Monday, November 24, 2014

Multi-Entity Models.... Baker, Coyle, Petiya

Multi-Entity Models of Resource Description in the Semantic Web: A comparison of FRBR, RDA, and BIBFRAME
by Tom Baker, Karen Coyle, Sean Petiya
Published in: Library Hi Tech, v. 32, n. 4, 2014 pp 562-582 DOI:10.1108/LHT-08-2014-0081
Open Access Preprint

The above article was just published in Library hi Tech. However, because the article is a bit dense, as journal articles tend to be, here is a short description of the topic covered, plus a chance to reply to the article.

We now have a number of multi-level views of bibliographic data. There is the traditional "unit card" view, reflected in MARC, that treats all bibliographic data as a single unit. There is the FRBR four-level model that describes a single "real" item, and three levels of abstraction: manifestation, expression, and work. This is also the view taken by RDA, although employing a different set of properties to define instances of the FRBR classes. Then there is the BIBFRAME model, which has two bibliographic levels, work and instance, with the physical item as an annotation on the instance.

In support of these views we have three RDF-based vocabularies:

FRBRer (using OWL)
RDA (using RDFS)
BIBFRAME (using RDFS)

The vocabularies use a varying degree of specification. FRBRer is the most detailed and strict, using OWL to define cardinality, domains and ranges, and disjointness between classes and between properties. There are, however, no sub-classes or sub-properties. BIBFRAME properties all are defined in terms of domains (classes), and there are some sub-class and sub-property relationships. RDA has a single set of classes that are derived from the FRBR entities, and each property has the domain of a single class. RDA also has a parallel vocabulary that defines no class relationships; thus, no properties in that vocabulary result in a class entailment. [1]

As I talked about in the previous blog post on classes, the meaning of classes in RDF is often misunderstood, and that is just the beginning of the confusion that surrounds these new technologies. Recently, Bernard Vatant, who is a creator of the Linked Open Vocabularies site that does a statistical analysis of the existing linked open data vocabularies and how they relate to each other, said this on the LOV Google+ group:
"...it seems that many vocabularies in LOV are either built or used (or both) as constraint and validation vocabularies in closed worlds. Which means often in radical contradiction with their declared semantics."
What Vatant is saying here is that many vocabularies that he observes use RDF in the "wrong way." One of the common "wrong ways" is to interpret the axioms that you can define in RDFS or OWL the same way you would interpret them in, say, XSD, or in a relational database design. In fact, the action of the OWL rules (originally called "constraints," which seems to have contributed to the confusion, now called "axioms") can be entirely counter-intuitive to anyone whose view of data is not formed by something called "description logic (DL)."

A simple demonstration of this, which we use in the article, is the OWL axiom for "maximum cardinality." In a non-DL programming world, you often state that a certain element in your data is limited to the number of times it can be used, such as saying that in a MARC record you can have only one 100 (main author) field. The maximum cardinality of that field is therefore "1". In your non-DL environment, a data creation application will not let you create more than one 100 field; if an application receiving data encounters a record with more than one 100 field, it will signal an error.

The semantic web, in its DL mode, draws an entirely different conclusion. The semantic web has two key principles: open world, and non-unique name. Open world means that whatever the state of the data on the web today, it may be incomplete; there can be unknowns. Therefore, you may say that you MUST have a title for every book, but if a look at your data reveals a book without a title, then your book still has a title, it is just an unknown title. That's pretty startling, but what about that 100 field? You've said that there can only be one, so what happens if there are 2 or 3 or more of them for a book? That's no problem, says OWL: the rule is that there is only one, but the non-unique name rule says that for any "thing" there can be more than one name for it. So when an OWL program [2] encounters multiple author 100 fields, it concludes that these are all different names for the same one thing, as defined by the combination of the non-unique name assumption and the maximum cardinality rule: "There can only be one, so these three must really be different names for that one." It's a bit like Alice in Wonderland, but there's science behind it.

What you have in your database today is a closed world, where you define what is right and wrong; where you can enforce the rule that required elements absolutely HAVE TO be there; where the forbidden is not allowed to happen. The semantic web standards are designed for the open world of the web where no one has that kind of control. Think of it this way: what if you put a document onto the open web for anyone to read, but wanted to prevent anyone from linking to it? You can't. The links that others create are beyond your control. The semantic web was developed around the idea of a web (aka a giant graph) of data. You can put your data up there or not, but once it's there it is subject to the open functionality of the web. And the standards of RDFS and OWL, which are the current standards that one uses to define semantic web data, are designed specifically for that rather chaotic information ecosystem, where, as the third main principle of the semantic web states, "anyone can say anything about anything."

I have a lot of thoughts about this conflict between the open world of the semantic web and the needs for closed world controls over data; in particular whether it really makes sense to use the same technology for both, since there is such a strong incompatibility in underlying logic of these two premises. As Vatant implies, many people creating RDF data are doing so with their minds firmly set in closed world rules, such that the actual result of applying the axioms of OWL and RDF on this data on the open web will not yield the expected closed world results.

This is what Baker, Petiya and I address in our paper, as we create examples from FRBRer, RDA in RDF, and BIBFRAME. Some of the results there will probably surprise you. If you doubt our conclusions, visit the site http://lod-lam.slis.kent.edu/wemi-rdf/ that gives more information about the tests, the data and the test results.

[1] "Entailment" means that the property does not carry with it any "classness" that would thus indicate that the resource is an instance of that class.

[2] Programs that interpret the OWL axioms are called "reasoners". There are a number of different reasoner programs available that you can call from your software, such as Pellet, Hermit, and others built into software packages like TopBraid.

Tuesday, November 18, 2014

Classes in RDF

RDF allows one to define class relationships for things and concepts. The RDFS1.1 primer describes classes succinctly as:
Resources may be divided into groups called classes. The members of a class are known as instances of the class. Classes are themselves resources. They are often identified by IRIs and may be described using RDF properties. The rdf:type property may be used to state that a resource is an instance of a class.
This seems simple, but it is in fact one of the primary areas of confusion about RDF.

If you are not a programmer, you probably think of classes in terms of taxonomies -- genus, species, sub-species, etc. If you are a librarian you might think of classes in terms of classification, like Library of Congress or the Dewey Decimal System. In these, the class defines certain characteristics of the members of the class. Thus, with two classes, Pets and Veterinary science, you can have:
Pets
- dogs
- cats

Veterinary science
- dogs
- cats
In each of those, dogs and cats have different meaning because the class provides a context: either as pets, or information about them as treated in veterinary science.

For those familiar with XML, it has similar functionality because it makes use of nesting of data elements. In XML you can create something like this:
<drink>
    <lemonade>
        <price>$2.50</price>
        <amount>20</amount>
    </lemonade>
    <pop>
        <price>$1.50</price>
        <amount>10</amount>
    </pop>
</drink>
and it is clear which price goes with which type of drink, and that the bits directly under the <drink> level are all drinks, because that's what <drink> tells you.

Now you have to forget all of this in order to understand RDF, because RDF classes do not work like this at all. In RDF, the "classness" is not expressed hierarchically, with a class defining the elements that are subordinate to it. Instead it works in the opposite way: the descriptive elements in RDF (called "properties") are the ones that define the class of the thing being described. Properties carry the class information through a characteristic called the "domain" of the property. The domain of the property is a class, and when you use that property to describe something, you are saying that the "something" is an instance of that class. It's like building the taxonomy from the bottom up.

This only makes sense through examples. Here are a few:
1. "has child" is of domain "Parent".

If I say "X - has child - 'Fred'" then I have also said that X is a Parent because every thing that has a child is a Parent.

2. "has Worktitle" is of domain "Work"

If I say "Y - has Worktitle - 'Der Zauberberg'" then I have also said that Y is a Work because every thing that has a Worktitle is a Work.

In essence, X or Y is an identifier for something that is of unknown characteristics until it is described. What you say about X or Y is what defines it, and the classes put it in context. This may seem odd, but if you think of it in terms of descriptive metadata, your metadata describes the "thing in hand"; the "thing in hand" doesn't describe your metadata. 

Like in real life, any "thing" can have more than one context and therefore more than one class. X, the Parent, can also be an Employee (in the context of her work), a Driver (to the Department of Motor Vehicles), a Patient (to her doctor's office). The same identified entity can be an instance of any number of classes.
"has child" has domain "Parent"
"has licence" has domain "Driver"
"has doctor" has domain "Patient"

X - has child - "Fred"  = X is a Parent 
X - has license - "234566"  = X is a Driver
X - has doctor - URI:765876 = X is a Patient
Classes are defined in your RDF vocabulary, as as the domains of properties. The above statements require an application to look at the definition of the property in the vocabulary to determine whether it has a domain, and then to treat the subject, X, as an instance of the class described as the domain of the property. There is another way to provide the class as context in RDF - you can declare it explicitly in your instance data, rather than, or in addition to, having the class characteristics inherent in your descriptive properties when you create your metadata. The term used for this, based on the RDF standard, is "type," in that you are assigning a type to the "thing." For example, you could say:
X - is type - Parent
X - has child - "Fred"
This can be the same class as you would discern from the properties, or it could be an additional class. It is often used to simplify the programming needs of those working in RDF because it means the program does not have to query the vocabulary to determine the class of X. You see this, for example, in BIBFRAME data. The second line in this example gives two classes for this entity:
<http://bibframe.org/resources/FkP1398705387/8929207instance22>
a bf:Instance, bf:Monograph .

One thing that classes do not do, however, is to prevent your "thing" from being assigned the "wrong class." You can, however, define your vocabulary to make "wrong classes" apparent. To do this you define certain classes as disjoint, for example a class of "dead" would logically be disjoint from a class of "alive." Disjoint means that the same thing cannot be of both classes, either through the direct declaration of "type" or through the assignment of properties. Let's do an example:
"residence" has domain "Alive"
"cemetery plot location" has domain "Dead"
"Alive" is disjoint "Dead" (you can't be both alive and dead)

X - is type - "Alive"                                         (X is of class "Alive")
X - cemetery plot location - URI:9494747      (X is of class "Dead")
Nothing stops you from creating this contradiction, but some applications that try to use the data will be stumped because you've created something that, in RDF-speak, is logically inconsistent. What happens next is determined by how your application has been programmed to deal with such things. In some cases, the inconsistency will mean that you cannot fulfill the task the application was attempting. If you reach a decision point where "if Alive do A, if Dead do B" then your application may be stumped and unable to go on.

All of this is to be kept in mind for the next blog post, which talks about the effect of class definitions on bibliographic data in RDF.


Note: Multiple domains are treated in RDF as an AND (an intersection). Using a library-ish example, let's assume you want to define a note field that you can use for any of your bibliographic entities. For this example, we'll define entities Work, Person, Manifestation for ex:note. You define your note property something like:

ex:note
    a rdf:Property ;
    rdfs:label "Note"@en ;
    rdfs:domain ex:Work ;
    rdfs:domain ex:Person ;
    rdfs:domain ex:Manifestation ;
    rdfs:range rdfs:literal .

Any subject on which you use ex:note would be inferred to be, at the same time, Work and Person and Manifestation - which is manifestly illogical. There is not a way to express the rule: "This property CAN be used with these classes" in RDF. For that, you will need something that does not yet exist in RDF, but is being worked on in the W3C community, which is a set of rules that would allow you to validate property usage. You might also want see what schema.org has done for domain and range.

Saturday, October 25, 2014

Citations get HOT

The Public Library of Science research section, PLOSLabs (ploslabs.org) has announced some very interesting news about the work that they are doing on citations, which they are calling "Rich Citations".

Citations are the ultimate "linked data" of academia, linking new work with related works. The problem is that the link is human-readable only and has to be interpreted by a person to understand what the link means. PLOS Labs have been working to make those citations machine-expressive, even though they don't natively provide the information needed for a full computational analysis.

Given what one does have in a normal machine-readable document with citations, they are able to pull out an impressive amount of information:
  • What section the citation is found in. There is some difference in meaning whether a citation is found in the "Background" section of an article, or in the "Methodology" section. This gives only a hint to the meaning of the citation, but it's more than no information at all.
  • How often a resource is cited in the article. This could give some weight to its importance to the topic of the article.
  • What resources are cited together. Whenever a sentence ends with "[3][7][9]", you at least know that those three resources equally support what is being affirmed. That creates a bond between those resources.
  • ... and more
As an open access publisher, they also want to be able to take users as directly as possible to the cited resources. For PLOS publications, they can create a direct link. For other resources, they make use of the DOI to provide links. Where possible, they reveal the license of cited resources, so that readers can know which resources are open access and which are pay-walled.

This is just a beginning, and their demo site, appropriately named "alpha," uses their rich citations on a segment of the PLOS papers. They also have an API that developers can experiment with.

I was fortunate to be able to spend a day recently at their Citation Hackathon where groups hacked on ongoing aspects of this work. Lots of ideas floated around, including adding abstracts to the citations so a reader could learn more about a resource before retrieving it. Abstracts also would add search terms for those resources not held in the PLOS database. I participated in a discussion about coordinating Wikidata citations and bibliographies with the PLOS data.

Being able to datamine the relationships inherent in the act of citation is a way to help make visible and actionable what has long been the rule in academic research, which is to clearly indicate upon whose shoulders you are standing. This research is very exciting, and although the PLOS resources will primarily be journal articles, there are also books in their collection of citations. The idea of connecting those to libraries, and eventually connecting books to each other through citations and bibliographies, opens up some interesting research possibilities.

Sunday, October 19, 2014

schema.org - where it works

In the many talks about schema.org, it seems that one topic that isn't covered, or isn't covered sufficiently, is "where do you do it?" That is, where does it fit into your data flow? I'm going to give a simple, typical example. Your actual situation may vary, but I think this will help you figure out your own case.

The typical situation is that you have a database with your data. Searches go against that database, the results are extracted, a program formats these results into a web page, and the page is sent to the screen. Let's say that your database has data about authors, titles and dates. These are stored in your database in a way that you know which is which. A search is done, and let's say that the results of the search are:
author:  Williams, R
title: History of the industrial sewing machine
date: 1996
This is where you are in your data flow:

The next thing that happens (and remember, I'm speaking very generally) is that the results then are fed into a program that formats them into HTML, probably within a template that has all your headers, footers, sidebars and branding and sends the data to the browser. The flow now looks like

Let's say that you will display this as a citation, that looks like:
Williams, R. History of the industrial sewing machine. 1996.
Without any fancy formatting, the HTML for this is:
<p>Williams, R. History of the industrial sewing machine. 1996.</p>
Now we can see the problem that schema.org is designed to fix. You started with an author, a title and date, but what you are showing to the world is a string of characters are that undifferentiated. You have lost all the information about what these represent. To a machine, this is just another of many bazillions of paragraphs on the web. Even if you format your data like this:
<p>Author: Williams, R.</p>
<p>Title:  Williams, R. History of the industrial sewing machine</p>
<p>Date: 1996</p>
What a machine sees is:
<p>blah: blah</p>
<p>blah: blah</p>
<p>blah: blah</p>  
What we want is for the program that is is formatting the HTML to also include some metadata from schema.org that retains the meaning of the data you are putting on the screen. So rather than just putting HTML formatting, it will add formatting from schema.org. Schema.org has metadata elements for many different types of data. Using our example, let's say that this is a book, and here's how you could mark that up in schema.org:
<div vocab="http://schema.org/">
<div   typeof="Book">
<p>
    <span property="author">Williams, R.</span> <span property="name">History of the industrial sewing machine</span>. <span property="datePublished">1996</span>.
    </p>
    </div>
</div>
Again, this is a very simple example, but when we test this code in the Google Rich Snippet tool, we can see that even this very simple example has added rich information that a search engine can make use of:
To see a more complex example, this is what Dan Scott and I have done to enrich the files of the Bryn Mawr Classical Reviews.

The review as seen in a browser (includes schema.org markup)

The review as seen by a tool that reads the structured schema.org data.

From these you can see a couple of things. The first is that the schema.org markup does not change how your pages look to a user viewing your data in a browser. The second is that hidden behind that simple page is a wealth of rich information that was not visible before.

Now you are probably wondering: well, what's that going to do for me? Who will use it? At the moment, the users of this data are the search engines, and they use the data to display all of that additional information that you see under a link:


In this snippet, the information about stars, ratings, type of film and audience comes from schema. org mark-up on the page.

Because the data is there, many of us think that other users and uses will evolve. The reverse of that is that, of course, if the information isn't there then those as yet undeveloped possibilities cannot happen.



Wednesday, October 01, 2014

This is what sexism looks like

[Note to readers: sick and tired of it all, I am going to report these "incidents" publicly because I just can't hack it anymore.]

I was in a meeting yesterday about RDF and application profiles, in which I made some comments, and was told by the co-chair: "we don't have time for that now", and the meeting went on.

Today, a man who was not in the meeting but who listened to the audio sent an email that said:
"I agree with Karen, if I correctly understood her point, that this is "dangerous territory".  On the call, that discussion was postponed for a later date, but I look forward to having that discussion as soon as possible because I think it is fundamental."
And he went on to talk about the issue, how important it is, and at one point referred to it as "The requirement is that a constraint language not replace (or "hijack") the original semantics of properties used in the data."

The co-chair (I am the other co-chair, although reconsidering, as you may imagine) replied:
"The requirement of not hijacking existing formal specification languages for expressing constraints that rely on different semantics has not been raised yet."
"Has not been raised?!" The email quoting me stated that I had raised it the very day before. But an important issue is "not raised" until a man brings it up. This in spite of the fact that the email quoting me made it clear that my statement during the meeting had indeed raised this issue.

Later, this co-chair posted a link to a W3C document in an email to me (on list) and stated:
"I'm going on holidays so won't have time to explain you, but I could, in theory (I've been trained to understand that formal stuff, a while ago)"
That is so f*cking condescending. This happened after I quoted from W3C documents to support my argument, and I believe I had a good point.

So, in case you haven't experienced it, or haven't recognized it happening around you, this is what sexism looks like. It looks like dismissing what women say, but taking the same argument seriously if a man says it, and it looks like purposely demeaning a woman by suggesting that she can't understand things without the help of a man.

I can't tell you how many times I have been subjected to this kind of behavior, and I'm sure that some of you know how weary I am of not being treated as an equal no matter how equal I really am.

Quiet no more, friends. Quiet no more.

(I want to thank everyone who has given me support and acknowledgment, either publicly or privately. It makes a huge difference.) 

Some links about "'Splaining"
http://scienceblogs.com/thusspakezuska/2010/01/25/you-may-be-a-mansplainer-if/
http://geekfeminism.wikia.com/wiki/Splaining

Tuesday, September 23, 2014

The book you scroll

I was traveling in Italy where I spend a lot of time in bookstores. I'm looking not only for books to read, but to discover new authors, since Italian bookstores are filled with translations of authors that I rarely see in the few bookstores remaining in my home town of Berkeley, CA. While there I came across something that I find fascinating: the flipback book. These books are small - the one I picked up is about 4 3/4" x 3 1/4". It feels like a good-sized package of post-it notes in your hand.

 From the outside, other than its size, looks "normal" although the cover design is in landscape rather than portrait  position.

The surprise is when you open the book. The first thing you notice is that you read the book top-to-bottom across two pages. It's almost like scrolling on a web page, because you move the pages up, not across.

The other thing is that they are incredibly compact. The paper is thin, and some of the books contained entire trilogies, although only about 2 - 2.5 inches thick.

Because there is no gutter between the two pages, you essentially get a quantity of text that is equal to what you get on a regular book page. Oddly, the two contiguous pages are numbered as separate pages, although only the odd numbers actually printed, so you have pages 37, 39, 41, etc. However, the actual number of open pages is about the same as the paperback book.

The font is a sans serif, similar to many used online, so that whole thing feels like a paper book imitating a computer screen.

I haven't read the book yet, so I don't know if the reading experience is pleasing. But I am amazed that someone has found a way to reinvent the print book after all these years. Patented, of course.

There are few titles available yet, but an Amazon search on "flipback" brings up a few.

Thursday, September 11, 2014

Philosophical Musings: The Work

We can't deny the idea of work - opera, oeuvre - as a cultural product, a meaningful bit of human-created stuff. The concept exists, the word exists. I question, however that we will ever have, or that we should ever have, precision in how works are bounded; that we'll ever be able to say clearly that the film version of Pride and Prejudice is or is not the same work as the book. I'm not even sure that we can say that the text of Pride and Prejudice is a single work. Is it the same work when read today that it was when first published? Is it the same work each time that one re-reads it? The reading experience varies based on so many different factors - the cultural context of the reader; the person's understanding of the author's language; the age and life experience of the reader.

The notion of work encompasses all of the complications of human communication and its consequent meaning. The work is a mystery, a range of possibilities and of possible disappointments. It has emotional and, at its best, transformational value. It exists in time and in space. Time is the more canny element here because it means that works intersect our lives and live on in our memories, yet as such they are but mere ghosts of themselves.

Take a book, say, Moby Dick; hundreds of pages, hundreds of thousands of words. We read each word, but we do not remember the words -- we remember the book as inner thoughts that we had while reading. Those could be sights and smells, feelings of fear, love, excitement, disgust. The words, external, and the thoughts, internal, are transformations of each other; from the author's ideas to words, and from the words to the reader's thoughts. How much is lost or gained during this process is unknown. All that we do know is that, for some people at least, the experience is vivid one. The story takes on some meaning in the mind of the reader, if one can even invoke the vague concept of mind without torpedoing the argument altogether.

Brain scientists work to find the place in the maze of neuronic connections that can register the idea of "red" or "cold" while outside of the laboratory we subject that same organ to the White Whale, or the Prince of Denmark, or the ever elusive Molly Bloom. We task that organ to taste Proust's madeleine; to feel the rage of Ahab's loss; to become a neighbor in one of Borges' villages. If what scientists know about thought is likened to a simple plastic ping-pong ball, plain, round, regular, white, then a work is akin to a rainforest of diversity and discovery, never fully mastered, almost unrecognizable from one moment to the next.

As we move from textual works to musical ones, or on to the visual arts, the transformation from the work to the experience of the work becomes even more mysterious. Who hasn't passed quickly by an unappealing painting hanging on the wall of a museum before which stands another person rapt with attention. If the painting doesn't speak to us, then we have no possible way of understanding what it is saying to someone else.

Libraries are struggling to define the work as an abstract but well-bounded, nameable thing within the mass of the resources of the library. But a definition of work would have to be as rich and complex as the work itself. It would have to include the unknown and unknowable effect that the work will have on those who encounter it; who transform it into their own thoughts and experiences. This is obviously impractical. It would also be unbelievably arrogant (as well as impossible) for libraries to claim to have some concrete measure of "workness" for now and for all time. One has to be reductionist to the point of absurdity to claim to define the boundaries between one work and another, unless they are so far apart in their meaning that there could be no shared messages or ideas or cultural markers between them. You would have to have a way to quantify all of the thoughts and impressions and meanings therein and show that they are not the same, when "same" is a target that moves with every second that passes, every synapse that is fired.

Does this mean that we should not try to surface workness for our users? Hardly. It means that it is too complex and too rich to be given a one-dimensional existence within the current library system. This is, indeed, one of the great challenges that libraries present to their users: a universe of knowledge organized by a single principle as if that is the beginning and end of the story. If the library universe and the library user's universe find few or no points of connection, then communication between them fails. At best, like the user of a badly designed computer interface, if any communication will take place it is the user who must adapt. This in itself should be taken the evidence of superior intelligence on the part of the user as compared to the inflexibility of the mechanistic library system.

Those of us in knowledge organization are obsessed with neatness, although few as much as the man who nearly single-handled defined our profession in the late 19th century; the man who kept diaries in which he entered the menu of every meal he ate; whose wedding vows included a mutual promise never to waste a minute; the man enthralled with the idea that every library be ordered by the simple mathematical concept of the decimal.

To give Dewey due credit, he did realize that his Decimal Classification had to bend reality to practicality. As the editions grew, choices had to be made on where to locate particular concepts in relation to others, and in early editions, as the Decimal Classification was used in more libraries and as subject experts weighed in, topics were relocated after sometimes heated debate. He was not seeking a platonic ideal or even a bibliographic ideal; his goal was closer to the late 19th century concept of efficiency. It was a place for everything, and everything in its place, for the least time and money.

Dewey's constraints of an analog catalog, physical books on physical shelves, and a classification and index printed in book form forced the limited solution of just one place in the universe of knowledge for each book. Such a solution can hardly be expected to do justice to the complexity of the Works on those shelves. Today we have available to us technology that can analyze complex patterns, can find connections in datasets that are of a size way beyond human scale for analysis, and can provide visualizations of the findings.

Now that we have the technological means, we should give up the idea that there is an immutable thing that is the work for every creative expression. The solution then is to see work as a piece of information about a resource, a quality, and to allow a resource to be described with as many qualities of work as might be useful. Any resource can have the quality of the work as basic content, a story, a theme. It can be a work of fiction, a triumphal work, a romantic work. It can be always or sometimes part of a larger work, it can complement a work, or refute it. It can represent the philosophical thoughts of someone, or a scientific discovery. In FRBR, the work has authorship and intellectual content. That is precisely what I have described here. But what I have described is not based on a single set of rules, but is an open-ended description that can grow and change as time changes the emotional and informational context as the work is experienced.

I write this because we risk the petrification of the library if we embrace what I have heard called the "FRBR fundamentalist" view. In that view, there is only one definition of work (and of each other FRBR entity). Such a choice might have been necessary 50 or even 30 years ago. It definitely would have been necessary in Dewey's time. Today we can allow ourselves greater flexibility because the technology exists that can give us different views of the same data. Using the same data elements we can present as many interpretations of Work as we find useful. As we have seen recently with analyses of audio-visual materials, we cannot define work for non-book materials identically to that of books or other texts. [1] [2] Some types of materials, such as works of art, defy any separation between the abstraction and the item. Just where the line will fall between Work and everything else, as well as between Works themselves, is not something that we can pre-determine. Actually, we can, I suppose, and some would like to "make that so", but I defy such thinkers to explain just how such an uncreative approach will further new knowledge.

[1] Kara Van Malssen. BIBFRAME A-V modeling study
[2] Kelley McGrath. FRBR and Moving Images

Thursday, September 04, 2014

WP:NOTABILITY (and Women)

I've been spending quite a bit of time lately following the Wikipedia pages of "Articles for Deletion" or WP:AfD in Wikipedia parlance. This is a fascinating way to learn about the Wikipedia world. The articles for deletion fall mostly into a few categories:
  1. Brief mentions of something that someone once thought interesting (a favorite game character, a dearly loved soap opera star, a heartfelt local organization) but that has not been considered important by anyone else. In Wikipedian, it lacks WP:NOTABILITY.
  2. Highly polished P.R. intended to make someone or something look more important than it is, knowing that Wikipedia shows up high on search engine results, and that any site linked to from Wikipedia also gets its ranking boosted.
Some of #2 is actually created by companies that are paid to get their clients into Wikipedia along with promoting them in other places online. Another good example is that of authors of self-published books, some of whom appear to be more skilled in P.R. than they are in the literary arts.

In working through a few of the fifty or more articles proposed for deletion each day, you get to do some interesting sleuthing. You can see who has edited the article, and what else they have edited; any account that has only edited one article could be seen as a suspected bogus account created just for that purpose. Or you could assume that only one person in the English-speaking world has any interest in this topic at all.

Most of the work, though, is in seeing if you can establish notability. Notability is not a precise measure, and there are many pages of policy and discussion on the topic. The short form is that for something or someone to be notable, it has to be written about in respected, neutral, third-party publications. Thus a New York Times book review is good evidence of notability for a book, while a listing in the Amazon book department is not. The grey area is wide, however. Publisher's Weekly may or may not indicate notability, since they publish only short paragraphs, and cover about 7,000 books a year. That's not very discriminating.

Notability can be tricky. I recently came across an article for deletion pointing to Elsie Finnimore Buckley, a person I had never heard of before. I discovered that her dates were 1882-1959, and she was primarily a translator of works from French into English. She did, though, write what appears to have been a popular book of Greek tales for young people.

As a translator, her works were listed under "E. F. Buckley." I can well imagine that if she had used her full name it would not have been welcome on the title page of the books she translated. Some of the works she translated appear to have a certain stature, such as works by Franz Funck-Brentano. She has an LC name authority file under "Buckley, E. F." although her full name is added in parentheses: "(Elsie Finnimore)".

To understand what it was like for women writers, one can turn to Linda Peterson's book "Becoming a Woman of Letters and the fact of the Victorian market." In that, she quotes a male reviewer of Buckley's Greek tales, which she did publish under her full name. His comments are enough to chill the aspirations of any woman writer. He said that writing on such serious topics is "not women's work" and that "a woman has neither the knowledge nor the literary tact necessary for it." (Peterson, p. 58) Obviously, her work as a translator is proof otherwise, but he probably did not know of that work.

Given this attitude toward women as writers (of anything other than embroidery patterns and luncheon menus) it isn't all that surprising that it's not easy to establish WP:NOTABILITY for women writers of that era. As Dale Spender says in "Mothers of the Novel; 100 good women writers before Jane Austen":
"If the laws of literary criticism were to be made explicit they would require as their first entry that the sex of the author is the single most important factor in any test of greatness and in any preservation for posterity." (p. 137)
That may be a bit harsh, but it illustrates the problem that one faces when trying to rectify the prejudices against women, especially from centuries past, while still wishing to provide valid proof that this woman's accomplishments are worthy of an encyclopedia entry.

We know well that many women writers had to use male names in order to be able to publish at all. Others, like E.F. Buckley, hid behind initials. Had her real identity been revealed to the reading public, she might have lost her work as a translator. Of late, J.K. Rowling has used both techniques, so this is not a problem that we left behind with the Victorian era. As I said in the discussion on Wikipedia:
"It's hard to achieve notability when you have to keep your head down."

Saturday, June 28, 2014

Linking, really linking

Some time ago I posted a bit of wishful thinking about linking search engine results to library holdings. Now Overdrive has made this a reality, at least in Bing:

This appears in the "extended information" area of a Bing search for the Girl with a Dragon Tattoo trilogy. This is based on you having an Overdrive account and I believe you may also have to have given the browser permission to use your location.




I don't usually use Bing, and so I was unaware that Bing has made much better use of linked data (in part promoted by the use of schema.org standards) than Google. Here is the Google extended sidebar for the same book:
Google results
Now look at what Bing provides:
Bing results
 
If you can't see that there are advantages to linked data after looking at these examples, then, like the global warming deniers, you just don't want to be confused by the facts. Now to get on to how we can make library offerings as rich as this.

**HT to Eric Hellman for blogging this from ALA.

Saturday, May 10, 2014

Let's link!

Tom Johnson responded to a statement of mine in which I said that we need non-programmer tools for linked data work. He asked for case studies, and so I'm going to do some "off the top of my head" riffing here, just to see what might come of it.

First, let me say a little something about some analogous technology. Let's take HTML. I'm one of those folks who learned HTML many decades ago when all you needed was <p>, <i>, <b> and maybe <hr>. With these, you could create a web page. Web pages in those days didn't have banners, sidebars, tables ... adding an image was going all fancy. There were no WYSIWYG tools because it was very simple to create such a web page. Then we got more sophisticated formatting (server-side includes, side bars, tables(!)), and now there is a whole programming language called CSS to handle page creation (and destruction, since CSS is very complex.) It doesn't take much looking around to understand that most people today need TOOLS to create a web page and lots of tools exist, tools that cost little or nothing (WordPress, Drupal) and can be used by folks who've never written 50 lines of working code.

Essentially, on the web today, a few people are providing the structure and the tools, but most folks are providing content without knowing the guts of the technology. Content is, in my mind, the actual goal of the web; technology is the means.

I've spent a lot of time talking to folks about linked data, but linked data is not itself a goal. I've started trying to move the conversation from the underlying technology to what I see as the real goal: making connections -- connections between concepts, ideas, statements about things. This is inherently more social than technical, but of course it needs the technology behind it in order to work. The easier it is for people to make connections, the more connections they will make.

The problem that I'm seeing today in the linked data space is that we don't have an idea of what kinds of connections people will want to make, or what they will do with them. I don't think we're going to know until we apply some scientific method to the problem -- that is, try, fail, try again, rinse, repeat.

I created a really funky web page with one idea. The page looks like this:



It talks about having the ability to profile types of data that can then be selected in the content of the page. Each of these will then pull additional content (maps, term definitions, biographical information)  from available linkable data into the page. This could begin as a very simple application with only a small number of options, as a proof of concept. (WordPress?) I'm sure that others can greatly improve on it.  Take a look at the display in the FAO AGRIS catalog to get an idea of how this might look with a better display.

Have at it, please.

Friday, April 25, 2014

Works, Expressions, and the Bibliographic Universe

In 1545, Conrad Gessner set about to create a bibliography of everything ever published, the Bibliotheca Universalis. He nearly succeeded. In 2004, Google set about to create a digital universal library by digitizing millions of books from library shelves. Shortly thereafter the Internet Archive began a project, the Open Library, to create "One web page for every book ever written." Meanwhile, OCLC's WorldCat has grown to over 300 million records from 72,000 libraries. All of these exemplify the concept of some universal bibliography or database that make up what I am calling, for now, the bibliographic universe.

The other side of this coin is the catalog of an individual library (or library consortium). This bibliographic data set is expressly designed to be local, not universal. It does not include materials not available to the library user, and what it does include has been selected with that library's constituency in mind.

The FRBR study is a conceptual view of this "bibliographic universe"; that is, an abstract view of intellectual resources and their relations to each other. To say that Magic Mountain is a translation of Der Zauberberg is to make a statement that is unrelated to any actual example of that text, much less the holdings within any individual library. We can conclude that FRBR addresses bibliographic data universally because it claims to be agnostic to any particular usage of the bibliographic concept, and to be able to define any and all bibliographic resources and their relationships.

The problem that arises, as I see it, is where this universal view meets the individual library. Do you express these relationships between items in your library, or do you express the relationships in the bibliographic universe? If your library has Magic Mountain in English and Spanish, but not in German, should you organize your presentation of data around the original German version, even if your users will not be able to read or understand the title in that language? What if you only have the English version - should this be displayed as a manifestation of a translation of the original German text? Is there a purpose to creating bibliographic information about relationships between works, even if the library does not hold those works?

A key question that we need to ask is: what is the purpose of providing these bibliographic relationships? The answer that I believe I would get from catalogers today is that the purpose of this information is to provide an organized view of the library to the user. While in a large library, works and expressions may be organizing principles, in a small library with often only one version of the resource, the addition of information about the work and the expression could be more disorganization than organization, because it doesn't give the user a more organized view of the library's holdings but adds information that may be confusing. This doesn't mean that the information about works and expressions isn't useful - the work and expression could be used to provide interesting information to the user, but in this case they do not provide a useful organizing function in the catalog.

As far as I can tell, neither FRBR nor the cataloging rules (past and present) clearly differentiate between organizing a library, organizing a larger context within which the library operates, or describing the bibliographic universe. I'll accept that organizing the bibliographic universe is probably out of scope in today's world, but where does one draw the line between the individual library and the larger useful bibliographic context that might be useful to your library users?

This is not a new problem. AACR2 introduced the idea of the work by adding the uniform title to the catalog record. The uniform title turned out to be less uniformly used than perhaps was intended because it was not applied by catalogers to all records where it was applicable. In a library with only one (or a small number) of editions of a work, the uniform title was deemed to be either unnecessary or not worth the time of the cataloger. This worked fine in individual libraries but caused problems for sharing and in union catalogs. It also makes it more difficult to move from today's cataloging records to a FRBR-ized catalog, since the essential clues about works are not provided consistently.

FRBR is a conceptual model, but it isn't clear to me what context it is modeling: a library catalog, the conceptual collective of all or most library catalogs, or the bibliographic universe. The original task was to model the essential things of a library bibliographic record to respond to a set of user tasks.  However, at the 30,000 foot level at which FRBR operated, the questions about how one would serve users in a particular library is left open. The FRBR user tasks are a look at the existing concept of a library catalog and what one mythical "user" does when approaching it. It also is a look from the point of view of a large library catalog: no one on the study group was from a small or even medium-sized library. FRBR is very much a top-down look at the bibliographic world. If we look at the library bibliographic world from the bottom up -- either looking from the point of view of individual users, or of individual libraries, then we would need to see the FRBR concepts as possibilities, not requirements; possibilities to be used as appropriate for ones particular situation.

We know that a given library serves its particular users, not the "universe of users." The best service for a library's users is to allow the library to make choices that are appropriate for those users. For that reason, requiring libraries to present, in their catalogs, data that has the bibliographic universe as its context is going to be detrimental to library service. At the same time, it would be ideal to have a true catalog of the bibliographic universe available from which a library could draw information or could create links as a way to expand its catalog information for users who need more. For example, the user who looks up the Chicago Manual of Style should be able to learn if the copy the library holds is the latest version. The user looking for "Harry Potter" and seeing that the library has copies in English and Spanish should be able to ask if the book was translated into Vietnamese (yes) or Tegulu (no).

It would be naive to say that we have no use for a bibliotheca universalis. However, a bibliotheca universalis is not a library catalog. It would also be naive to say that every library has the same needs regarding bibliographic data. What we seem to be lacking is the way to bridge the gap between the 30,000 foot FRBR bibliographic view and the needs of the individual library. I think we have the technology to do this today, and some of the possible answers can be found in general databases like WorldCat or DBPedia. It's the connection between these that needs to be designed.

Tuesday, April 01, 2014

FRBR group 1: the gang of four

(This is a very delayed follow on to my earlier  post on FRBR groups 2 and 3. It's not that I haven't been thinking about it... and I hope soon to be able to post my talk from FSR2014 on FRBR, as well.) 

Parts vs. views

Each of the three FRBR groups is defined briefly in the introduction to section 3 of the FRBR document. The second and third groups have fairly concrete definitions:
group 2 "...those responsible for the intellectual or artistic content, the physical production and dissemination, or the custodianship of the entities in the first group"

group 3 "...an additional set of entities that serve as the subjects of works"
The definition of Group 1 is more complex and considerably less clear:
"The entities in the first group (as depicted in Figure 3.1) represent the different aspects of user interests in the products of intellectual or artistic endeavour." [FRBR, p. 13]
Where groups 2 and 3 are made up of similar but independent things (which is a common definition of a class of things), group 1 consists of aspects of a single thing ("intellectual or artistic endeavor"). The term "aspects" can be defined as either parts of something or points of view about something. The difference between "parts" vs. "points of view" is important. Parts could be defined as simple, observable facts, such as the parts of a particular automobile (motor, chassis, wheels). These are characteristics of the thing itself, independent of the observer. Points of view, of course, vary for each viewer and perhaps each viewing. This would fit with the FRBR document's statement on the work:
"The concept of what constitutes a work and where the line of demarcation lies between one work and another may in fact be viewed differently from one culture to another. Consequently the bibliographic conventions established by various cultures or national groups may differ in terms of the criteria they use for determining the boundaries between one work and another." [FRBR p. 17]
That the FRBR document states that the entities are aspects of user interests rather than aspects of an intellectual endeavor implies that the entities of group 1 are not parts of the endeavor, but constructions in the minds of users. From the remainder of the FRBR document, in particular the areas where the attributes are defined for each entity, it is clear that the FRBR Study Group chose to define the bibliographic description of intellectual endeavors as a single point of view. For each entity, the Study Group has a provided a set of elements that are each defined for only that one entity, with no concession made for different points of view or interests. This is, however, in spite of the statement above that communities may have different views.

To reinforce the view of group 1 as parts of a whole, there exist dependencies between the group 1 entities such that, with the exception of work, each can only exist in combination with certain others to which it is linked. Therefore none represents a whole on its own. (In fact, there is no concept of a whole bibliographic description in FRBR. That would need to come from a different analysis.) The definitions of the entities express these dependencies.
"work: a distinct intellectual or artistic creation." [FRBR, p.17]
"expression: the intellectual or artistic realization of a work in the form of alpha-numeric, musical, or choreographic notation, sound, image, object, movement, etc., or any combination of such forms." [FRBR p. 19]
"manifestation: the physical embodiment of an expression of a work." [FRBR p. 21]
"item: a single exemplar of a manifestation" [FRBR p. 24]
as does the description of the cardinality of the relationships:
"The relationships depicted in the diagram indicate that a work may be realized through one or more than one expression (hence the double arrow on the line that links work to expression). An expression, on the other hand, is the realization of one and only one work (hence the single arrow on the reverse direction of that line linking expression to work). An expression may be embodied in one or more than one manifestation; likewise a manifestation may embody one or more than one expression. A manifestation, in turn, may be exemplified by one or more than one item; but an item may exemplify one and only one manifestation." [FRBR pp. 13-14]
This directionality, or fixed order, of the dependencies is the source of the image of group 1 as a hierarchy, where each entity connects to the entity "above" it. But there is more than one interpretation of these definitions. Taniguchi [taniguchi] reads the description of the entities as a "Russian doll" with each succeeding entity containing the previous ones. In the definitions of expression, manifestation, and item, above, each entity appears to encapsulate the one or ones above it in the diagram ("the physical embodiment of an expression of a work"). When diagrammed, this view would look like:

(Note that this does not exclude the one-to-many and many-to-many relationships as long as both expressions and manifestations can be part of more than one nested structure.)

The other interpretation, which is the most common interpretation of the entity-relation diagrams, is similar to a database design where each entity represents a single set of data elements that can be shared in one-to-many or many-to-many relationships. This view presents the four aspects as separate entities with strict relationships between them.

I perceive a contradiction between the verbal definitions in the document and the diagrams, which one presumes are intended to represent the information in the text. The decision to represent the group 1 entities as separate parts and without any overlap in data elements is a conceptual reduction of the definitions that are given early in the document, and no where does the document state that such a decision was made. There could be good reasons to implement the FRBR group 1 concepts in a particular technology as a simplified structure, but it is clear to me that the model in the diagrams is not as rich as the concepts in the text would allow.

Process

Some interpretations of FRBR treat the work, expression and manifestation as a process or continuum, moving from the idea in the creator’s mind, to an expression of that creation, and then to a manifestation where the expression becomes "manifest" or physical in nature.
"Content relationships can be viewed as a continuum from works/expressions/manifestations/items. Moving left to right along this continuum we start with some original work and related works and expressions and manifestations that can be considered ‘equivalent,’ that is, they share the same intellectual or artistic content as realized through the same mode of expression." [tillett p. 4]
The FRBR group 1 "continuum of entities" runs into problems when faced with the reality of publishing and packaging. While the line from work to expression to manifestation may follow some ideal logic, it may have been more functional to separate the description of the package from its intellectual contents. Instead, manifestation, as described in FRBR, is still based on the traditional catalog entry that mixes content and carrier by including creators, titles, and edition information, which better fit the definition of expression than manifestation.

Most explanations of work, expression, manifestation, item (WEMI) move from work to expression, then to manifestation, in that order, and most give only a slight nod to item. But in terms of cataloger workflow, WEM is a single unit that is encountered with the item in hand. While you may be able to store information about a work or expression separately in a database, you cannot separate the work from the expression or the manifestation in real life.

Agency

FRBR provides a static view of the bibliographic resource with little agency. The entities simply exist, they are not described as created as the result of an action. In fact, the entities seem to be actors themselves, as when the expression "realizes" the work. It would make more sense to say that the expression entity is the realization of the work, and that some sentient being acts to create the expression. Instead, in FRBR some unnamed magic occurs between the work and the expression. The same is true of the manifestation, which should be the result of some action that produces a physical manifestation of the expression of the work.

This static view is compatible with library cataloging, which is mainly interested in describing the item in hand as a single unit. The development of a model that emphasizes relationships between creative outputs begs for a more actor-centered view of the bibliographic universe. One could argue that it is precisely the intervention of specific actors that creates a differentiation between entities. The same music piece performed by different musicians, or the same musicians at a different time, must be a different expression. The "studio cut" and the "director's cut" of a film are either different expressions or different works (depending on your definition of work), based on the agent in control. Adding agent intervention to the model could be useful in developing clearer rules for the determination of separate entities during the cataloging process.

Conclusion

While FRBR groups 2 and 3 are composed of real world things (in the semantic web sense), group 1 appears to be an analysis of the current data of bibliographic records. The division of attributes into the four "boxes" of WEMI does not introduce new data elements but partitions the existing bibliographic record among the entities. The resulting group 1 picture resembles ISBD rather than AACR/MARC in that it is a static view of a bibliographic "done deal" with no indication of agents or subjects. Others have noticed that there are neither creator no subject attributes among the listed attributes for the work -- instead, these are included in the model as relationships defined between groups 2 and 3 and group 1. This is a logical outcome of the use of the database design methodology where data is stored for subsequent use but is not part of a data creator or data user workflow.

In the past bibliographic description has been unitarian, with one record representing one, indivisible bibliographic thing. FRBR posits a quatritarian view of the same data. The difficulty, however, is that the FRBR group 1 is not like the division of an automobile into chassis, motor and wheels; instead, where one draws the line between the separate aspects of the FRBR quaternity -- or whether one prefers a unitarian, duotarian or even a quintitarian approach -- is not based on empirical data, but on one's particular point of view. That point of view is not arbitrary, but has many factors based on material type, organization type, and the needs of the users. FRBR's four-part bibliographic description is one possibility. It may represent a particular bibliographic view, but one cannot expect that it represents all bibliographic views, either in libraries or beyond them.


[FRBR] IFLA Study Group on the Functional Requirements for Bibliographic Records. (2009). Functional Requirements for Bibliographic Records. Retrieved from http://archive.ifla.org/VII/s13/frbr/frbr_2008.pdf

[taniguchi] Shoichi Taniguchi. “A conceptual model giving primacy to expression-level bibliographic entity in cataloging”, Journal of Documentation, Vol. 58 Iss: 4, 2002. pp.363 – 382. http://dx.doi.org/10.1108/00220410210431109

[tillett] Tillett, B. What is FRBR? A conceptual model for the bibliographic universe. (p. 8). Washington, DC. 2003. http://www.loc.gov/cds/downloads/FRBR.PDF

Monday, February 24, 2014

The FRBR Groups

FRBR has three groups of entities, numbered 1-3. Each group, however, has its own set of characteristics that are very different from each other, so different that they really are different kinds of groups. These differences make it hard to speak of them in one breath.

One of the key things to know about the groups is that they aren't classes in the data modeling sense. Why they are therefore grouped at all is not clear, except perhaps it was a convenient way to speak of them. The IFLA FRBR Review Group maintains that the groups do not represent classes and that the ten (or eleven, with family) entities represent the highest organizational level recognized by FRBR. Unfortunately the treatment of them as groups throughout the document tends to contradict this statement. This just adds to the confusion about the meaning and nature of the groups.

Group 2

I'm going to take Group 2 first because it is the simplest. Group 2 is a group of "agents" or "actors" that perform actions in the bibliographic environment. The original entities of the group were person and corporate body, but family has been added through the work that was done on the Functional Requirements for Authority Data (FRAD). In most kinds of data modeling the members of Group 2 would be members of a class, and the class would have certain characteristics that define the kinds of things that could be members and their shared characteristics. For example, one could say that all members are people or groups of people, that they generally have names, they perform certain actions, etc. These characteristics would therefore not need to be defined separately for each member of the class, and definitions of members would only include characteristics unique to that class. Because no classes exist in FRBR, each Group 2 entity is described separately through its own attributes, although there is a fair amount of overlap between them.

Note that each of the Group 2 entities stands alone with no dependencies on any other entities. (This matters when we get to Group 1).

Group 3

Group 3 is an odd grouping because it has a rather miscellaneous nature. The entities that are described as Group 3 are ones that are needed for subject description in bibliographic data: subject; object; place; event. Not much is said about them because FRBR, not unlike cataloging rules (AACR, ISBD), does not really address subject assignment. It isn't clear to me where these particular entities come from because they are not equal to the "facets" of LCSH (form, chronological and geographic). It would be interesting to know how this particular set came about, since the FRBR Study Group was looking beyond North American practice.

What makes Group 3 odd, though, is not its composition but that it is only a partial listing of the subject elements; the full set also includes all of the members of groups 1 and 2. So the actual meaning of group 3 is: all of the subject entities that are not in other groups.

It remains to be seen what will happen to these entities when FRBR and the Functional Requirements for Subject Authority Data (FRSAD) are merged. FRSAD takes a 30,000 foot view of subjects, and essentially concludes that if you can call it a subject and give it an identity and a name, it's a subject. This aspect of the original FRBR study, which was specifically directed at the charge of defining the elements of a core bibliographic record, could change as the model becomes more generalized, which seems to be what is happening.

Group 1

All that I will say about Group 1 here is that it is a group that represents one thing divided into four levels of abstraction. Group 1 deserves its own post, rather than making this one overly long. That post will be next.

Wednesday, February 19, 2014

FRBR goals: entities, relations, and a core level record

The FRBR study was motivated by a 1990 international seminar on cataloging held in Stockholm. The charge to the study group was approved by the IFLA Standing Committee of the Section of Cataloguing in 1992. That document, called the Terms of Reference for a Study of the Functional Requirements for Bibliographic Records, stated:
Today the expectations and constraints facing bibliographic control are more pressing than ever. All libraries, including national bibliographic agencies, are operating under increasing budgetary constraints and increasing pressures to reduce cataloging costs through minimal level cataloging. [1]
Or, as Olivia Madison, the chair of the FRBR Study Group from 1991-1993 and 1995-1997, put it:
The Stockholm Seminar addressed the general question: "Can cataloging be considerably simplified?" [2]
The Standing Committee decided that consultants with particular skills in the area of cataloging were needed in order to approach the task, and three (later four) consultants were engaged. The primary charge to the consultants was:
1. Determine the full range of functions of the bibliographic record and then state the primary uses of the record as a whole.
This is at the very least a daunting task. However, the Terms of Reference gave the consultants some guidance about how to go about their work. The remaining tasks for the consultants were:
2. Develop a framework that identifies and clearly defines the full range of entities (e.g., work, texts, subjects, editions and authors) that are the subject of interest to users of a bibliographic record and the types of relationships (e.g. part/whole, derivative, and chronological) that may exist between those entities.
3. For each of the entities in the framework, identify and define the functions (e.g., to describe, to identify, to differentiate, to relate) that the bibliographic record is expected to perform.
4. Identify the key attributes (e.g., title, date, and size) of each entity or relationship that are required for each specific function to be performed. Attribute requirements should relate specifically to the media or format of the bibliographic item where applicable.
The notions of entities, relationships, and attributes don't appear in traditional cataloging theory; they come instead from the world of database design, and in particular relational database design. Because these concepts were expected to be unfamiliar to members of the committee and perhaps also the consultants, the Terms of Reference provides definitions, using as its source the 1984 book Data Analysis: the Key to Data Base Design, by Richard C. Perkinson. (Note, some of this is re-iterated in the FRBR final report, in the section on methodology, where four books are cited as sources of information on entity-relation methodology.)

Those were the tasks for the consultants, the selected experts who would do the analysis and present the report to the Study Group. The Study Group itself had this task:
5. For the National Libraries: for bibliographic records created by the national bibliographic agencies, recommend a basic level of functionality that relates specifically to the entities identified in the framework the functions that are relevant to each.
It appears to be this last charge that directly addressed the needs expressed in the Stockholm seminar: the need for a core level record that would help cataloging agencies reduce their costs while still serving users. I read the charges to the consultants as mainly providing a working methodology that would allow the consultants to focus  their energies on what amounts to a general rethinking of cataloging theory and practice.

The Terms of Reference is a rather bare bones statement of what needs to be done, and it says little about the why of the study. According to Tillett's 1994 report [3], some of the concerns that came out of Stockholm were:
"the mounting costs of cataloging," the proliferation of new media, "exploding bibliographic universe," the need to economize in cataloging, and "the continuing pressures to adapt cataloguing practices and codes to the machine environment."
The FRBR document states the motivation as:
"The purpose of formulating recommendations for a basic level national bibliographic record was to address the need identified at the Stockholm Seminar for a core level standard that would allow national bibliographic agencies to reduce their cataloguing costs through the creation, as necessary, of less-than-full-level records, but at the same time ensure that all records produced by national bibliographic agencies met essential user needs." [4] p.2
At this point, it is worth asking: did the FRBR study indeed result in a "core level standard" that would reduce cataloging costs for national bibliographic agencies? It definitely did define a core level standard, although that aspect of the FRBR report is not often discussed. Chapter 7 of the FRBR document, BASIC REQUIREMENTS FOR NATIONAL BIBLIOGRAPHIC RECORDS, lists the "basic level of functionality" for library catalogs:
Find all manifestations embodying:
  • the works for which a given person or corporate body is responsible
  • the various expressions of a given work
  • works on a given subject
  • works in a given series
Find a particular manifestation:
  • when the name(s) of the person(s) and/or corporate body(ies) responsible for the work(s) embodied in the manifestation is (are) known
  • when the title of the manifestation is known
  • when the manifestation identifier is known
Identify a work
Identify an expression of a work
Identify a manifestation
Select a work
Select an expression
Select a manifestation
Obtain a manifestation
This of course looks quite similar to the goals of a catalog developed over a century ago by Charles Ammi Cutter:
Section 7.3 of the chapter lists the descriptive and organizing elements (headings) that should make up a core bibliographic record. This chapter should be a key element of the FRBR study results, yet it isn't often mentioned in discussions of FRBR, which tend to focus on the ten (or eleven, if you add family) entities and their primary relationships to each other (is realization of, manifests, etc.), and the four user tasks (find, identify, select, obtain).

While most people can hold forth on the FRBR entities, few can discourse on this outcome of the report, which is a basic level national bibliographic record. Admittedly, the report itself does not emphasize this information. The elements of the basic level record use the terminology of ISBD, not of FRBR, which makes it difficult to see the direct connection with the rest of the report. (I haven't had the fortitude to work through the appendix comparing FRBR attributes with ISBD, GARE and GSARE but I assume that a matching was done. However, this does make the recommended core record hard to read in the context of FRBR.) For example, there are core descriptive elements relating to uniform titles ("addition to uniform title - numeric designation (music)") yet uniform titles are not mentioned among the FRBR attributes and the term "uniform title" is not included in the index.

It is not clear to me what has happened to the goal to provide a solution for cash-strapped cataloging agencies. The E-R model, which in my reading was offered as a methodology to support the analysis that needed to be done, has become what people think of as FRBR. If the FRBR Review Group, which is currently maintaining the results of the Study Group's work, does have activities that are aimed at helping national libraries do their work more effectively while saving them cataloging time, it isn't nearly as visible as the work being done to create definition of bibliographic data that follows entity-relation modeling. In any case, I, for one, was actually surprised to discover Chapter 7 in my copy of the FRBR Study Group report.



[1] Terms of Reference for a Study of the Functional Requirements for Bibliographic Records. (1992). Available in: Le Boeuf, P. (2005). Functional Requirements for Bibliographic Records (FRBR): Hype or Cure-All?. New York: Haworth Information Press.
[2] Madison, Olivia M.A. The origins of the IFLA study on Functional Requirements for Bibliographic Records. In: LE BŒUF, Patrick. Ed. Functional Requirements for Bibliographic Records (FRBR): Hype, or Cure-All? [printed text]. Binghamton, NY: the Haworth Press, 2005.
[3]Tillett, B. B. (1994). IFLA Study on the Functional Requirements of Bibliographic Records : Theoretical and Practical Foundations, (April), 1–5.
[4] IFLA Study Group on the Functional Requirements for Bibliographic Records. (2009). Functional Requirements for Bibliographic Records. Retrieved from http://archive.ifla.org/VII/s13/frbr/frbr_2008.pdf

Friday, February 14, 2014

FRBR as a conceptual model

(I have been working on a very long and very detailed analysis of FRBR, probably more than anyone wants to know. But some parts of that analysis might be generally helpful in understanding FRBR, so I'm going to "leak" those ideas out through this blog.)

The FRBR document, in its section on Methodology, gives the reasoning behind the use of entity-relation modeling technique:
The methodology used in this study is based on an entity analysis technique that is used in the development of conceptual models for relational database systems. Although the study is not intended to serve directly as a basis for the design of bibliographic databases, the technique was chosen as the basis for the methodology because it provides a structured approach to the analysis of data requirements that facilitates the processes of definition and delineation that were set out in the terms of reference for the study.
E-R modeling is a multi-step technique that begins with a high-level conceptual analysis of the data universe that is being considered. Quoting the FRBR document again:
The first step in the entity analysis technique is to isolate the key objects that are of interest to users of information in a particular domain. These objects of interest or entities are defined at as high a level as possible. That is to say that the analysis first focuses attention not on individual data but on the "things" the data describe. Each of the entities defined for the model, therefore, serves as the focal point for a cluster of data. An entity diagram for a personnel information system, for example, would likely identify "employee" as one entity that would be of interest to the users of such a system.

This is a very good description of conceptual modeling. So it is either puzzling or disturbing that most readings of FRBR do not recognize this difference between a conceptual model and either a record format or a logical model. In part this is because few have done a close reading of the FRBR document, and unfortunately it is easy to view the diagrams there as statements of data structure rather than high level concepts about bibliographic data. (It's not surprising that people get their information about FRBR from the diagrams, rather than the text. There are three very simple diagrams in the document, and 142 pages of text. Yet even if a picture is worth a thousand words, those three are not equal to the text.)

One of the main assumptions about FRBR is that the entities listed there should be directly translated into records in any bibliographic data design that intends to implement FRBR. For example, there is much criticism of BIBFRAME for presenting a two-entity bibliographic model instead of the four entities of FRBR. This reflects the mistaken idea that each Group 1 entity must be a record in whatever future bibliographic formats are developed. As entities in a conceptual model there is absolutely no direct transfer from conceptual entities to data records. How best to create a record format that carries the concepts is something that would be arrived at after a further and more detailed technical analysis. In fact, the development of a record format might not seem to be a direct descendent of the E-R model, since the E-R modeling technique has a bias toward the structure of relational database management systems, not records, and the FRBR Study Group was not intending its work to be translated directly to a database design.

There are innumerable ways that one could implement a data design that fulfills the conceptual view of FRBR. In E-R modeling there are subsequent steps that build on the conceptual design to develop it into an actionable data design. These steps are actually more detailed and imposing than the conceptual design which is often used to bridge the knowledge gap between operational staff and the technical staff that must creating a working system. The step after the conceptual model is usually the logical design step that completes the list of attributes, and defines the types of data values that will be stored in the database tables (text, date, currency) and the cardinality of each data element (mandatory, optional, repeatable, etc.). It then normalizes the data to remove any duplication of data within the entire database. It also resolves relationships between data tables so that one-to-many and many-to-many relationships are correctly implemented for the applications that will make use of them. Although this is couched in terms of database design, an equally rigorous step would be needed to move from a conceptual view to a design for a format that could be used in library systems and for data exchange.

As an illustration, here is a logical design for the bibliographic system MusicBrainz that stores information about recorded music. It has many of the same concepts as FRBR (works, performers, variant expressions), and must resolve the complex relationships between albums, songs, and performances (not unlike what a music library catalog must do):


With perhaps some difference in details you could say that this implements the concepts of FRBR. Still, this is a database design, and not a record format. For many databases, there is no single record that represents all of the stored data. Business databases are generally a combination of data from numerous departments and processes, and they can often output many different data combinations as needed.

It does say something about the state of technology awareness in the library profession that once a presumably successful conceptual model was developed there was no second step to make that model operational. What was the ultimate goal of FRBR, and did it fulfill that goal? Look for another post soon on that topic.