The Corporate Librarian

Sarah Houghton-Jan and Shannon M. Staley co-presented. While we’re not using SharePoint, a lot of companies are and I wanted to get the real story on how easy/difficult it is to develop with, how good search functionality is, etc.. Apologies for the notes – I was trapped on an end seat with a not-very-good view of the slides. I’ll add a link to the online version later.

The audience was largely corporate and legal librarians, from the show of hands.

First was an explanation of what Microsoft SharePoint was, and Sarah explained the difference between Windows SharePoint Services (free lightweight version) and Microsoft Office SharePoint Server (standard enterprise version).

Next up Shannon gave a quote from FASTForward illustrating the general market reaction – it has a lot of functionality but the components are not best-in-breed (you can get better blogs, wikis, etc.). Sites created with SharePoint are very customizable, as the example screenshots demonstrated.

Similar products include Alfresco, Central Desktop, Clearspace, ???, ???, ??? Collaboration Suite.

Key benefits included the ability to include Web 2.0 functionality with one software service, facilitation of collaboration, increased communication, integration and compatibility with Microsoft Office, ability to post large documents while avoiding an email attachment glut and the ability to update intranet content without adding to the webmaster’s workload.

They picked SharePoint for the joint San Jose Public and State University Library because of the difficulty in communication between the two organizations, the administration’s desire to increase institutional communication and collaboration, the ability to standardize on support and on a single interface and the ability for teams to manage their own updates.

Designing the SharePoint presence involved IT, the Web team and input from the organization. Several smaller sites were piloted. Not all sites were permanently transferred to SharePoint. After the architecture was planned out SharePoint was opened up to a wider group.

By clicking on an intranet link, users are taken to the SharePoint site. All of the sites are read-only to everyone, for greater organizational transparency, communication and collaboration. Not all intranet content requires SharePoint for content management – the criteria they used were:

• Frequently-updated content
• Site requires group collaboration
• Which of SharePoint’s features would be useful in supporting a particular group’s content area

People using Internet Explorer and logged into Windows can access the SharePoint sites seamlessly from their own computers, otherwise they can login manually.

A new database of usernames and passwords was created so people didn’t have to remember them.

Site management features are not as intuitive as end-user features (e.g., method for adding an agenda is different from the method for adding minutes). Nomenclature is nonintuitive.

Key features:

• Managing documents (They encourage smaller more focused document libraries rather than one huge document library)
• Meeting workspaces (Calendar, agenda, minutes, supporting documents)
• Announcements
• Calendars (Classroom visits, collection development deadlines, storytimes, staff absences, etc.)
• Blogs
• Wikis
• Discussion forums
• Notifications (email or RSS alerts, each section has to be separately signed up for due to authentication process)

Other features include surveys, task lists, etc.

They’ve been using SharePoint for branch program calendars, sharing of statistics and reports, branch pages, etc.

IT manages the server, Web team can help with site set-up and ongoing support, each site needs a site owner, external online training resources as well as in-person training classes.

[various things raced through]

10/25/08 UPDATE: Sarah’s put the presentation and associated handouts up on her site.

Technorati Tags: IL2008

This was a joint presentation by Christopher Connell, Karen Draper, Emily Shem-Tov and Amy Affelt. Given that there was just one hour for four speakers, I’m not going to go into all-consuming detail. I’ll link to the presentations once they’re up.

Christopher Connell talked about using Vivisimo to ‘vivisect’ the OPAC and build a federated search tool for the enterprise. Vivisimo provided clustering of topics and the capability to federate the OPAC with other internal and external resources. He had the advantage of an in-house SQL programmer to extract bibliographic data from the catalog database and create a union query from the different catalog templates on an ongoing basis for Vivisimo’s index. The OPAC now integrates Library of Congress subject heading, provides on-the-fly topic clusters and groups subject headings and other metadata facets (corporate authors, formats, date, etc.). The catalog is repositioned as an integral part of the Institute for Defense Analyses’ intellectual capital.

Benefits, results and next steps:

• Researchers can connect people, projects and project deliverables.
• 25% of researchers are using the OPAC now.
• Plans for SharePoint 2007 integration and a document management system.
• Vivisimo will introduce tagging, commenting and other Web 2.0 features in its next release.

Karen and Emily presented on innovative ways the Adobe Library uses to communicate with users about events and training workshops past and upcoming, print and digital resources, etc.. How to let people know about relevant things they don’t know about/don’t realize are relevant?

Communications channels:

• Events offer an opportunity to publicize services in email newsletters and intranet articles
• Window displays
• Reports to management
• Blogs and wikis (both the Library wiki and other groups’ wikis)
• Tagging items in the company social bookmarking tool within the firewall
• Annual surveys as means of obtaining feedback and raising awareness
• “Best bets” in the research portal
• RSS starter feeds
• Free T-shirts and cookies for National Library Week, signs all over the place

No one channel is best – elevator signs was a big channel generally for getting internal event information, but so were the company intranet, internal email newsletters, email distribution lists, etc..

“Summer blockbuster” campaign with signs based on popular movie posters.

Ripple effect – standing in front of the cafe led to teams in Europe and Asia all the way up to VPs they never would have reached.

Amy Affelt talked about clipping and alerting services – marketing the information center, rainmaking for new clients, competitive intelligence, making the users look smarter.

It’s not enough to send feeds, you have to customize and personalize, demonstrate the value with the service. Don’t wait for people to ask for alerts, just do it. Best sources: Wall Street Journal, New York Times, Financial Times, niche blogs from industry insiders, requesters’ favorite publications, table of contents.

Delivery and timing – on the commute, via handheld, articles as separate documents with one sentence summaries, with customized informative subject line.

Be indispensable, minimize the work your users have to do.

Questions:

• What does Chris think about the quality of Vivisimo’s clusters?
• How long does it take Amy and her staff to create alert (I think that was the question)
• Did Chris have to get permission from vendors to federate their products?
• How far up the chain of command did Amy and Karen get up with reports?

10/23/08 UPDATE: Karen and Emily’s presentation is up (link direct to PDF)

Technorati Tags: IL2008

Deb Hunt presented this session. The first two case studies were for corporate environments (an environmental engineering firm and an architectural/design firm) , the third for a non-profit (The Exploratorium). WiFi went up and down throughout the session.

Universal challenges:

• Too many information silos
• Dirty data
• No metadata/classification nor taxonomy in place
• Differing needs of different groups
• Multiple physical locations with differing protocols for storing information
• Loss of intellectual capital when people leave
• [One other thing I didn't capture in time, but the presentation is available online at ]

Environmental firm had a traditional library of externally-published documents (both print and digital), five offices in California with plans to expand to Nevada and Oregon. Security of proprietary information was key.

Deb interviewed the staff to understand the information-seeking behaviors of different groups and levels and what issues those groups and levels had, then presented at the annual all-staff meeting. She only spoke for a few minutes, the majority of the presentation was staff people talking. Two people had created their own databases, yet nobody else knew about them. One person had a LexisNexis subscription which nobody else knew about.

She used a number of sources to research appropriate systems (Capterra’s Library Automation Software Finder, Marshall Breeding’s Library Technology Guides, article in October issue of Computers in Libraries) and created an RFI and spreadsheet for vendors to fill out. Narrowed candidates down from 12 to 3, only one candidate allowed the company to retain proprietary documents in its system. The firm is using a cataloger to catalog print and digital items (with items prioritized), instituted a Dumpster Day, the intranet is being redesigned and marketed to staff. The firm will continue to train staff, market the portal and get buy-in for contributing documents to be catalogued. Education needed (not everything is digital).

The architectural/design firm has six U.S. offices and one in Asia. The staff is young and uses Google to find images and information (despite the firm’s already owning similar information). No information professionals on staff, lots of silos, badly-designed intranet which nobody used, and each office had its own culture for sharing and maintaining documents.

The CEO wanted a simple “Google-like solution,” hates the end-user searching in Canto Cumulus’ image database. He had a very naive view of searching. She got advice not to use Google Search Appliance (not cheap, hard to get support) and brought that to the CEO. She knew she was looking for enterprise search solutions, and identified 43 possibilities from Capterra’s main enterprise software directory and input from colleagues. After sending an RFI to the 43 vendors, she narrowed the set to 19, and then to 8 solutions in 3 tiers. The number one choice was a partial solution, as it only handled project management, but it did have the architectural/design focus.

Deb recommended they hire a librarian/information professional, the firm wanted to hire another IT person but she talked them out of it. She wrote a job description and had it posted to multiple job sites, and the firm hired a recent graduate of San Jose State University’s SLIS program who had worked part-time at an architectural firm for 10 years. The firm is on the verge of implementing a solution, despite the CEO initially not having much faith in finding a good solution the first go-round.

They are looking at open-source content management systems and portals.

The Exploratorium was an early adopter of the Web and has 577,000 in-person visitors/year. They began looking into knowledge management in 2003. Challenges included multiple format contents (including print, images in both digital and print, Hi8 video, VHS, U-Matic, audio cassettes, audio reel-to-reel), some of which no longer had players in production. Content was also in multiple places. The digital asset archive has a clunky end-user search but staff use Canto Cumulus (both a client and an internal database). 1.8 FTEs in media archiving, 0.45 full-time equivalents in knowledge management and 2.75 full-time equivalents in the Learning Commons. The Exploratorium is still struggling with enterprise search, but the intranet is now the main source for information internally and is well-marketed.

Key takeaway – there is no one-size-fits-all solution.

10/25/08 UPDATE: Deb has her presentation up (direct link to PowerPoint).

Technorati Tags: IL2008

As he did last Internet Librarian, Danny Sullivan talked about trends in search and what he saw as future developments. Before that, Cindy Shamel presented the AIIP Technology Award to 10KWizard (and Marie Varelas from my local chapter of SLA accepted the award).

Danny sees Google as ruling the search roost for at least the next five years, and being the dominant tech company generally. He talked about the transition of tech leadership from Microsoft to Google, similar to the transition from IBM to Microsoft. Danny gave the example of Cuil.com as a failed challenger, along with Powerset and Microsoft itself.

Google has a 60%-70% share in the U.S. and it’s higher in many other countries. 80-90% share of website traffic is coming from Google.

But there are some areas other search engines handle better, and they might get more attention.

• Summize searches Twitter in real-time.
• Urbanspoon is a mobile search engine which randomly selects restaurants near you (using cellphone tower triangulation). A similar tool is Chowhound.
• For events, Eventful and Upcoming (I wondered about microformats and whether Google would be able to drive adoption).
• Yelp offers local reviews – Google Maps is trying to grow the equivalent.
• Trulia and Zillow for real estate information.
• Kayak is a multi-site travel search engine.
• Farecast offers predictions of fare prices.
• Craigslist for buying/selling items, personals. Google tries to compete with Google Base, but not there. Looks pretty, but the information isn’t there.

And there are even more options out there:

• Jobs (Indeed.com)
• People (Pipl.com. Spock.com)
• News/Discovery (Digg.com)
• Video (Blinkx.com, VideoSurf.com)
• Gas prices (GasBuddy.com)

Hard to remember all the smaller search engines out there and what they do.

Yahoo! innovates (mobile search, BOSS, SearchMonkey) but its future is uncertain.

Microsoft is focused more on ads than search and is exclusively focused on consumer search, has major branding problems. There is some good stuff, but will people notice and will Microsoft search grow?

Google’s master plan? Look at Danny’s article on the Google hive mind. Google might focus more due to the current economic problems, and of course ads everywhere.

Other Google search initiatives:

• Google Video has been a video meta-search for about a year.
• Universal search mixing of results continues.
• Google Trends continues to grow and provides data on website traffic now.
• Community editing on maps grows but so is map spam.
• Google Blog Search clusters stories and identifies top headlines.

Google is customizing searches based on geographic location, previous queries, Web history. It’s done some of this for a while, but it’s being more explicit about this and will likely be doing more customization going forward.

Mobile and vertical search do offer new opportunities for search, but generally you’ll have Google dominance in search with Microsoft a very distant number two.

Technorati Tags: IL2008

Tom Reamy, who presented on folksonomies and tagging last year, led this session. Since it’s in the San Carlos Ballroom, back to no Wi-Fi access. Then again, from what I’m hearing, even if there’s Internet access in a given room it’s spotty at best.

His talk is about integrating semantics, taxonomy and faceted navigation, with a look at what works and what doesn’t for media sites.

Facet – Orthogonal dimension of metadata
Taxonomy – Aboutness of documents
Ontology – Relationship between entities and facts
Software – Text analytics and auto-categorization
People – Tagging and evaluation of tags, fine-tuning rules and taxonomies, social tagging and suggestions

Facets are metadata attributes (e.g., people, location) and are not identical with categories (which are limited in number and involve aboutness). By orthogonal Reamy means mutual exclusivity –> an event is not a person is not a document is not a place. Facets have a variety of units and structures, and are designed to be used in combination.

Faceted navigation is more intuitive to end-users and allows for dynamic selection of categories (rather than forcing the user to go down a single path to find information). It involves fewer elements – 4 facets of 10 nodes can yield a 10,000 node taxonomy, it’s flexible and easier to maintain.

Taxonomies deal with semantics (meaning, aboutness) and documents and are complementary to facets. Taxonomies support multiple meanings and purposes, and can be relatively small if combined with facets. Formal taxonomies work better (is-a-kind of, is-a-part-of) than broader classifications.

Ontologies deal with relationships between entities – e.g., Vice Presidents have employees and bosses. They can be represented with XML, RDF, OWL, inference rules.

Best approach is dynamic search and browse. Reamy gave the panda, monkey, banana example (which two terms go together best) which I seem to recall Dave Snowden using at KM World last year.

Sample sites:

• Wine shop (pure facets – uses sliders, which is increasingly common)
• Search engine (Source facet – news, video, pictures, Web – with a few selection filters)
• CNNMoney.com (Source facet again, doesn’t allow for detailed drilldown, too many sponsored results)
• Search engine (includes Source and Date facets)
• New York Times (uses semantic technologies to suggest search terms and to offer related stories, has a Source facet identifying section of the ‘newspaper,’ but semantic technologies have issues – Obituaries is one section Obama appears in)
• Forbes.com (chaotic interface)
• Factiva (true faceted navigation – uses Source facet but has sub-categories, graph showing number of documents in a given date range, clusters of co-occurring terms displayed in a tag cloud, multiple traditional facets such as company but also a Subject facet
• Financial Times (limited number of results, but auto-summarization. Standard facets but also taxonomic elements – Topics)

Common themes:

• Balance of commerce and information
• Basic facets – Source and Type
• Standard (People, Companies, Place, Industry)
• Interactive interface (sliders, date ranges)
• Keywords vs. simple taxonomy
• Tag clouds/clusters – how genuinely useful are they? Reamy hints the answer is not very, but doesn’t go into details.

Common issues:

• Advertiser dominance
• Auto-ads
• Non-orthogonal facets (Topics and Issues for one client)
• One or two filters (don’t provide enough intersections)
• Semantic component is still the hardest
• Good information architecture – Summary or full facet display? Simplicity vs. research power

Design issues:

What is the right combination of elements? What is the right balance of elements (Are all facets treated equally?) When should elements be combined – before or after search?

Tools and approach:

Text analytics software extracts entities and noun phrases from a document or set of documents, and can be designed to feed facets,signature, ontologies.

Auto-categorization software feeds subject facets, can be ‘taught’ using training sets, a set of terms (with literal strings, stemming and a dictionary of related terms), simple rules such as position in text, saved search queries, Boolean. Advanced features can include fact extraction, sentiment analysis.

Entity extraction can be dictionary-based, rules-based. A collection of entities can be the aboutness of a document.

Documents are more complicated than products! Facets can’t be an add-on. There has been some progress on semantics. Future of search will be smarter ways to refine results, not better relevance.

When do you add metadata and how? Depends on the environment. Using content management in the enterprise one can balance taggers, results from software and company policies. Software can suggest categorization and facet values. Relevance is best based on ontologies.

10/27/08 UPDATE: Tom’s presentation is up.

Technorati Tags: IL2008

Sophia Guevara and Qin Zhu presented this session. As Jane Dysart had announced before the keynote, WiFi was available in every room except the San Carlos Ballroom (where the keynote was). So back to liveblogging.

There are many definitions of digital libraries, so they suggested starting with an understanding of a given library’s users, an understanding of what the goals are of the organization and of the users and how the library could align its services with those goals, and an understanding of digital content and the digital collection.

Physical and space constraints, changing user information needs and changing information behavior may all drive digital libraries. Digital content may include e-journals, e-books, e-reference, image collections, digital audio and video and electronic databases.

The lifecycle of digital content, as expressed in a 2005 article written by Tamar Sadeh and Mark Ellingsen (available as a PDF):

• Discovery
• Trial
• Selection
• Acquisition
• Access
• Renewal or cancellation

Questions to ask include bibliographic details, terms of access and pricing. Analysis of content – granularity, dates of coverage, subject focus and overlap with existing sources.

Content analysis resources:

• Content provider sites
• Electronic Resource Management systems
• ILS
• Comparison tools, either existing (CUFTS, databases, Excel) or home-built

Assuming everything checks out, user feedback is positive and a vendor is found that can provide an option for filling a given information need, some questions to consider:

• Can content be purchased a la carte or only as part of a package?
• Can a better deal be negotiated by agreeing to a multi-year purchase?
• What kind of access is available (per-seat, per-site, etc.)?
• Try to negotiate out auto-renewal clauses

Licensing 101:

• Contract template available at Yale’s LibLicense site
• Connect with procurement and legal departments and agree on what will and will not be acceptable
• Understand access and pricing options
• Be flexible

Resources:

• LibLicense site
• two other things I didn’t catch because the presenters went too fast

How will you provide access to users? Be prepared to complete routine maintenance and encourage feedback from users.

There are different methods for content authentication – IP authentication, URL referral, username and password. There may be a single-sign on for your institution, the Athens access management system or a proxy server.

You might want to provide several different information access points – A to Z lists, pathfinders, subject guides, journal/e-journal lists, your library catalog, RSS feeds, etc..

Is content searchable at the top level and as full-text? Is it findable at the publication level? How will you integrate it and aggregate it with content from different providers? Journal lists? Library catalogs? Federated search or alerts? RSS feeds?

Electronic Resource Management Systems, SUSHI or COUNTER can help collect usage statistics. Measurements include site visits, page visits, time spent on site, downloads, etc.. Study your log files, understand user behavior and how they access content.

Criteria for deciding whether or not you will renew a resource:

• High cost per use ratio? If so, does another vendor offer the resource at a better cost/use ratio?
• What value is coming from price increases?
• How frequently do users have problems with the product and how does the vendor respond to problems?

Questions:

• At what point do you decide keeping print resources is a better deal than trying to choose between different electronic resources with multiple interfaces?
• What do you do if you don’t have the final say on purchase decisions in your organization?

Apologies if I missed anything, the speakers went very fast. Hopefully at least some of the missing content will be available whenever the slides are posted.

Technorati Tags: IL2008

Howard Rheingold spoke on “Communities & Communication in a Social World.” Lawson and I got there early enough to grab seats near the front. Equally importantly, the seats were by a power strip.

I took a photo of Howard’s massively-stickered laptop before things got started – Lawson was smart and actually went up to the podium to take his photo.

Through black magic he also took a photo of me which I actually like.

Howard sees the increasing importance of cooperative arrangements and complex interdependencies. He talked about some of the things which led him to this conclusion. Several examples were from Smart Mobs, but he also talked about the Spanish protests against their government blaming the Madrid bombings on ETA and the Penguins’ Revolution in Chile. But there were also negative examples – the SMS-organized protests in Syria and Egypt against Denmark because of the Mohammed cartoons, the racist riots in Australia, the riots in Nigeria around the Miss World contest.

Howard traced the development of social forms based on collective action.

• Prehistoric hunting strategies (family groups and allies)
• The development of cities
• Alphabetic writing (limited to an elite)
• Development of the printing press in Europe (mass audiences)

This is still going on – new kinds of cooperation and wealth creation (e.g., IBM’s embrace of open source, Eli Lilly’s opening up of research problems to the wider community, Google opening up its ad network, Amazon.com opening up its API). Howard thinks this will definitely spread beyond the computer industry.

ThinkCycle – nonprofit means of tapping the world’s design student community – came up with a vastly-improved hydration system. SETI@Home is the most powerful computer in the world (40 teraflops). Emergent collective response to disaster – Asian tsunami blog for organizing aid, Katrina people finder wiki. Trying to find Jim Gray’s sailboat using photographs from Google and NASA, Microsoft and Amazon.com put up sections of the images on Amazon.com’s Mechanical Turk and 12,000 volunteers looked through them.

Barriers to participation continue to be lowered. Technologies which enable cooperative and sharing economies are easy to use, enable connections and group formation, are open, are self-instructing and leverage individual self-interest.

Participatory media are social media which open up knowledge and wealth creation. There is nothing innate in knowing how to apply skills to education, politics, etc.. Education must recognize a new way of learning and teaching, not just tack on new skills to an existing curriculum. Howard and others are developing curricula, syllabi and teaching notes as well as a social media classroom and a community of practitioners.

Don’t try to keep up with the technologies, keep up with the literacies.

http://socialmediaclassroom.com, or contact Howard at his first name at his last name dot com.

Technorati Tags: IL2008