�ݺ�ߣ

Editor's Notes

• #3: Let’s spend a couple minutes doing some group brainstorming. Let’s keep these actionable tasks in mind but for now, I’d like to transition into talking about competencies.
• #4: Competencies are behaviors or tasks that must be completed in order to show adequate job performance. Competencies are usually not concrete in nature. For example, being a performance artist requires balance—balance is a competency. How do we know that a performer possesses good balance? We look for performance indicators. For example, you may observe a bagpiping Sith Lord riding a unicycle and you may draw a conclusion as to whether that performer has demonstrated the competency of balance adequately. On their own, neither the unicycle, nor the bagpipes, nor the costume are competencies; they’re tools of the performer’s trade. Similarly, MARC is one of many tools catalogers use to perform their jobs. Consistent and correct usage of MARC is evidence that a cataloger has a met one of the key competencies for catalogers: the structuring of data for machine manipulation.
• #5: When “traditional catalogers” perform metadata work, I don’t automatically assume that existing competencies must change (new competencies occasionally arise—I’ll give an example later) I do however, expect tasks and performance indicators to change. When introducing metadata work into daily duties of a cataloger, the highest concentration of change is more likely to be in the realm of tools and tasks. This is what I’d like to address next—how to reimagine performance indicators in light of expanded metadata tasking
• #6: This is, of course, not an exhaustive list! I’ve selected a few competencies for the sake of illustration.
• #7: Standardizing data values, by using tools like content standards, helps us normalize data for indexing, display, etc. Being familiar with a set of content standards such as AACR2 and RDA primes us for cataloging work even when we’re in unfamiliar territory--we don’t have to start completely from scratch every time we’re met with something new, we have a framework to build upon The benefit of learning new content standards and how they are applied in a user community, is that you begin to see the biases of standards you know well For example, I refuse to use RDA to describe archival materials. Why? How do you reconcile the idea of FRBR Expression with a handwritten letter? RDA has a built-in publication bias. This doesn’t work for archival materials. Danger of using/knowing only one standard is that you’re unaware of how metadata you produce “plays” with other metadata in The Wild. Be aware of your blinders. You can’t know everything but you can try to know what you don’t know.
• #8: Parsing a resource for descriptive metadata is one of the first things catalogers are trained to do. Experienced catalogers do this so naturally, we’re not aware we’re doing it. What happens when the resource is predominantly visual, rather than textual, in nature? How does one parse an image? Here’s an instance where there may be a new competency to be considered! Visual literacy is required for successful image description—this is hardly new territory for media catalogers but it not so with those who only cataloged text-based resources.
• #9: Catalogers’ brains seem to be hard-wired to recognize structural patterns and hierarchy of information. This competency is needed to master taxonomies that contain broader and narrower terms, such as Library of Congress Subject Headings. This competency is also useful for encoding textual content for machines to parse and display; has been useful in text encoding projects using TEI and EAD. Is domain expertise (in a subject or language) a competency?
• #10: Recognizing and recording relationships has always been important in cataloging work; in AACR2-speak, think of all of those main and added entries we supply during cataloging and subject analysis. RDA really pushes us to focus on this particular competency. Think of all of those 7XX and $e relator terms that RDA wants you to add… The identifier at the center is for James Whitcomb Riley. We have physical materials (books as well as letters, papers, and photographs in which Riley is the subject); many of these objects are being digitized. By declaring relationships, one can discover that Lew Wallace once posed with Riley for some photographs; and that Riley wrote letters to the man who would become is biographer, John Marcus Dickey--our special collections library hold’s the biographer’s papers This linked data stuff REALLY hinges on authority data!
• #12: Change is nothing new to catalogers. Technology has always seemed to be changing out from under us—standards changed and continued to develop (MARC), we implemented OPACs, etc.
• #13: Technology doesn’t fundamentally change people; however, it can have a significant impact on their behaviors and expectations. And then there’s that semantic web thing…
• #16: Schema.org was the brainchlid of Google, Yahoo, Bing! Introduced in summer of 2011, it was implemented to structure web content semantically to support searching. This means that not just headings and paragraphs are being marked up; ingredients and cook times are encoded for recipe pages, actors, directors, and synopses are encoded in movie pages, and location, hours, and contact information are encoded in restaurants’ webpages. Shown here is a search result from a search on events in Las Vegas. Dates and shows can be displayed because that data is marked up semantically on this particular website.`
• #17: Because most emails are coded in HTML, semantic markup is appearing in our inboxes too.
• #18: Google’s Knowledge Graph was introduced in 2012. Although Google hasn’t divulged much about the Knowledge Graph’s inner workings, the word “graph” suggests linked data graphs (triple statements). Relationships between people, places, things, and concepts are being encoded semantically. I have to imagine that some sort of taxonomy is being built to map complex concepts (such as broader, narrower, and related terms).
• #19: You know those cards that pop up at the right of your web browser, next to the search results list? That is data is being pulled in from the Knowledge Graph. A keyword search cannot do this; this feature is powered by taxonomies much like the controlled vocabularies we use in libraries. The difference is the Knowledge Graph is expressed in linked data. In DaVinci’s card, related artists and works of art are displayed.
• #20: Drilling down into works of art related to the search.
• #21: And then Google pushes things even further. Google quietly launched it’s new Hummingbird search algorithm in late August 2013 and announced the release a month later.
• #22: This algorithm is a complete rewrite; gone now is the keyword search, which, although powered by sophisticated algorithms, was still essentially a machine matching strings of text and displaying results that seemed to be the most relevant.
• #23: Instead of keywords, the new search algorithm focuses not on text strings but on concepts. Searching is meant to be conversational. Now, even when you type in keywords, Google seems to be answering natural-language question you didn’t ask.
• #24: This is where Google’s Knowledge Graph comes in—a rich taxonomy, expressed in linked data, provides powerful platform on which to build algorithms that tease out what searchers REALLY want to know. In this case, Google guessed that I wanted to know more about the dog breed (Wikipedia and American Kennel Club is at the top of list), or perhaps that I’m interested in adopting a retriever. The one-box approach to searching was once a “blind” matching of keywords; that one-box search now operates like a machine-facilitated reference interview
• #25: How serious is this change to semantic search?
• #26: Like it or not, Google is EVERYWHERER. It’s not just in search browsers that Google will be implementing semantic search; Google is in the OS and hardware markets too, meaning that semantic search is in our pockets, on our faces, and around our wrists. User expectations are already beginning to shift; our users will never be happy with keyword search again—this is very good news for catalogers.
• #27: How often have we heard this argument, usually offered as a reason why structuring data is a waste of time? It is no longer all about keyword searching. Please learn as much as you can about semantic search and start educating your staff and your administrators.
• #28: Rest assured catalogers already possess the core competencies needed in order to participate in the semantic web. The challenge will be in reimagining the tools we use to do our work and figuring out what those performance indicators will be.