The Corporate Librarian

While I continue to look for work (interview next week), I try and stay in touch with the profession. Except that’s not as easy as it sounds.

SLA has the Annual Conference (June 2008, in Seattle) and Leadership Summit (January 2008, in Louisville). I have no idea when the next regional chapter meeting will be - I’ve volunteered to help with the website, to do more in the association and to meet more people, but that’s still largely up in the air. I don’t hear from the divisions or sections I’m in much, though I try to chip in on the IT Division’s Blogging Section blog and I made sure to meet Jill Hurst-Wahl at Internet Librarian. I’ve attended two conferences this year, which is more than many people get to, but it’s a big financial hit. I’m forced to carry a balance on my credit card, something I really hoped I’d never have to do.

Most of the big-name library blogs are for academic and public libraries, so I don’t really feel a part of the conversation there. Ditto for LSW on Meebo. Hell, this blog pretty much only gets comments from vendors wanting to push their products. At least Internet Librarian has sessions for corporate librarians, and I think a lot of Library 2.0 people could learn a lot from attending KMWorld & Intranets and learning from what’s gone on before.

Mailing lists? I’m on actKM, BUSLIB-L, the one for my SLA chapter and the one for AIIP members. I’ve chipped in on BUSLIB in the past, and may do so on AIIP if a topic comes up I feel comfortable with. But answering reference requests isn’t really what I’m after.

Where’s a venue for corporate librarians and knowledge management types to just hang out, talk shop and commiserate, regardless of geographic location on an informal basis?

Tom Reamy’s full presentation on folksonomies and tagging from Internet Librarian is available on the KAPS Group website.

I’ve got pushback on it in a few areas, but check it out and see what you think.

Technorati Tags: il2007

Not all of them by any means, but a chunk are posted online (mostly as PowerPoint slideshows) on the Internet Librarian 2007 site.

Technorati Tags: il2007

TouchGraph is a neat Java app which graphs the connections between websites (blogrolls definitely are one factor, I’d have to see what else). For the heck of it, I decided to compare my connections with those of a real library blogger:

Discuss.

I figure now that I’ve had some time to reflect, I’d post some thoughts on the KMWorld & Intranets conference last week.

For me personally, it was worth it on the whole. I still need to work on the whole networking thing, but at least I didn’t grab people and force my business cards on them - I listened to them, spoke up when it made sense and so on. There was a decent range of backgrounds for attendees, though I suspect there’s a heavy corporate IT contingent. Because it was at the San Jose Convention Center, I just had to hop on the Light Rail and cross the street - no hotel needed.

I liked getting a lot of the conference presentations in printed form, so I could focus on the discussion and the points brought out not in the PowerPoint. Even if my note-taking does suffer from the “Must organize things in bullet points” mode I learned in consulting.

That being said, I did find some things I’d improve in future conferences. Wi-Fi access in the conference rooms, not just in the hallway outside. A meeting board, so people could arrange to meet up, post job listings, etc.. Power strips, and lots of them. More publicizing of the wiki (beyond a brief notice at the bottom of a page in the middle of the conference guide) and announcing tags for the conference for Flickr, Technorati, etc..

While I worry about how it might affect participation in sessions, I’d like to see some experimentation with official backchannels, whether they’re in the form of IRC, web chats or what have you. I think they could get a lot of basic questions by people answered while leaving time for more complex back-and-forth.

The conference sessions I got the most out of tended to incorporate interaction with the audience while the session went on, rather than limiting things to a Q&A session at the end. They also were not just product pitches - I’m beginning to come around, reluctantly, to Dave Winer’s way of thinking on vendor pitching in conferences. The Google rep pooh-poohed ‘discovery’ and presented search as the be-all and end-all. Which is obviously only true if you know what you’re looking for. A lot of what he said resonated with previous presenters (publishing and sharing what you can, the importance of collaboration) but it was framed as “And this is why you should use Google Apps!” Swell, the average number of keywords in a Google search is down to 1.7. What does that actually mean?

One of the things which fascinated me about the BEA session was that the vendor took a class of technology based on trust and collaboration and stressed the advantage of being able to stop experimentation by users and lock down what people could see.

In some respects a lot of what I heard was similar to discussion in the 1990s - the need to get out of users’ way and act as facilitators, rather than imposing a system on them (though, as David Snowden noted, some boundaries are needed). Innovation was a frequent topic, I suspect due to companies worrying about their prospects for sustainable growth. What fascinated me was the tension between people’s discussions - Are communities of practice highly useful or no better than chance would suggest? What’s the balance between organizing and understanding what we already know and looking outside the walls of the organization?

I left with interesting questions I wouldn’t necessarily have come up with on my own, and I hope discussions continue after the conference.

Technorati Tags: kmw07

Cyrus Mistry gave this presentation on what next-gen collaboration looks like and how organizations can maximize productivity. I was honestly a bit nervous that this would be a product pitch, but Google is so important in the areas I’ve worked in and hope to work that I felt obliged to attend.

Why can’t you have cool fun technology at work? Why is it limited to the consumer space? Business technology user satisfaction is declining over time (so he says), 75-80% of tech spending is maintenance. Most of the top startups (80%) are in the consumer space.

Google does not have laborious deployment of IT - experimentation and iteration.

Technology must be simple to implement (appliances, Software-as-a-Service, minimize customization)

Platforms have to be able to deal with constant change.

OK, he’s getting into product pitching. Grr.

• Manual categorization and hierarchies are dead (Um, no, they’re not - mix of structure and tagging, guy). He shows an array of email folders versus Gmail (you can search your inbox), files on desktop versus Google search. Internal Google startup page is search bar plus mashups bringing in content. Search is a way of navigation.

• He dismissed discovery - said you generally want an aggregate - you don’t want all the pages on weather, you want to know the weather in your location, in a simple display (but how is that context derived?)

• The goal is to answer the user’s question with search = average number of keywords in Google search is down from 3-4 something to 1.7

• Applications (pitch for Google Apps) - simple interface, collaborative, going for the 80% common tasks (so yes, Google Docs doesn’t have mail merge, as an example he gave)

• How to deploy to enterprise - start small, with official intranet. Add a directory search, then high-value (can’t make out my writing here - battery had died and I was hand-writing notes on my program).

Questions:

• Why is there this gap in usability between consumer tech and the enterprise? (Designing for features rather than users)
• What will the impact of natural language searching be? (Increasing impact)
• Timeline for blogs/wikis available as part of Google Apps? (Doesn’t know)

Technorati Tags: kmw07

Chris Rivinus, Director of Knowledge Systems for Parsons Brinckerhoff talked about visualization (surprise!) and the advantages Parsons Brinckerhoff had found in using it.

Selling the vision (marketing) - Pictures to moving pictures –> conceptual phase of project –> concepts handed over to design team –> models becoming more exact –> parametric object modeling.

(showed video of a visualization of Dubai’s Palm Island project, would be used in pre-sales to public and officials)

Exact model captures and transmits project details across discipline, experience, language barriers.

(showed video of a visualization of the Fulton Street Transit Center project)

Digital models using CAVE technology, hand-me-down gaming environments to experience the project before it’s built and to work through design issues with clients. Merging of engineering/graphics design.

(showed video of a walkthrough of the integration of the various World Trade Center designs, 1/4″ accuracy, can pull off skins)

2d to 3d design, can work out design conflicts, communications tool between engineers and construction managers. More common in Europe, where much more of a vertical market due to legal issues around handoff.

(showed video of visualization of the World Trade Center, showing conflict between 2 designs)

Normally, conflict between designs would be discovered on-site, weeks of delay. Process of resolving the conflict only took about 3 hours.

4d modeling - 4th dimension is time, works out phasing conflicts and problems, full integration of project schedule with phased renderings.

(showed video of visualization of the WTC, showing scheduling conflict)

HIgh-density knowledge exchange - Israeli Ministry of Finance, risk management tool, exchange of highly-specific high-density technical information, shared experience. Visualization requires shared inherent value of symbols.

Social network analysis - why hasn’t it caught on? Managers can’t understand it without explanation.

(showed video of visualization of Alaskan Way Viaduct project, several other videos)

Questions/Comments:

• What applications do you use? How do you make it work? (Evangelism, good clients, people with reputation. GSA, Intel, Disney other examples of organizations using this.)
• How do you handle duration of phasing? (Target areas where you’re going to achieve savings/ROI. Contractors and subcontractors don’t pad bills as much when bidding, because they have better information to work with.)
• Have you been able to overlay SNA with actual work-in-progress and achieve benefits? (No, they’re separate. SNA is around back-office stuff. Can push visualization to project managers more in tight-budget situations.)
• How do you deal with the shift in skill sets? (It’s just not that hard, you need a champion/mentor but it’s not hard to pick up the basics. Schedule integration is a trickier bit, because you need to get people to put stuff down, traditionally scheduling and design are separate.)

Technorati Tags: kmw07

Ted Graham of McKinsey & Company presented. I was curious to see his take, since I remember back in the 90s attending an SLA consultants’ group where everyone waited for someone else to talk about their organization’s KM. He did say he’d be able to discuss some info about McKinsey, but his focus was on his experience as Chief Knowledge Officer of Hill & Knowlton.

• Early failure - 1998 intranet (most downloaded item was vacation request form).
• Passive capture of email - all emails sent to an email address was posted in a threaded e-mail archive.
• Web cam gave visibility into what others were doing around the world (London, Hong Kong, New York, Toronto cams).
• Photo display and profile with interesting facts featured on web site of person who had updated their profile most recently (preference for new people).

2002-2003 looked at ways of recording relationships with companies, media outlets.

• Old way - Nag people to reveal relationships.
• New way - Employed BranchIt to uncover relationships, recorded active relationships and one-time contacts (influence network map) from email interactions, used LinkedIn to leverage alumni and as a “return to H&K recruiting tool,” Collective Conversation (blogging platform, visible externally) - people tracked their influence (Google ranking for “communications,” etc.), social bookmarking-ish (tried to leverage del.icio.us, but didn’t want to reveal what H&K was reading, so developed Cogenz as an internal social bookmarking tool to bookmark intranet sites as well as external sites).

Intern got job as full-time employee of H&K because he was a best-seller, there was a worst-sellers list which wasn’t publicized but was used for private coaching and trimming content.

Social network web survey - trying to understand why information flows stopped at particular people. Connected people with good ‘network hygiene’ with experts with poorer ‘network hygiene.’ One of most connected people was in an entry-level position with high turnover.

Sharepoint can show new colleagues, external contacts, keywords based on recent activity and add it to the profile for people to view (opt-in).

(nTAGS?) let people beam each other contact info, identify shared interests, locate each other. At end of conference sent list of people who’d beamed info to each other (so you could automatically get a list of contacts with contact info).

What are you doing to enable connections? He closed with a TouchGraph visualization of his Facebook network.

Questions/comments:

• Identifying hub communicators as well as experts and connecting them - how well did it work? (There has to be some sort of social connection, encouraging connectors to talk about what works and what doesn’t)

Technorati Tags: kmw07

Greg Pepus of Flex Analytics talked about different advanced indexing methods for dealing with text and multimedia content and providing context. He was speaking largely from an intelligence & security aspect.

For text, a variety of tools can be brought to bear to generate metadata. Often systems of systems are used in parallel to generate the “cleanest” metadata for a repository.

For multimedia (video, images, audio) there are a variety of tools - he mentioned:



Pixlogic analyzes, indexes and searches images and video files using a variety of tools to compare elements, identify relationships between elements, etc. and generate XML.

3VR offers video surveillance analysis capabilities (not just facial recognition, but motion analysis, location, etc.) as well as case management.

BBN offers speech-to-text capabilities, and can be combined with Language Weaver’s statistical analysis. Handles foreign language - best to derive context from original language, otherwise lose idiom, dialect.

Call Miner identifies patterns in telephone calls and allows for ad hoc searches.

In response to a question from the audience, he said that currently in the analyst community there are not standardized frameworks and methodologies for combining these tools, though the government is making moves in this direction.

Technorati Tags: kmw07

I know of David Snowden from his work introducing storytelling and narrative approaches to knowledge management, so I was curious what his take would be on tagging.

First Nancy Garman introduced Taxonomy Boot Camp, a conference overlapping KMWorld & Intranets.

Experiment - slide with cow, chicken, grass - close our eyes and raise our hands for the one which is the odd one out. Anglo-Saxon world - 2/3rds of people eliminate grass (not an animal), Southern Europe - tend to eliminate chicken because cow has relationship with grass.

Dave makes the point that social computing is being used as a categorization device (cf. Weinberger - we don’t have to put people in boxes) when it’s really about relationships. Social computing environments are ecologies, evolve.

Old ways of thinking persist longer than we think.

Worldwide creation myth: Patriarchal gods bring order out of chaos (Australia an exception).

Complex adaptive systems - Mutual constraint of system and agents (inherently unpredictable and volatile). In the blogosphere, influenced by proximate relationships. Sensitive dependency to initial conditions, which means you can’t replicate outcomes. Manage the emergence of attractors within boundaries.

Not a matter of choosing one system and getting it right, you let people experiment within boundaries. Examples of children’s party, roundabout.

Not true that everything is miscellaneous, everything is fragmented. Human brain makes decisions quickly, based on 2-8% of what they see, runs it against patterns. Humans are not information processors, we are pattern recognition devices. We assemble and produce fragments.

Fragments have the advantage of being context-independent - so you ask ambiguous questions, get answers from a range of perspectives, assemble your own answer. Reducing ambiguity is not what our goal should be.

Traditional approach to metadata: Created by profession or authors, top-down, removing ambiguity and static.

User-created approach to metadata: Keyword tagging, works well within boundaries. Nouns without grammar or agreement. There are no deep structures in languages - language changes subtly in context.

We’re moving from searching to serendipitous browsing.

Issues with user-created metadata - used by techies and early adopters at the moment. Not thinking of the next generation. Too focused on individuals - the “village idiots” phenomenon (”village idiots” cluster), how do you force cognitive diversity.

Semi-structured tagging (funded by DARPA/Singapore government, anti-terrorist technology lets people look at narratives and view areas that can be changed - managing attractors)

• Why are things the way they are, need to understand the boundaries.
• Types of tags (names, multiple options, key words*, free text, filters)

Implications for KM:

• No communities of practice, no portals - there are not recipes
• Provide the tools (blogs, wikis, freeware support), work with key influencers and senior people, amplify/dampen behaviors as needed. Add tools and training (utilities, HTML training, common document systems) and encourage collaboration, find a project –> start with a blog —> preload wiki with content of blog —> publish.
• Start using narrative, linking and connecting.
• Stimulate networks (5 things you can remember/15 max number of people you can trust/150 acquaintances you can manage).
• Ban e-mail attachments (lose coherence, bad security) - use linking.
• Restrict firewalls - more aware. Example of removing traffic lights, makes people more aware as drivers.

Get people connected. Let them do what the hell they want - they’ll use different things, experiment.

Three big issues:

• Validation (example of Wikipedia, tyranny of crowds)
• Scale (why do we assume tech will scale?)
• Community development (does virtual environment permit necessary experience and development, do we create autistic communities through overstructure)

Three heuristics

• Knowledge is volunteered, not constructed
• Human knowledge is contextual
• We know more than we can say, say more than we can write down (importance of narrative and experience)

*60-70% of key words associated with content not in the content, systems of meaning

Technorati Tags: kmw07