Issues and Ideas for the Next Generation Catalog

Issues and Ideas (cooked)

Scenarios

• Scenario 1 - Catalog as Inventory Only

See also http://futurelib.pbwiki.com/Scenarios

Issues and Ideas (raw)

• What is the role of classification? Topic maps?
• Can we change the catalog without changing cataloging? Look more at how they are interrelated.
• In the digital environment we need controlled vocabularies that are systematic and rational, that allow users to slice in different ways, that allow browsing, navigation, and profiling in addition to intentional single-class-at-a-time-retrieval; and that provide human-readable rather than numerical-notational-labels, ideally providing for multiple labels (multi-lingual?), and that dont' assume any item has to go in one and only one bucket.
• Social tagging
  • private tags vs. public tags.
  • Tags of general interest vs. tags that might be particular to a single copy, single individual, or close-knit group. (e.g. who gave it, where it 'lives').
  • Statistical use of tags to identify characteristics of an item.
  • Privacy issues.
  • "can it work"--if we make it available, will they tag.
  • Concern that a folksonomy is not as powerful as professionally-applied controlled vocabulary-based classification system.
• The "all things to all people" problem -- is there a single solution?
  • Perhaps the solution is to allow significant amounts of customization by the user. The user can then put together the solution that works for them. How do we find just the right amount of customization? http://librarianinblack.typepad.com/librarianinblack/2007/02/user_control_an.html implies that there can be too much
  • Low-barrier "profile" system, possibly based on FOAF, could allow dynamic presentation based on preferences, audience, etc.
  • Create many points of deployment, such as portable apps, lucene index of catalogue, and feeding desktop indexers.
• user-centric vs systems-centric
  • the system should be providing the raw data in many forms (i.e. indexes, thesauri, etc) the user should then have the tools to use what works best for them. Think Google Desktop, Yahoo! Widgets.
  • the system should seek to efficiently replicate data in the most malleable form possible, and make it possible to seamlessly keep this data up to date.
  • perhaps the catalogue contains the virtual DNA of an object, and could position the description for endless augmentation
  • will users be able to use the tools we provide for them to access our data? Library patrons often lack rudimentary technical skills -- can we build tools that support those patrons?
• Will the catalog be built around the inventory of the library, or will it be built around the needs of the patron? How would the world look if we start with the premise that the user wants to access some information for a specific purpose, and they don't care where it is, who owns it, whether it is in a catalogue at all...etc? Should we stop thinking in terms of libraries altogether, and think instead of an inclusive social network of individuals and organizations?
  • What if we built a catalog around the idea that the most important element of a book is what social interaction it can bring for our patrons? -- for example, how well it will enable them to contribute to a group discussion, and thus give them credibility with their peers? An item in this scenario becomes a social gathering point as well as a piece of inventory.
  • What if our catalogue is nothing more than a data source which feeds the user's catalogue, and their own personal catalogue is the only thing that matters to them?
  • What if the catalog provided not just a pointing device telling patrons where on the shelf to find a book, but was able to pass the patron through to the content itself? Patrons often want not the item, but the content. They also might want to search not just the library collection, but also their own libraries, their friends' libraries, local bookstores, and the collections of strangers who might be willing to share.
• Browsing
  • Is browsing a requirement of the data format?
  • A system feature?
  • Or not needed (because it would be better to use topic maps)?
  • Browsing will always be needed -- human beings cannot avoid doing it; it's a built-in part of the organism.
  • Linkages from classification schemes to subject/topic maps/thesauri, etc. must faciliate browsing.
• Humans will automatically start to orient themselves within whatever schema is presented, so will seek out terrain which has some kind of recognizable shape (could be a word, but surely not necessarily one). Patterns emerge (again, could be word patterns, but could be vistas of other kinds). We should try to work together with, and alongside, human functioning.
• We need to stop thinking in a flat, two dimensional print-like space. We should use shape. Color. Motion. Texture. I would even say Smell and Taste if there were technologies that tapped into those. Just for sake of argument: what stops us from putting the 'author entities' "space" on a "texturized" frame of a certain shape? Again, working in concert with human ability to recognize things. Reduce cognitive overhead at the same time. Diversify the lead-in mental vocabulary, if you will. More points of contact.
• Should relevant commercial content (e.g., book reviews from Amazon; search results from Google) be linked from or incorporated into the catalog?
  • Issue: The decision to incorporate commercial content into services funded by tax dollars is not one to be taken casually. Doesn't mean it can't be done, but public institutions have to tread carefully.
  • Problem: When we provide a service, particularly a visible, heavily-used service that is central to our public perception, we need to know how to fix it if it's not working and have a reasonable assurance beyond "heck, this will never go away" for a service's availability.
  • Issue: We have to make sure we don't give away our patrons' privacy or the public's content.
  • Can we attract value-added user-content? (social tagging, individuals' reviews, ...)
• Display of search results
  • Ranking and clustering
    • "There's no reason a FRBRized OPAC shouldn't provide clustered and relevance ranked results to the query "show me the editions of Hamlet". Clustering based on mode of expression (script, performance, popup-book) and language of expression and relevance rank based on availability (to the searcher), popularity, "quality" of annotation, .. would probably be useful." -Kent Fitch to NGC4LIB 2/26/07
    • Perhaps cluster search results by call number area, to give searchers likely places to start shelf-browsing.
      • At the same time, give the searcher a list of results that are in areas where shelf-browsing would seem to be a waste of time.
  • Rich browsing and navigational capabilities