CIL2006, Day 3: The Future of Catalogs

Roy Tennant:

What catalogs do well:

  • inventory control system – what you have, where it’s located
  • known item searching
  • … for items within a particular collection

Doesn’t do well (his “short list”):

  • any type of search beyond known item
  • anything beyond books and journal titles
  • not good at displaying results by logical groupings (FRBR)
  • Faceted browsing
  • relevance ranking
  • recommending similar titles

How we got in this mess:

  • automation began in the back room – automated circ functions, acquisitions,etc
  • moved to the public as an afterthought
  • system was optimized for librarians
  • failed to exploit full power of computer systems and catalog infrastructure

key problems:

  • mixed management and discovery purposes – bad mashup
  • stovepipe system – hard to get information back out
  • abdicated responsibility to the vendor
  • slow to exploit new opportunities
  • reluctant to collaborate on proffesion-wide level


  • library catalog is one finding tool among many
  • acknowledge the good and bad
  • users want a variety of information sources – not just books
  • we can do better!

Future of catalog:

  • one system among many that must interoperate
  • refocused on local inventory only
  • functions well alone
  • won’t be the most important finding tool in a library

Signs of Life:

  • gapines
  • redlightgreen
  • OCLC Curiouser project
  • X9 csusm
  • Andrew Pace’s new catalog, which leads us to …

Andrew Pace

You never get what you don’t ask for.

Next Generation OPAC examples out there: vivisimo, aquabrowser, endeca, exlibris, etc.

Sirsi/Dynix has partnered with FAST?

Endeca – gave a NCSU catalog tour

They went with Endeca because it had:

  • speed
  • relevance ranking
  • faceted browsing
  • true browsing
  • spell checking
  • stemming
  • “did you mean” search suggestions

They partnered with Endeca

Tech overview: Endeca coexists with their SirsiDynix Unicorn ILS and the Web2 catalog

Indexes MARC exported from Unicorn. Index is refreshed nightly

They have to reformat MARC so Endeca can parse it

Endeca doesn’t understand MARC – it gets turned into a flat text file for Endeca

Who is Endeca? they built the search engines that Walmart and Barnes & Nobles uses, among others

challenges: using LCSH like it’s never been used before

Future plans – getting rid of authority searching (because none of their users use it)