Eric Lease Morgan is Head of the Digital Access and Information Architecture Department at the University Libraries of Notre Dame. He was previously at North Carolina State University from 1991-2001 where he created the MyLibrary feature.
Gives handouts to help him process what is in his head and creates a structure for himself to see what is going on in libraries.
Can talk about trends but really wants to focus on opportunities. Focus on the How rather than the What.
Loves libraries. Makes his own books, archives his personal items.
Look at higher level trends
- Not iPhone – mobile computing
-
Not Prius – green technology
Small bits of information are trend. Not whole books, facts within them.
Web-enabled mobile devices – ubiquitous computing
Semantic Web
Information for computers to read because of linked data
Computer finding various relationships that others have created from many different perspectives.
Graph these relationships
Walden example
Name, dates, people who have cited it.
Common relationships and rare relationships will show up.
Middle relationships will be true
Rare will be questionable
Examples of linked data (using RDF’s)
Infomotions – RDF File
About Walden – originally from Freebase.com – made by MIT
rdf.freebase.com/ns/en.walden
Shows URI that refers to thoreau.
Libraries have created some really cool technologies but are now superceded.
MARC record was created in the 60’s, predated SGML, HTML
first five characters tell how long the record is. There is an ASCII character that indicated the end.
Shared office with Frank Kilgour
Said MARC was really stupid!
MARC has been taken over by XML.
Z39.50 was good then, but OpenSearch is better now for the current information environment.
MARC8 was good then, but now Unicode is a better fit. Especially for international, and even has Klingon!
People don’t want to browse, they want to search
There is too much to look at now.
Point of browse is to find “more like this”
Search is not meant to retrieve comprehensive results, just want a couple good ones.
Relevancy Ranking
Wisdom of Crowds
Internet was designed to withstand Nuclear attack (from DARPA – military)
distributed, no central authority
If a lot of people agree than it is probably true.
How to deal with this environment
See the big picture
We are straddling two worlds which is a challenges
Build collections for historical records
organize it into more manageable chuncks
Preserve them
very challenging becaues they must foresee changes
Provide access
Competitors in this environment – fashionable to be into information
enables democracy – now many places to get information
not able to assume that we are exclusviely ”public good”
What can we do?
Save the user time
Play and innovate
Items are free for the taking
Find ways to harness items on the Web to help your community.
Cannot be exclusively about books
Collect and organize
doesn’t recommend we use MARC, use what everyone else does XML
Controlled Vocabularies
not as important as they used to be
can use wisdom of crowds, full text computer data mining also
Use Indexers
Databases are good at creating and maintaining
but user must know structure of each – search commands
Index is about finding
Search words
Use them together for best result.
Go beyond just giving them the “thing” they want
Add value, everyone else can find and give
Provide people with tools to use the content
Once I get the book help them:
Read it, cite it, compare, summarize, trace idea through it
Not just books, other items
datasets – create graphs, tables, charts
Other services
steer and recommend items to people
http://infomotions.com/musings/xml-in-library/user
Dirctory of Open Access Journals
He created a program to harness information from the Open Archives Initiative (OAI)
While Google Books is good for UMich, not for everyone because Michigan gets a scanned copy everyone else is at the mercy of Google
Internet Archive is truly free.
Has been working with Univeristy of Toronto
His own library of books – Alex
21000 public domain and open access documents from American/English Literature as well as Western Philosophy.
Can provide services against the Index – limit by subject, author, type of item (article, book), repositories (Internet Archive, DOAJ)
Demo of collection of various files of the same text from Internet Archive.
Can search within PDF’s in Internet Archive (because of dirty OCR’s), not Google Books.
He uses Perl. Files ending in .pl
Concordance in a general sense, counts words and gives context. Perl command make-concordance.pl – Creates an index.
All these examples, what do you envision all librarians doing with this?
He envisions people who go and do this for specific topics.
Raw text that he harvested can provide services against it.
Look up word in Dictionary, Look up in Catalog, Translate
Concordance of title, can see top 10, 50, 100 terms. then locate those terms where they appear.
Core skills necessary to do this.
Library collection, organization, dissemination,
Computer Skills
Relational Databases,
XML
Indexing
Programming