The bibliotek-o Framework: Principles, Patterns, and a Process for Community Engagement
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Formal Metadata
| Title |
| |
| Title of Series | ||
| Number of Parts | 15 | |
| Author | ||
| License | CC Attribution - ShareAlike 3.0 Unported: You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal and non-commercial purpose as long as the work is attributed to the author in the manner specified by the author or licensor and the work or content is shared also in adapted form only under the conditions of this | |
| Identifiers | 10.5446/47594 (DOI) | |
| Publisher | ||
| Release Date | ||
| Language |
Content Metadata
| Subject Area | ||
| Genre | ||
| Abstract |
|
10
00:00
Lecture/ConferenceMeeting/Interview
01:10
Computer animation
05:38
Computer animation
06:13
Computer animation
07:43
Computer animation
10:50
Computer animation
12:43
Computer animation
16:26
Computer animation
17:12
Lecture/ConferenceMeeting/Interview
19:05
Meeting/Interview
20:22
Computer animation
Transcript: English(auto-generated)
00:01
Okay, so my name is Jason Gavarri and today Stephen Folsom and I will be talking about BiblioTeko which is in many ways an implementation of bibframes, so I'm really glad that we went following Nate and Ray. The work today is part of the
00:22
Linked Data for Libraries Labs and Linked Data for Production Andrew W. Mellon Foundation funded projects, which you had an introduction to this morning by Regine, so thank you for saving us that effort. But that really is, but more specifically, this is an effort of the ontology group, the joint ontology group within those two projects.
00:43
So that is really to say that this work is the effort of significantly more people than Stephen, myself, and our co-presenter Rebecca, or co-author Rebecca. And way too many people to name individually, but they represent six institutions, Columbia, Cornell, Harvard, Princeton, Stanford Universities, as well as the Library of Congress.
01:05
We're going to keep it fairly superficial because we are limited on time, so with that I will move forward. So what is BiblioTeko? BiblioTeko is a very simple model for representing bibliographic, the bibliographic domain, and
01:24
the core concepts and models are based on bibframe, a work instance item you can see in the diagram there, and building upon bibframe is the biblioTeko ontology itself, which in many ways supplements bibframe with an extension, as well as in places, replaces patterns in bibframe
01:46
in our attempt to demonstrate alternative modeling, which we'll talk about in a moment. On top of that, are additional ontologies such as schema, DC terms, FOAF, etc. All of these are available at the biblioTeko.org site. All of these links are in the slides later,
02:04
and you can you can find them through there, if you want to know exactly which terms we're using from which ontologies. Alongside of the ontologies are Shackle application profiles, which are in development currently, although there are some that are currently available in GitHub already,
02:24
that define implementation of the use cases that we are trying to to describe or to create data around within the projects, as well as mappings are also available within the GitHub repo, specifically currently the BIPCO
02:41
MARC bibliographic standard record mapping to biblioTeko. So what is the purpose of this? Why did we bother creating yet another core model? This really began as an attempt for the ontology group to engage in assessment of bibframe 2.0.
03:01
Conveniently, these two projects began the same month as bibframe 2 came out. And when we said, okay, we're going to use bibframe in these projects, we started by saying, we don't actually know what's in bibframe, because we knew bibframe 1 pretty well, and now bibframe 2, and we have no idea what they changed, and we have to assess. And in doing so, that assessment sort of took on a life of its own.
03:25
It was evaluation and engagement, because we believe that as people who are going to use these models, we need to deeply engage with them. And it was also an attempt to look towards extension of bibframe itself. We went further to deviate from bibframe and demonstrate these alternative patterns
03:44
with the goal of trying to identify whether or not there are other ways to go about creating our data that could be more queryable, could be more useful for us. The verdict is out on some of those, because we haven't tested yet.
04:02
But the idea here is that we are not in any way creating a competitor to bibframe. We are simply demonstrating these alternative patterns so that they can be assessed by the community, such as yourselves, and also by Library of Congress, who are obviously the bibframe developers,
04:23
to be determined whether or not some of them could be folded into future iterations of bibframe, future versions of bibframe. And all of this alongside trying to assess whether or not we can accept our legacy data and bring it into a real-world orientation alongside creating new data.
04:44
Now, future plans for BiblioTecco are actually frozen. We are not continuing development of this model, and in fact, not even all of the six institutions who have created this model are using the model.
05:00
The goal here, again, really is assessment and evaluation to demonstrate alternatives. We defined a rather robust change management process available in the GitHub repo and have a ton of work that we identified but never completed because of time constraints. But we believe that we did succeed in our goal of assessment.
05:23
So we are very proud of that. And we also really want to engage with the community around this. We want issues on GitHub. We want people to say this is great or terrible. Why did you do this? We really like the latter, particularly. So there's plenty of places to get documentation.
05:42
Perhaps a few too many. But there are many. There's bibliotecco.org. There is a wiki for this as well as a GitHub repo. We try to cross-link as much as possible. But between these spaces, there is actually plenty of rich documentation so that the community who wishes to either implement or analyze has the tools to do so.
06:05
There are OWL files, shackle application profiles, mapping spreadsheets, narrative pattern descriptions, and diagrams. And speaking of pattern descriptions, these are in a way of rather at times very lengthy
06:23
analysis of what exactly is within the pattern itself. These merge modeling and implementation to demonstrate diagrams of how we believe the pattern should be profiled, what exactly should be, what kinds of data
06:40
should be produced through this. And the patterns that were created where the bib frame patterns are either supplemented or replaced follow a few principles. And in discussing three of these patterns in a minute, we'll outline which principles sort of align with which.
07:00
Those in general being reused and aligned with existing vocabularies and ontologies. And also inversely for anything that we do meant ourselves doing so broadly enough that we encourage reuse by others. Trying to use OWL axioms in moderation so that we do not over constrain the ontology itself.
07:23
Preferring object properties and structured data over unstructured literals in not worrying so much about and not being mired in our legacy data in hopes that one day we will have more future data created natively in these models. And so we're going to focus on three patterns.
07:41
I'll take the first and then kick it over to Stephen. The first being activities. The principles that this aligns with are preferring the simplest and most general model capable of faithfully representing the data. And the second being using OWL constructs in moderation.
08:02
So for a just very two second primer, within the bib frame model, activities in a way are divided between provisions and contributions with provisions being only on the instance level and contributions I thought were only on the work level.
08:21
But apparently they're now working since our item used with or it says it in the, it says it last time I checked at least. And we feel that this is unnecessary complication in that the two profile the same. The two have the same kind of, it's the same concept. We saw the semantics as being at least having a superclass there that would allow for it.
08:47
And also on the other hand, bib frame also uses BF role to define the agent's interaction with the resource. In bibliotheco, there is a single pattern for representing activities, whether they be in the bib frame mode, provisions or contributions.
09:05
And that activity uses subclassing as the inherent way of identifying how the agent is related to that activity or that resource. The significant features that we feel that this model changes or at least aligns with are that it's a consistent model across the board for activities on a resource.
09:25
There is a superclass and a shared superclass and a shared property for all types of activities. And there aren't any formal constraints about which type of BF core class that something aligns with. So the bibliographic resource is not specified as being this activity is only related to
09:46
BF instance, for instance. And I think that's all I'm gonna say about activities. And with that, I'll pass it over to Stephen. Okay. So jumping into the next pattern, bib frame uses two distinct patterns for capturing
10:05
content carrier and media type. So if you're familiar at all with RDA, these concepts are sort of an embodiment of an implementation of RDA. And so what you see here is you see two types of ARCs.
10:20
You see RDF type and you also see content carrier and media. And so really what surfaces multiple query paths that you then have to take into account in your applications. And so there's the subclassing of works instance items and then linking also to content
10:42
carrier and media resources that essentially captures the same, if not similar, if not the same ideas. So at least in the lifetime of this grant, we're proposing that we commit to subclasses. And so this is sort of like what Valentin was saying with their schema.org extensions.
11:06
So schema has creative work. And then the types of resources can be extended through subclassing. And so RDF type, everybody knows, is a pretty ubiquitous property. Systems know how to use it. It's a fairly safe way to implement and extend bib frame.
11:26
And the other thing to point out here, or a couple of things to point out here, an alternative model would be to not ignore these content carrier and media vocabularies that are often Scoss vocabularies. But even with that, we don't necessarily need a different property per resource.
11:45
So CDOC CRM has P2 where you can classify resources and use a single property to say this entity has this type. But we're not actually linking to a class. We're linking to a Scoss concept. It gets around that sort of punning question.
12:02
The other thing to point out here is that we haven't extended media type. And we're trying to push the boundaries here on instances so that we say we define instances and we extend instances in such a way that in their definition is this idea of what it takes to play back.
12:21
So the bib online resource here is a subclass of electronic. And electronic in its definition says requires a computer for playback. So having another pattern that says here's a media property and computer off from the resource, we're really striving for some simplicity there.
12:44
So moving on, maybe a harder pattern to describe is the relations document. I'm only going to touch on this in surface. There's, as Jason said, a document that goes in much more detail. But in this case, we decided that so for some of the properties that Ray was talking
13:01
about with expression of and related to in the sub properties, they reach a limit. Like they've held themselves to being a framework and ask that you extend. And so this is sort of an approach to extending sort of the core relationships between principal or primary entities in bib frames.
13:25
So an example of this would be the derivative work. So there's a pretty deep hierarchy of derivative works in RDAU. And we're suggesting to use that. And I'll acknowledge OSMA here and say, yes, the OPIC URIs are a pain.
13:40
But we'd rather not mint them ourselves. Some other motivations behind this or another motivation behind this was that we felt that we wouldn't have to train catalogers necessarily. I mean, they're getting a lot of RDA training. So hopefully they'll recognize some of these terms. And the other is the RDAU properties. If you don't know, RDA unconstrained has open domains and ranges.
14:03
So RDA wants to use the Ferber stack work expression manifestation and item. And this lets us get around that in the way that bib frame strives to do. So some of our more recent work is trying to understand with the definitions that we
14:21
or the properties that we've chosen to use, how do we use those in this sort of bibliotheco environment? So we've identified properties like DC terms description and the domain is open. And we want to build these applications that use the properties that we've identified in deliberate ways. And the ontologies themselves aren't enough to build an application.
14:44
You have to specify how you're going to use each of these properties. And Huda Khan in the audience has great presentations about ontology to app. And I'd recommend going out and looking at those. And so with Shackle, the shapes constraint recommendation out of W3C,
15:04
there's lots of ways that it can be used. A lot of people are considering a validation language. We were happy to find axioms in the model that allow for form building.
15:21
So and really, as the Princeton folks pointed out, we got really far behind and bogged down in modeling. So we got behind in building an application that would let catalogers catalog. And so this is our way to a standard way to define what the cataloging tools would use.
15:43
Examples of this you can find in the repository, the most mature being the hip hop repository. And really, this is coming from a very content specific set of decisions. So for audio works, what properties and classes do we want in that form so that
16:00
you don't have all of the frame and you don't have all of the other extensions in a big heavyweight editor. So really trying to tune it for a content type and moving images coming along. And one of our goals is to have it for any project that wants to use vitro lib, which is the editor that we're building to define some shackles
16:21
so that we can transform the shackle into what the editor needs. And so I'm going to leave this out. I'll just use this slide as a plug for a breakout session, because we don't have necessarily the time to go into it. But I'm really interested in the semantics of form building and what shackle can do versus
16:41
what it isn't able to do. But we know that there are certain semantics and form building that we have in JSON configuration files and elsewhere. So there's a lot of experience in the room around building forms. So I'd love to have a conversation about a standards based way of doing that. And given that shackle is RDF based, we can have URIs and share shapes and share some
17:06
of these definitions so that across applications, we don't have to recreate some form building. So leaving time for questions. Really, we do hope that people will engage with our work.
17:22
And either here at the conference or online, as Jason said, there's lots of opportunities. Please comment publicly. It's sure we'll have a side conversation. But the more that this stuff can be out in the open, I think the more useful it is. And if you have your own assessments, I know we're not the only groups looking at BIBFRAME
17:42
and trying to make these sort of evaluations. We'd love to hear from you. And lastly, though, BiblioTeko isn't going to progress as something that we maintain into the future. We're really interested both with our engagement with BIBFRAME and other ontologies,
18:00
your experience about building community based open process for communicating and maintaining ontologies. Thank you. So thank you. Thank you for your presentation. Do we have any questions or comments from the audience?
18:27
We have time. Come on. OK, there's one.
18:43
Hi, my name is Christina. I'm from the Danish Library Center. I was just wondering, actually, what your motivation behind this was. Was it like you saw some data that wasn't really well represented in BIBFRAME? Or was it you mentioned something in the beginning about making it more searchable?
19:01
Or what was your motivation for this? Was it presenting it in a user interface? Then this doesn't make sense in BIBFRAME. What were your thoughts about this? I think it would depend on the pattern itself or the fragment that we were concerned with.
19:21
I can't offhand think of an individual instance where it was like, this was about querying, and this was about display or something like that. But overall, it was a sense that because our assessment was going class by class, property by property through, it's easy to, especially when you have 20 people in the
19:42
room, suddenly say, oh, we really wish it was doing this instead of that. And we really wanted it to be simpler in this one area or more complex in this other or whatever the case might be. And it took a life of its own prior to us really having a mode of operation in that
20:05
way. We wasn't thought of as being a massive undertaking, a whole framework when we began the project. It was really just, let's document some questions that we have and some issues we have. OK, thank you once more.