Author: Jon

  • The RDA Toolkit == The RDA Data Model

    Sitting in the CC:DA meeting at ALA last week, listening to the RDA report and the questions it generated, I was struck yet again by the disconnect between the understanding by the audience of RDA as a content model — the Instructions as embodied in the Toolkit — and the presentation by the speakers of the changes to both the RDA Toolkit and the RDA Registry in terms of the changes to the RDA data model.

    In the case of the Toolkit, a significant part of the reorganization is intended to better align the Toolkit with the Registry, the data model. In the case of the Registry, a significant portion of the changes are intended to continue the current data model’s status as an implementation of FRBR, and to adjust the model to accommodate the consolidation of the various FRBR models into the IFLA LRM.

    From its inception, the Toolkit has reflected the RDA instructions, an upgraded and internationalized version of AACR2, originally intended to provide content instructions for populating the MARC21 carrier, while conceptually describing the data in terms of the FRBR-oriented RDA carrier. The changes to the Toolkit are a necessary evolution away from the slowly devolving MARC21 data model toward the RDA data model. But the audience doesn’t, can’t, easily see that distinction, and the ‘questions’ swirl around expressions of discomfort with the changes that reflect a lack of understanding of the broadening disconnect between the Toolkit and the Registry, and between RDA and MARC21.

    This cognitive dissonance must be addressed by the RSC and its representatives at every future opportunity when discussing the rationale for both the alignment of the RDA data model with IFLA LRM, and the alignment of the RDA content model in the Toolkit with the RDA data model. In that context, the global bibliographic community needs definitive guidance in understanding the relationship between the entities and elements of the RDA data model and the MARC21 data model in which they currently create and maintain their metadata.

  • A Manifesto for Managing Metadata in an Open World…

    Metadata is produced and stored locally, published globally, consumed and aggregated locally, and finally integrated and stored locally. This is the create/publish/consume/integrate cycle.

    Providing a framework for managing metadata throughout this cycle is the goal of the Dublin Core Abstract Model and the Dublin Core Application Profile (DCAM/DCAP).

    The basic guidelines and requirements for this cycle are:

    • Metadata MUST be syntactically VALID and semantically COHERENT when it’s CREATED and PUBLISHED.
    • Globally PUBLISHED Metadata SHOULD be TRUE, based on the domain knowledge of the publisher.
    • PUBLISHERS of metadata MUST publish the semantics of the metadata, or reference publicly available semantics.
    • CONSUMERS of published metadata SHOULD assume that the global metadata is locally INVALID, INCOHERENT, and UNTRUE.
    • CONSUMED metadata MUST be checked for syntactic validity and semantic coherence before integration.
    • AGGREGATED metadata SHOULD indicate its PROVENANCE.
    • CONSUMED metadata MAY be considered TRUE based on its PROVENANCE.
    • Locally INTEGRATED global metadata MUST be syntactically VALID, semantically COHERENT, and SHOULD be TRUE based on local standards.

    The DCAM takes as its base the rdf data model because of its simplicity and flexibility of that model. The DCAM refines the rdf data model in order to support syntactic interoperability between the rdf data model and non-rdf data models.

    A DCAP defines a distinct subset of the set of all things and defines the domain-specific knowledge of the properties of those things, and the relationships those things have to other things. It expresses that knowledge through the DCAM and a set of related documentation (the Singapore Framework). A complete DCAP should provide the necessary domain-specific infrastructure to fully support the create/publish/consume/integrate cycle for any system, any domain, and any data model. A DCAP based on the DCAM should be able to be used by a machine to generate a valid data model in any modeling syntax, and any modeling syntax should be able to be expressed as a DCAP.

    We strongly recommend that metadata be published and consumed using the RDF data model. The strength of the rdf data model for publishing and aggregating metadata lies in its extreme flexibility of expression, it’s self-describing semantics, and its assumption of validity. These strengths become weaknesses when creating and validating local metadata that must conform to a set of local restrictions. An effective DCAP will provide metadata creators with the ability to create locally valid metadata, publish it as rdf, validate consumed rdf against local standards, and ultimately integrate global metadata into the local knowledge base that contains local descriptions of the things being described.

    There are many systems, many data models, many publish and subscribe models, many storage and validation models. There are many paths to integration. There are very few that provide a generic and neutral method for modeling and documenting inter-system metadata integration. The DCAM/DCAP specification has the potential to be one of those few.

  • Non-sequiturs

    Corey Harper seems to have started a collection of statements taken out of context, some of which I’ll share here:

    “If it’s not fish it has to be linked data” –Jennifer Bowen

    “I’m too far underground to know where I am” –Jon Phipps

  • Curmudgeon redux

    When I used the word “curmudgeon” in my previous post, I had an apparently uncommon definition in mind: unflinching truth teller. I’ve actually taken minute pleasure in thinking of myself that way ever since, in the not-so-long ago I asked my new boss what my new role on the team would be and, knowing me, he said “curmudgeon?”. I had to hunt a bit to find a definition that wasn’t the Andy Rooney-esque “crusty, ill-tempered, old man”, but I kind of liked this one:

    …Nature, having failed to equip them with a servicable denial mechanism, has endowed them with astute perception and sly wit.

    Curmudgeons are mockers and debunkers whose bitterness is a symptom rather than a disease. They can’t compromise their standards and can’t manage the suspension of disbelief necessary for feigned cheerfulness. Their awareness is a curse.

    Perhaps curmudgeons have gotten a bad rap in the same way that the messenger is blamed for the message: They have the temerity to comment on the human condition without apology. They not only refuse to applaud mediocrity, they howl it down with morose glee. Their versions of the truth unsettle us, and we hold it against them, even though they soften it with humor.

    – JON WINOKUR

  • The curmudgeon’s table

    Today I participated in a Linked Data Unconference at ALA 2010 in Washington DC, which was remarkably successful. Organized by Corey Harper from NYU and ably moderated by Karen Coyle, about 50 of us held two sets of three hour-long, highly engaging breakout discussions with reports back to the larger group. I participated in a discussion of some of the practical difficulties encountered trying to implement Application Profiles in a far from perfect Linked Data environment (I strongly recommended creating a local mirror of inadequately expressed in RDFS/OWL, but otherwise useful, data models) and a discussion of scholarly Linked Data use cases and data reuse (we looked hard at VIVO and VIAF and discussed metering data usage). It was fun! I was surprised.

    Afterwards a few of us had a fine lunch together, discussing the nuts and bolts of RDA and the future of cataloging, a lively and fascinating discussion in the best possible sense of ‘lively’ and ‘fascinating’. It strikes me in retrospect that we formed a kind of curmudgeon’s table — Diane Hillmann, Corey Harper, Eric Hellman, Ed Summers, Karen Coyle, and me. All of us sharing strong opinions, agreeing and disagreeing whole-heartedly and with considerable good humor and affection. A table of warm and engaged people, knowing that it’s too late to save the world, maybe too late to save cataloging, but it would still be really interesting to try. It was great fun. I wasn’t the least bit surprised.

  • We don’t need no stinking Access Points

    …in the RDA Ontologies. Do we? After all, they’re a big part of the ‘Access’ in Resource Description and Access (RDA). But they’re not particularly semantically meaningful, especially if you have the component parts available. An Access Point is just a structured string. For instance a ‘Publication Statement’ Access Point for “The Daytona daily news” might look like:

    “Daytona Beach, Florida : Geo. F. Crouch, 1903-1926”

    It has a formal syntactic structure, and semantics derived from adherence to that structure when the string is created:

    “Place of Publication” : “Publisher’s Name”, “Date of Publication”

    Note that the punctuation is part of the formal grammar that helps parse a grammatically correct statement into its semantically meaningful constituent parts.

    And this is the way we’ve been doing things forever (well it seems like forever) — semantics is derived from proper use of a syntax that everybody who is creating and using shared data has agreed upon in advance. And as long as everybody uses precisely the same syntax this works great. It works really, really well with structured syntaxes like MARC21:

    260 $a Daytona Beach, Florida
    260 $b Geo. F. Crouch
    260 $c 1903-1926

    …and hierarchical syntaxes like XML:

    <publicationStatement>
      <placeOfPublication>Daytona Beach, Florida</placeOfPublication>
      <publisherName>Geo. F. Crouch</publisherNam>
      <dateOfPublication>1903-1926</dateOfPublication>
    </publicationStatementt>

    The RDA documents go so far as to call an Access Point an Element and its constituent parts Sub-Elements, again clearly thinking of this nice syntacticly-defined semantics.

    But what if your data says, semantically, that “Place of Publication” isn’t the Name of the place or the Label for the place, but a URI that identifies the Place itself; a resource rather than a string. The Access Point rules don’t let you stick a URI in the Publication Statement where a string is supposed to be.

    What about “Publisher’s Name”? That’s clearly going to be a string no matter what — names tend not to be resources. But there’s probably a Publisher resource out there, somewhere, with a URI that identifies the Publisher and probably has a Name or a Label property that provides a string that you can stick in the Publication Statement.

    We’ll just ignore “Publication Date” for the now, since that’s a very different can of worms: slimy, smelly worms.

    At the moment, RDA doesn’t acknowledge the existence of the resources supplying the strings for an Access Point, and it lets substantial ambiguity sneak in with property names like “Place of Publication” rather than “Place of Publication Name” like they resolved with “Publisher’s Name”. But that ambiguity didn’t exist when all the data was strings — strings you used for indexing, and displayed to the user, and didn’t have to go fetch from somewhere because they were right there in that 260 field.

    I listened to a radio program that referred to all of the money that everyone in the world has available to invest as “The Global Pool of Money” and I think that applies quite nicely to the Linked-Data notion of the Semantic Web — “The Global Pool of Data”.

    The open world model of the Semantic Web assumes that you will never have all of the available data that describes a resource, and the RDF data model supports this. Resources often exist, outside of traditional library data, available from the Global Pool of Data, that can supply the necessary labels.

    But of course we usually just have the labels. This is library data made for cards that need to be put in the correct order and read by a person. And the Global Pool of Data usually just has a bunch of resources. This is linked data, in no particular order at all, meant to be read by a machine.

    So, specifying Access Points as pre-coordinated strings actually provides us with a major opportunity when defining the ontologies; several opportunities actually:

    • We can formalize each Access Point specification into what Dublin Core calls a Syntax Encoding Scheme (SES) and say that each Access Point has a datatype.
    • We can clarify the semantics of using a label rather than a resource for properties (sub-elements) like “Publisher’s Name”
    • We can clarify the semantics of using a resource for “Place of Publication” and say that the label used in an Access Point must be the Name of the Place and this is distinctly different.

    So, refined to use properties that are a bit more semantically clear, we have a slightly modified Publication Statement:

    “Place of Publication Name” : “Publisher’s Name”, “Date of Publication”

    …we tie these properties specifically to a FRBR Manifestation that RDA says must be what they describe, and in RDF the supporting ontology looks like:

    rda:placeOfPublicationManifestation a owl:ObjectProperty
    rda:PlaceOfPublicationName a owl:DatatypeProperty
    rda:publisherManifestation a owl:ObjectProperty
    rda:PublisherName a owl:DatatypeProperty

    Here’s our sample instance again (by the way, this data is from a real linked data resource):

    “Daytona Beach, Florida : Geo. F. Crouch, 1903-1926”

    <http://chroniclingamerica.loc.gov/lccn/sn93063916>
      rda:publisherManifestation <http://???> (blank, but we know one must exist)
        rda:PublisherName "Geo. F. Crouch"
      rda:placeOfPublicationManifestation <http://dbpedia.org/resource/Daytona_Beach%2C_Florida>
        rda:PlaceOfPublicationName "Daytona Beach, Florida"
        rdfs:label "Daytona Beach, Florida"

    This tiny chunk of data was gathered by hand and mapped (by me) from the existing resources and the labels supplied by those resources.

    Someday there will be services that comb through linked data looking for missing data like that Publisher resource, and will perform a search on the Global Pool of Data looking for resources with labels matching that library data, expressed in RDA/RDF, fill in the missing pieces and present the lucky cataloger, and ultimately the user, with all that rich linked data.

    The eXtensible Catalog project is working on services that do just that kind of thing, so someday may not be too far off.


  • LCSH, SKOS and pre-coordinated headings

    This week, Karen Coyle wrote a post about LCSH as linked data: beyond “dash-dash” which provoked a discussion on the id.loc.gov discussion list.

    It seems to me that there are several memes at play in this conversation:

    LCSH and SKOS

    As Karen points out, LCSH is more than just a simple thesaurus. It’s also a set of instructions for building structured strings in a way that’s highly meaningful for ordering physical cards in a physical catalog. In addition, each string component has specific semantics related to its position in the string, so it’s possible, if everyone knows and agrees on the rules, to parse the string and derive the semantics of each individual component. The result is a pre-coordinated index string.

    These stand-alone pre-coordinated strings are perhaps much less meaningful in the context of LOD, but this certainly doesn’t apply to the components. I think what Karen is pointing out is that, while it’s wonderful to have a subset of all of the components that can be used to construct LC Subject Headings published as LOD, there’s enough missing information to reduce the overall value. As I read it, she’s wishing for the missing semantics to be published as part of the LCSH linked data, and hoping that LC doesn’t rest on its well-earned laurels and call it a day.

    Structured Strings

    Dublin Core calls the rules that define a structured string a "Syntax Encoding Scheme" (SES) and basically, that’s what the rules defining the construction of LC Subject Headings seem to be. It’s structurally no different than saying that the string "05/10/09", if interpreted as a date using an encoding scheme/mask of "mm/dd/yy", ‘means’ day 10 in the month May in the year 2009 using the Gregorian calendar. Fascinatingly, that same ‘date’ can be expressed as a Julian date of "2454962", but I digress.

    As far as I can tell, no one has figured out a universally accepted (or any) way to define the semantic structure of a SES in a way that can be used by common semantic inference engines, and I don’t think that anyone in this discussion is asking for that. What’s needed is a way to say "Here’s a pre-coordinated string expressed as a skos:prefLabel, it has an identity, and here are it’s semantic components."

    Additional data

    So…

    "Italy--History--1492-1559--Fiction"

    …is expressed in http://id.loc.gov/authorities/sh2008115565#concept as…

    @prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
    @prefix skos: <http://www.w3.org/2004/02/skos/core#> .  
    @prefix terms: <http://purl.org/dc/terms/> .  
    @prefix owl: <http://www.w3.org/2002/07/owl#> .
    
    <http://id.loc.gov/authorities/sh2008115565#concept>
        skos:prefLabel "Italy--History--1492-1559--Fiction"@en ; 
        rdf:type ns0:Concept ;    
        terms:modified "2008-03-15T08:10:27-04:00"^^<http://www.w3.org/2001/XMLSchema#dateTime> ; 
        terms:created "2008-03-14T00:00:00-04:00"^^<http://www.w3.org/2001/XMLSchema#dateTime> ; 
        owl:sameAs <info:lc/authorities/sh2008115565> ; 
        skos:inScheme
            <http://id.loc.gov/authorities#geographicNames> , 
            <http://id.loc.gov/authorities#conceptScheme> ; 
        terms:source "Work cat.: The family, 2001"@en . 

    …and has a 151 field expressed in the authority file as…

    151 __* |a *Italy* |x *History* |y *1492-1559* |v *Fiction

    …which has the additional minimal semantics of…

    <http://id.loc.gov/authorities/sh2008115565#concept>
        loc_id:type "Geographic Name" ; #note that this is also expressed as a skos:inScheme property
        loc_id:topicalDivision "History" ;
        loc_id:chronologicalSubdivision "1492-1559" ;
        loc_id:formSubdivision "Fiction" ;
        loc_id:geographicName "Italy" .

    …and this might also be expressed as…

    <http://id.loc.gov/authorities/sh2008115565#concept>
       loc_id:type http://id.loc.gov/authorities/sh2002011429 ;
       loc_id:topicalDivision http://id.loc.gov/authorities/sh85061212 ;
       loc_id:formSubdivision http://id.loc.gov/authorities/sh85048050 ;
       loc_id:geographicName http://id.loc.gov/authorities/n79021783 ;
       dc:temporal "1492-1559" ;
       dc:spatial http://sws.geonames.org/3175395/ ;
       dc:spatial http://id.loc.gov/authorities/n79021783 .

    Making sure that those strings in the first example are expressed as resource identifiers is also something that I think Karen is asking for. (BTW, The ability to lookup a label by URL at id.loc.gov is really useful)

    I should point out that Ed, Antoine, Clay, and Dan’s DC2008 paper detailing the conversion of LCSH to SKOS goes into some detail (see section 2.7) about the LCSH to SKOS mapping, but doesn’t directly address the issue that Karen is raising about mapping the explicit semantics of the subfields.

  • Embrace the chaos

    I’m at the code4lib conference this week and I’m writing this as a reminder to myself for next year that, well, I just don’t care…

    I don’t care about the latest faceted search. I don’t care about the latest Open Source OPAC replacement. I don’t care about the latest Open Source ILS. I don’t care about project updates or software demos. I don’t care about API/standards/vendor/LC/OCLC bashing. I don’t care about reinventing the 20th century training wheels that library coders are so obsessed with (it’s their job after all).

    The people here are great, and I care about them both individually and as a class — I have a genuine affinity for library geeks of all kinds and this is, after all, a conference that tends to filter out everyone else. So it’s not like it’s not nice to be here, but the information being presented is useless to me to the point of irritation. And despite the fact that I like so many of the folks here, there’s definitely a cliquish feel to the whole thing with the hugely active IRC back channel of ongoing commentary (which really should be displayed where everyone including the presenters can read it), and the in-jokes (dongles). However welcoming, it can be a hard place to be a newbie.

    I want to stand up and shout “Folks. Listen. The world has moved on!! Deal with it.” I haven’t felt the need to set foot in a library, except for a meeting, in the last 10 years. I haven’t used a library OPAC or ILS, except for the purpose of researching the system itself in at least a year. I have logged into a university library to get access to subscription-only journal articles, but I actually found those articles in Google. My library’s only utility to me is as the proxy gatekeeper for ACM or Elsevier, and these days it only keeps me out.

    Many of the presentations and discussions are about alternatives: full-text vs. metadata; cataloger-created metadata vs. user-created metadata; Open Source vs. commercial; Google vs. the OPAC; data distribution standard A vs. standard B; marc vs. everything else. There’s a corollary notion that interoperability means everyone has to agree at some level. What really needs to happen is that everyone needs to get used to the idea that the ideal world in which everyone supports a few standards and data formats is the one we used to have, or imagine we had, and it’s gone. And it’s not coming back.

    What we have now is chaos. Many, many people are working very hard to maintain control over increasingly chaotic data and less-and-less predictable user behavior. They’re trying to hang on to users, even scholars, who are increasingly getting their information through Google or YouTube. They’re trying to figure out how to deal with steadily increasing digital collections, steadily decreasing user satisfaction, tone-deaf vendors, and sinking budgets.

    Let it go. Chaos is good. Keep your systems open and flexible. Watch. Listen. Integrate instead of compete with Google. Integrate data from the social networks. Share everything. Aggregate. Watch. Listen.

    Embrace.