Info Grid 2005 - Monday 26th - Deploying Services, Not Libraries

Deploying Services, Not Libraries
(or, Staying out of the middle of the road)
Peter Brantley
University of California, California Digital Library

Presentation: ppt (3.9M); pdf (1.2M)

Last Generation Systems

"OPACs are wonderful card catalogues, online, not much more"

For most users, libraries are places with books and comfy chairs.

Digital library silos

content is wonderful... and invisible to the user community

[...]

Change happens

Most of the new developments in our domain are not really *about* libraries

= repositories
= publishing services
= learning object repositories
= Google Scholar, Scopus, CiteSeer

Interesting times

* Scholarly work and communication are being transformed
* Users are seeking not just content discovery, but content re-use

Participate in change!

* Libraries must engage and seek solutions, understand faculty and students
* Create informed advocacy for waht users what, not what we think they should want

On Our To-Do List

* Engage all stakeholders [about] the worth of digital assets and service

*** Explode the roles of libraries within the enterprise

* Find new ways to make library services available to the users -- "flattening" the library

The road to services

* Libraries cannot achieve these things without altering themselves radically
* Participate actively in open-source communities
* Deploy services-oriented architectures, not build more silos

SOA

SoA means
- a set of self-contained, functional components (of higher level apps) encapsulated as services
- interaction between services through well defined interfaces
- [...]

CDL's Common Framework

The CF is
- a philosophy governing software devel
- a conceptual design for CDL services
- specific technical architecture
- on the wire services
- apps ?

[what to do]
- composite, modular, lightweight good
- design and implement quickly
- replace need for application-specific kludges
- make replacement and enhancement easy
- encourage staff training and devel
- quests for perfection are not permitted

CF Design

- elemental search concepts - search, admin, harest
- apps independent of services
- design atomic services to enable easy construction and rebuilding
- new app = reuse existing services

CF Tech Strategy

- common dev environ
- web services (XML, XSLT, Java)
- preserve generic design in spec
- cleanly separate api
- services agnostic human/machine
- externalize data modifications

CF Progress

- redefining continuously
- UC Digital Preservation Repository (DPR) is first production CFL app using CF
- Reciprocal design/functional review with stakeholders
- CF permits rapid dev cycles

Not just back-end

- SOA enables new modes of user interaction
- if sufficiently pervasive, it fundamentally changes the meaning of apps and ways in why data can be used
- an app is now sum of functions (services) addressible through APIs

Web 2.0

- web 2.0 extension of soa concepts (very loosely)
- data and systems made avail through apis
- Ajax uses service apis
- mashups

Consequences

* accentuates value in data and service provision but you lose branding

Expectations

* Newer generation apps are raising expectations (Flickr, delicious, LinkedIn, MySpace)
* Users expect to be part of the creative flow of content, not just discoverers or consumers
* [...]

Social bookmarking

* Connotea [guy from Nature is here]
(wishes it would support Shibboleth)

Flickr

Web 2.0 Impact

* Leveraging The Long Tail
* Small pieces, loosely joined
* Self-service and participation
* Radical decentalization
* Emergent behavior
- Dion Hinchcliffe's Web 2.0 Blog, Sept 2005

Web 2.0 Meme Map

Tim O'Reilly, Foo Camp, 2005 (Flickr)

mapping of concepts

"to use Google, generally, it's a fun experience... who can say that about their library OPAC?" (got a laugh)

Lines of Business

* Open services suggest libraries might not need to be engaged in every line of business
* Why not use Flickr through API

use the best services for your needs

Collaboration Essential

* As more people develop service platforms, why not use them
* e.g. CDL uses Lucene, made some changes, returned enhancements to Apache
* OCLC WorldCat may be the best union catalog available (in the US)

quote from Lorcan's blog

"libraries will have to source shared processes and infrastructure through common platforms - the opportunity costs of not doing so... are too high"

Future

Imagine new types of CONTENT SERVICES
- collaborating communities maintain shared lists or resources
- apps permit linking of disparate content sources into one presentation
- support redefinition of both publishing and the products of creative work

The New Library

* combine content from multiple sources
features could include
- multiple sites for access and preservation
- automatic replication
- different branding on each site
- granulated access rights via Shibboleth
- published APIs to enable mashups
- easy to personalize (use some opaque identifier) e.g. "other users also looked at..."

Next Gen Bib Services

- assessing using Lucene XML-based indexing and discovery
- how to order results for relevance?
- feasibility for recommendation and personalization
- explore alternative UI strategies

Q: There are lots of science fiction reviews on Amazon, but not many for science books.
A: not a common practice in our current scholarly community
Explore: create a community where people can work together without a lot of noise.

Q: Redefine the funding model for digital libraries?
A: These are very hard questions, I don't have an answer.
Libraries need to engage upper management to determine what models may be.
Educate the organization about the value and potential of your library.

Q: [footnote] Imagine where author makes book free, lives off ads and presentations etc.
if all books will go open access...
Libraries will have to catalogue everything that is on the net.
A: There is still a role for publishers.  Even as authors can increasingly easily self-publish.
Presumably self-published work will go into archives?

Where does the money come from?  (It costs money to run an archive / repository.)
Academic AdSense?

Q: Business model for those providing Web Services?
A: If you just have APIs, you can't demonstrate.  Need app to show use of Web Services.
Need some new ways of adding revenue streams.  Value-added services?