we're doing. And those who are in a position to move along with us in parallel can. If we do some sort of coordinated plans for national test pilot that I don't see happening before next year just to give my best guess. >> So we can play with other toys as well. >> BEACHER WIGGINS: Play with other toys. Yes. Because of the variety of the presenters this morning, we are hoping we are stressing that there is no one way to proceed to get libraries into the linked data world. Phil do you have any? >> PHILIP SCHREUR: So I guess I would just quickly add I think one of the things we are focused on on this project with five academic libraries was the realization that we all have different systems. Different ILS's, different workflows. So I think trying to work through it from a variety of different perspectives means that whatever we work through, you know what Stanford does may not be applicable to Davis, but it could be applicable to UCLA or some other people. So we're sort of developing multiple pathways to it and hopefully others will be able to fit into one of those pathways that we will be working on. >> There are over 11,000 small towns across the country. Communities that do not have public libraries. Am I to assume that this also will benefit them and will point to them or will it also just only point back to a physical space? >> BEACHER WIGGINS: Do you want? >> TED FONS: I guess. >> BEACHER WIGGINS: You two tackle that one. >> TED FONS: Yeah certainly I can start. I mean at the end of the day, you know, we rely on metadata you know to represent what is available. You know things that have been published and then we rely on some statements that say you know this organization or person or you know has something to offer you know and that's the essence. I have something to offer. I can lend it. I can make it available. So at some point for you know for those things to be available for humans. Somebody has to record that somewhere. I know this is a very abstract answer. By at some point somebody has to there is no, Link Data is a lot of things, but it is not mental telepathy, or you know ESP or something. You know, so at some point somebody has to record that an offer is being made by an organization. So I think like we see in the cooperative cataloging environment and the authorities environment where people say, "I'm willing to record things on behalf of somebody." I think that's really where the hope is or the need is I think for very small libraries is that somebody is willing to record those things on behalf of somebody and then make them available to the web. >> So it could be the size of a chicken coop. >> TED FONS: Yeah, so the comment was it could be the size of a chicken coop. Yeah, at that. And I think the size almost doesn't matter. You know whether it's one book or many. You know is somebody willing to say, you know, "I will record somewhere that this thing is that this readable thing or you know whatever it is, is available for lending." Somebody has to make that statement. So if very small organizations can make agreements with larger organization who have that ability to make those statements then there is progress for very small organizations. >> ERIC MILLER: Oh be careful. Let's not electric ourselves before. So I think that's a great question, but one of the things that we looked for in terms of sort of initial Libhub sponsors was and sort of partners for experimentation was making sure that these libraries reflected different kinds of community needs. So you have very large Denver public, but additional regional in smaller libraries are sort of part of this. Different ILS systems, different budgets, different needs. The notion of at a certain level being able to surface that information up to the web to Ted's point size actually is less of a consideration. But being part of this larger shared space is what's important in terms of linking. So whether it's a large library, or a small library, or a rural library, or basically a K-12 library. You know, we think of libraries in perhaps a lot of new ways. We can start thinking about schoolhouses and their libraries or churches in their libraries. At a certain level it's not just about necessarily lending it's about being able to describe an asset and sort of move forward. And to Beacher's point there's a lot of different ways of sort of looking at this. You know, we're focusing on visibility as a way to help understand the potential impact that BIBFRAME has to offer. And drive additional interest in terms of new vendors and new capabilities that frankly haven't been in the library space. Working on Link Data, working on RDF, there are a lot of new services and service providers that don't come from the library community when you start looking at applying this in global web standards. And the implications of what this means to smaller and smaller quote unquote, "libraries" or new kinds of entities that begin to want to share information. That's just an interesting aspect of what we hope actually will be part of what happens when you make these kind of assets available to the web. I think libraries have a very important role on this in terms of the credibility that they offer. But the difference between sort of starting to expose data in the credibility that comes over that I think is actually where libraries as a powerful center have a huge differentiator over others that are starting to explore Link Data, and Link Data Publishing, and Link Data services. ^M01:20:01 When we start surfacing that credible information. I believe the trust that we have in the physical space starts to basically apply to the web space. And that's where I think a huge opportunity will emerge next. And we can start rethinking than what libraries roles are in terms of physical locations you know from the large all the way down to the small as small as a chicken coop. >> BEACHER WIGGINS: Yes? >> That's actually a lot lovely segue. I come from a small public library. And when I leave here and go to the exhibits and I walk up to my ILS because I can't do it without them. What is the message I need to take to them? >> ERIC MILLER: Help me be visible on the web. >> TED FONS: Yeah I mean you could refer them to people like us, you know on this panel. The people that understand these things. I mean, I think there's some you know some learning, you know, and the traditional systems. I think that there's some you know collaboration they can do you know with people like you know OCLC. As a cooperative we work in the whole environment. You know with the commercial providers and open source and that whole environment. So, you know, so we can help. >> BEACHER WIGGINS: Did you want say anything about that? On this piece? >> PHIL SCHREUR: No. Well so [laughter] the only thing I wanted to say is as far as the project that we are thinking about. We think the vendor aspect of it is a huge aspect that we need to deal with from the very beginning. So as I mentioned OLE is something a system that Davis is working with. We have not actually reached out to any ILS yet but we are reaching out to vendors. For instance, that supply cataloging like Casellini [assumed spelling] and MARC Now [assumed spelling] and our authority vendors to try to get the whole community engaged in the effort as well. Because I have to admit a lot of the data that we come in comes from these vendors. And it's not data that we actually create ourselves. >> I had a question for Philip from Stanford. >> BEACHER WIGGINS: She has, Sally had one. >> Oh I'm sorry. >> SALLY MCCALLUM: I was just going to say I to get there early for some vendors and not so early for others. But the message to them is I hope you're watching. I hope you're going to take an opportunities that come up for you to embrace some of this or to participate or whatever comes up. You know vendors have a very difficult time. They have a bottom line they have to worry about. And they don't have extra people to do new things so much as to fulfill the requirements that they are already obliged to do. And so we just want to sort of push them to stay awake I think. >> Okay. Thank you. No, I was just curious you mentioned in your presentation that you're going to focus on the music profile because you have a lot of expertise in that area. And I was wondering will you be I guess working with your other institutions looking at I guess like a basic I guess profile or template for like a single volume monograph. Is there going to be like some consensus on you know what kind of identifiers, what kind of info? I mean, I really have a difficult time visualizing all of this but I mean, I understand we're moving away from maybe a lot of the text, but I guess I'm just wondering will there be some discussions on you know what's important to have like in a very you know basic you know profile? Thank you. >> TED FONS: Yeah, I think what is. Also, I didn't realize this until yesterday when the group got together. We're all focused on very different, very different domains as far as the project go. So I think this creation of profiles will be extremely important. So for instance as we deal with our vendor records from Casellini we'll be dealing with books, but there was another institution that was interested in working with art objects like sculptures. Another was one interested in working with maps. Another one was is interested in working with data that had original script included. So yes I think it's important to approach these variety of formats at the same time and then develop those profiles through the profile editor so that when people work with this type of information that that profile is there to help you fill in the data elements which are considered important. So that is also why as far as music goes, I am anxious to reach out to MLA into PCC just to make sure that there is some sort of society buy in to what those standards should be. >> NATE TRAIL: Yeah, we all like books, but I think they're the easiest to do and I'm really excited that we are dealing with a lot of different formats and scripts because that's going to really stretch our understanding of what we need to do in the model. So it looks like it's going forward. >> I'm grappling to understand the relationship of Eric's statement about the libraries are going to contribute by just keeping on cataloging and that Eric is talking about or I'm getting people mixed up. >> TED FONS: Yeah, I said it. >> Anyway the OCLC model library should just keep cataloging, this is a fair model library should get in there and get their sleeves rolled up. What's the relationship between those two, you know, a library has to choose to actively participate or else just keep doing what it's doing and OCLC will do things for us. Can you talk about that relationship? >> TED FONS: yeah. What I meant there was recognizing that the landscape of the local systems, it doesn't change very fast. You know, and that's not a criticism it's just an observation. Right so local systems where people are doing cataloging or you know, creating things doesn't change very fast. And people are trained in those systems, and are used to using those systems, and that activity or that sort of model will be with us for a long time. And so I am recognizing that so you know, my concern is as we talk about the possibilities of the future and the things we can do with big data tools and all those things that people will be discouraged. Oh my goodness, what am I going to do you know I just what I have is my local system today and people trained and you know. And so my point there is that if you, you know, so don't panic. That's number one. Despite the weather don't panic. You know, and you know we at OCLC our approach is, we can mind this data from these legacy formats like MARC, and we can continue to do that. In parallel, so it's not frozen but in parallel we can modernize the tools that are used to do the you know, metadata creation and metadata management, quality management. So what that means is you know those tools will change as well. Then the tools that OCLC can control are you know the services that we build on top of WorldCat. Now as local systems evolve then you know we can recognize that data coming in as well, as people do experiments we can recognize that data as well. So I hope this is helpful. You know what I was saying is don't be discouraged and think you have to you know suddenly change your entire tools are overnight. You can continue doing what you're doing. We can mine from that in the evolution that you heard about across the panel today will continue. >> ERIC MILLER: I'm saying roll up your sleeves and get into it [laughter]. I mean, these are two sides of that spectrum right. And change is the scary thing. That change process. But you know I'm very much you know we at OCLC you know, our editors, our backend databases, our transformation tools, these are all designed to allow people to sort of get into this and explore and understand what the implications of this are. The training and assessment are about helping that sort of transition. And the work that we do in a wide range of different projects is a way of sort of looking at this and sort of saying look how can we start without necessarily breaking existing workflows. How can we start basically becoming more you know, effective, in that. But that's not necessarily us telling you. That's you and us and everyone else having a dialogue. In that part of that is rolling up our sleeves and starting to explore. One of the interesting projects that came out of the training and into the Libhub initiative was the Denver Public Library sort of realizing that look they had more information about Molly Brown than anyone else. And the web doesn't know this. So the notion of the linkable Molly Brown as a way of starting to explore how we might connect these additional resources. It isn't something you can describe in MARC. But it is something that they want to start using in BIBFRAME to start exploring how to connect all these things together. And it just takes rolling up your sleeves, and exploring this, and making this happen. And they started it and the folks of George Washington sort of jumped in and said this is how we do it, and the folks at Zepheira sort of said these are the tools you might need. And it's sort of snowballed into sort of a really nice story that complements a lot of the existing exploration is going on. But shows the value of looking at BIBFRAME, not just as a MARC replacement bar as a cataloging framework for starting to explore new ways in which their patrons are interested in finding information. And new ways in which libraries are interested in engaging their patrons to sort of to say "Look, these are the kinds of things that I have that you might be interested in." So I don't think either of us are saying stop cataloging right. ^M01:30:00 But I do believe that we're in a fantastic place in terms of and you know that inflection that change between sort of you know what, how we've done things today and what the art of the possible is going forward. And you know, we are looking at the same problem from different perspectives. Everything is her lover the same thing but I will air with the side of you know, let's roll up our sleeves and let's try things out and let's talk an experiment they were goes. >> BEACHER WIGGINS: will wrap up with these last two questions. >> I just like to ask a simple and quick question when we talk about RDA as a descriptive standard you know BIBFRAME as a communication standard. From what I've seen today. There could be various descriptive standards that we can use for Link Data purposes. We are not just limited to RDA. We as a community we've chosen to follow RDA bar in your opinion really is RDA the best choice here? >> ERIC MILLER: You're breaking up [laughter]. >> BEACHER WIGGINS: I guess I have to answer that at least the RDA part. No. RDA isn't the only one. It certainly is the one that the Library of Congress and a large part of the library community followed because we reached that conclusion and we helped develop that. But as Eric pointed out on some of his slides. There are other content standards that could be used. And that is going to be the beauty of BIBFRAME, or any Link Data schema is that it should be able to accommodate a variety. Or else we could do other things that we been talking about. >> ERIC MILLER: I mean, if I can just complement and amplify it even more. It's actually the power of BIBFRAME in that context. So it allows those that are interested in a particular cataloging practice to expose that cataloging process in a way that actually connects with other cataloging practices. And one of the interesting benefits of this was, you know, showing how RDA cataloging in BIBFRAME and Museum cataloging in BIBFRAME. Even though they didn't know they were doing it in BIBFRAME sort of lines up in which they can sort of build off of each other's assets. And what was even more interesting was in fact not just the asset side of it, but the personal side of it. When you realize in fact, you could kind of connect with somebody, that you didn't know that it was down the hall. That was working on the same problem about the same thing, but you were just looking at it from different perspectives. So we sometimes focus on the technology and a lot of the modeling and things like that but the interesting aha moment for me in the context of this was just watching six months ago. You know some of the folks in the museum community looking at this and some of the people from the library community looking at this. And realizing that this was a way in which they could actually talk to each other and didn't even know they were working on the same problem. And then that kind of collaboration and building off of each other's work. You know, I mean that to me is that personal connection was a really powerful one. So I don't think it's going to be anyone but if there's a way of accelerating those connections. That's a powerful thing. >> BEACHER WIGGINS: And one last question. >> Does OCLC have a long-term plan to expose the full detail of metadata in WorldCat to the world in BIBFRAME or some other Link Data format or keep with some smaller subset as expressed in schema.org or schema.org with supplements? >> TED FONS: Yes. So our goal here is to like I said, focus first on the top level. Make sure we're doing that very well as well as we possibly can. And then kind of work down from there to get to deeper and deeper granularity. And once we get there, we recognize there's probably multiple vocabularies to do that kind of thing. Things like BIBFRAME, which are you know, capturing a deeper level of granularity. So yes, that's the plan. >> Into the world and not just to member? >> TED FONS: Well, yeah I mean that's what we do on the web. I mean that's what you know so WorldCat.org already exposes Link Data as you know, as part of its offerings so that's, so as we mature that and as we mature the modeling that will follow. >> BEACHER WIGGINS: Thank you all very much. ^M01:34:48 [ Applause ] ^M01:34:52 >> This has been a presentation of the Library of Congress. Visit us at loc.gov.