The XQuery Exposé, Kevin Clarke
28 Feb 07
The XQuery Exposé:Practical Experiences from a Digital Library
Kevin Clarke
Princeton University
http://diglib.princeton.edu/
XQuery
– an XML Query Language
– SQL of the XML world
– integrates XML data from a variety of sources
– focus on data, not dbs; higher level than SQL
– does not include fulltext searching or update
– not a full-fledged programming language
– currently working through Lucene; SOLR very enticing
XQuery is a functional language built entirely on expressions
uses native XML types and XPath
can be “loosely typed”, e.g. can omit declaration of data types
– Elsevier has done some work on code conventions
let 10,000 FLWORs bloom!
primarily an image-based site
some search, but mostly browsing
Posted by pzed on February 28, 2007 at 2.47pm
Forget the Lipstick, Fabien Tiburce, Peter Giansante, Beth Jefferson
28 Feb 07
Forget the Lipstick: This Pig Just Needs Social Skills
Fabien Tiburce, Peter Giansante, and Beth Jefferson
BiblioCommons
Beth Jefferson
extensive user research
BC, AB, ON provided proof of concept funding
today’s focus is on architecture and user experience
OPAC user’s not complaining, just going elsewhere
nextgen catalogue more than tech
social search is key
– discovery, relevance, connections to community and other users
Fabien Tiburce
tech – ajax widgets, rdbms mixed with xml repository
– object factories to aggregae representations of data
– decouple data in transit from data in storage
– rdbms to preserve most important data
Peter Giansante
data model
Beth Jefferson
personalized relevance (ratings, reviews, demographic filtering)
building trust in the social environment, adding other users to a network optionally based on a domain of expertise
Fabien Tiburce
social data coupled with the user
– user preference subsystem
– user generated data associated with bib record and other users
Peter Giansante
user prefs as part of data model
Beth Jefferson
thinking outside the box of the library
want to get away from “sorry, no match”
– opportunities to use ILL in local interface
– suggest to purchase
– suggest share from other user’s personal collection
– refer to community generated interests (groups, discussions, etc.)
engaging users to take us halfway
– use library metadata to structure discussions
– can then put community resources, events, etc. on the right pages
when a question is asked, discover who is likely to have the answer and present it to them
finally, offer an online answer service
– people in the OPAC
– users connecting with other users realtime in OPAC
Peter Giansante
more data model
Beth Jefferson
need critical mass of users
data must tie back to them
must engage a significant percentage of users
maximize breadth of implementation
– in the flow of existing activity
– approaching costless
– provide motivations to contribute
Fabien Tiburce
ILS integration with low footprint web services
will be open source
data flow diagram!
whew!
Posted by pzed on February 28, 2007 at 2.25pm
Library Data APIs Abound, Richard Wallis
28 Feb 07
Library Data APIs Abound!
Richard Wallis
Talis
http://www.talis.com/home/
libraries are used to obscure protocols, starting to use less obscure
all have an insular view of the world
– I have data, and you can have it if you use the right protocol
– mixing that data with other data is your problem
Talis platform uses Bigfoot
– easily queried
– can query bib data, pass search to augment search with holdings data, pass again to augment with deep linking to OPAC content and return results with links
– each “store” (bib, holdings, book jackets, etc.) has an API that can be search or used to augment
– can also augment XML data from other sources by transforming into a stream that can talk to API augmentation
Bigfoot APIs
– Items
– Augment
– Facet
– OAI-PMH (coming)
– Config (coming)
– On-demand stores (coming)
– transform service (built in to item query, coming for others)
Possibilities
– transform output to WordPress compatible output
c.f. the 20 minute union catalogue
Posted by pzed on February 28, 2007 at 2.10pm
Smart Subjects, Tito Sierra
28 Feb 07
Smart Subjects: Application Independent Subject Recommendations
Tito Sierra
NCSU Libraries
Input user search query and spit out a list of related library subjects
based on search log analysis done as part of QuickSearch development
– found lots of topical subject queries
NCSU also provides browsable subject portal
– locally developed classification, approx. 100 subject nodes in 12 top-level catagories
– nodes influenced by local curriculum
– subject specialist mapped resources to the subject page
also based on OpenSearch
started with data from course catalogues to gather terms for mapping; also used text snippets from papers in faculty publications archive
– text extracts used to create search indexes
– run keyword search on the index; return, rank, dedupe results
– crosswalk to classification map
strengths
– application and collection independent
– subject recommendations can be integrated into any library search application
– broader, serendipitous resource discovery
weaknesses
– false positives (bad recommendations)
– zero hits
future plans
– database advisor
– increase size of subject indices using article tables of contents and backlog of course descriptions
– guage interest in a less-specialized release to the community
http://www.lib.ncsu.edu/dli/projects/smartsubjects
Posted by pzed on February 28, 2007 at 1.00pm
On the Herding of Cats, Mike Rylander
28 Feb 07
On the Herding of Cats
Mike Rylander
Georgia Public Library System
General overview of planning, etc.
Posted by pzed on February 28, 2007 at 12.40pm
Free the Data, Emily Lynema
28 Feb 07
Free the Data: Creating a Web Services Interface to the Online Catalog
Emily Lynema
NCSU Libraries
Architecture – front end web application to an Endeca back end; considered advantageous: a “freedom” feature that allowed the creation of an XML interface to the catalog
began as a web service to speak to other systems like WorldCat and return availability results
REST API for querying catalog
– looking for RSS feeds, particularly new books
– integrating catalogue results into website QuickSearch
. . . alphabet soup. . .
wanted search results and facets
OpenSearch add-on; c.f. A9 aggregator
also had to integrate with NCSU’s QuickSearch product
facet data
– used OpenSearch query role="subset" custom:facet="4617264627"
– using numbers for facet values makes for cleaner URLs
– do aggregators provide support for this query role? not sure
– multiple elements would slow down results
http://www.lib.ncsu.edu/endeca
http://www.lib.ncsu.edu/catalog/ws/
Posted by pzed on February 28, 2007 at 12.19pm
MyResearch Portal, Andrew Nagy
28 Feb 07
MyResearch Portal: An XML based Catalog-Independent OPAC
Andrew Nagy
Villanova University
completely ILS agnostic portal for research activities
– catalog, DBSs, digital library
– single search interface
most resources are in XML
– digital library: METS
– Metalib XServer: XML
– catalog: MARCXML
– web site: XHTML
data store advantages in XML
– native XML stores easily, easily entered
– no need for RDBMS
eXist
– open source XML dbs
– full-text searching available
– platform independent: java backend; API through REST or SOAP
– inherent directory structure
– LDAP support
– however, not really meant for a library search type of system
Berkeley DB XML
– proven
– wide range of platforms
– good performance
– commercial backing, decent help
– no full text extenstions, no inherent directories
commercial solutions? – too expensive, more complex
implementation challenges
– eXist innappropriate
– DB XML capable, but slow, even with some reconfigurationm
– MARCXML very complex, lots of unhelpful data, field names actually in attributes
moving MARC field number into tag improved response times considerably
technologies don’t tend to understand the complexity of library data
queries not well optimized, if at all
– needed to develop very basic query processing
SOLR/Lucene looks like the answer
Posted by pzed on February 28, 2007 at 12.03pm
code4lib: Karen Schneider keynote
28 Feb 07
state of emergency in the profession (graphic shows homeland security level red!)
– have given away our collections (digitization by 3rd parties, journal content)
– don’t build or own the tools
– complex, poorly marketed systems
– function like a monopoly service when competition is thriving
quote from Andrew Abbott, The System of Professions
– digital library work, library software, is “the heartland of work”
– gets back to memory work: get beyond gaming, etc. to cultural memory
– rather than “don’t be evil”, libraries strive to do good
5-3-1 rule
– pick five ideas
– work on three
– make one happen
five things
– digital preservation
– standards adoption
– sucky library software
– 3rd party content hegemony
– scholarly awareness of key library issues
fold into three: digital preservation, sucky software, scholarly awareness
one: sucky software
four nifty things in softare: Evergreen, Umlaut, Scriblio, SOLR
beginning to restore the balance of power
reinstate the direction – this is our work
emphasis on library as memory organization
sends the message that we maean business
creative decoupling of components
interesting of re-use of other tools
resocialization of librarian artisans
#1 today is Evergreen
– timing is perfect: market consolidation at the same time as the centrality of the ILS is weakening
useful over-generalizations
– nobody cares about open source
– nobody cares about standards
– nobody cares about usability
– nobody cares about Evergreen
they care about books, videos, etc. NOT SOFTWARE
we don’t even want them thinking about Evergreen
more over-generalizations
– ARL still counts books on shelf as number one indicator
– IT directors have no resources to take on unfunded mandates
– most libraries can’t provide developer time (exceptions tend to be in higher ed.)
how directors see the world
– how much does it cost?
– what are we getting for the money?
– what are other directors doing?
– what problems will it create?
– why spend on this rather than something else?
– is this thing “fully baked”?
directors do not generally have a good opinion of OSS
– one car accident away from orphan software
– no support model
– cheesy make-do quality
– arcane, developer oriented (tip: do not design for yourself)
– nobody else is doing it
free beer vs. free kittens
last-gen librarians love complex systems they need to teach
– our users need it, etc.
every library needs a developer
every developer needs a library
Posted by pzed on February 28, 2007 at 12.01pm
Breathing Light: The World of Tokihiro Sato
21 Feb 07
I’ve been saying (tongue in cheek) to anyone interested in listening, that photography is dead. Like a parasite that has consumed its host, a population that has denuded its ecosystem and has nowhere to go, photography is dead precisely because there’s so goddam much of it. I stand corrected.
I’m really excited by the work of Tokihiro Sato. Admittedly, much of the work in this “Jacobins” show dates from the late 80s and early 90s, arguably before photography’s death, but I’m still looking forward to seeing Photo Respiration: Tokihiro Sato Photographs, a 2005 book that has been ordered by my library.
Takuo Komatsuzaki, the curator who wrote the brief entry on the above mentioned website, says:
Sato’s photographs give us a strong feeling of space, depth, and, through the artist’s process of applying light, even a sense of time. The aesthetic effect of these sensitively expressed works is certainly pleasant. However, we have to realize that if we simply stop there, we are doing no more than scratching the surface. The photographs of Tokihiro Sato breathe in an altogether deeper space.
It’s that sense of time that excites me, like we’re looking at ghosts, or spirits (kami?), or simply movement and process.
Sato on the web
- “Jacobins”: Museum of Fine Arts, Agen, France, 1993. I found this one first, wrote this post, and then went looking for some others.
- Tokihiro Sato, Leslie Tonkonow Artworks + Projects. New York. I’m a little less excited by the recent colour works, but the 1996 Buddha Head is very cool.
- Tokihiro Sato, Haines Gallery. San Francisco. Here we see some of the colour images installed using light boxes; not a good sign, but again you have to see them up close to decide. Clicking on the image leads to a number of older black’n'whites. I especially like #349 Kashimagawa, 1998.
Posted by pzed on February 21, 2007 at 11.26am
Akamaru
17 Feb 07
Jodi posted a too-cute picture of Akamaru (Kiba’s canine sidekick in Naruto). For fun, and because it’s Saturday morning and the reference desk is calm, I followed Jodi’s akamaru tag to look at other pictures of the cute little guy. Much to my surprise, I discovered instead something that looked suspiciously like ramen: one from flickr user shiokuma, the other from flickr user rhosoi.
A g-search for akamaru ramen recipe brings back scads of stuff from Naruto fan sites, but I was lucky enough to find near the top this review of a Tokyo restaurant called Ippudo. According to the review, akamaru ramen is a “newer recipe”, made with red seasoning oil. My 12-year-old daughter Claire, the anime nut who has the whole family watching Naruto, tells me that “aka” is Japanese for “red”, and “maru” is a suffix that denotes a masculine name. Is this ramen named after Kiba’s puppy? Hard to say, but I’d like to think so.
Posted by pzed on February 17, 2007 at 1.07pm
