They did say that in recent years virtualization has matured enough to be a university-wide solution for user computing. GA Tech showed that over 2 years students started being regular users of the system with approximately 30% of the users of the system becoming regular users (51-100) uses per semester. This led to a few issues including the need to scale up their infrastructure. One of the neatest features they discussed was the ability to automatically scale citric clients to ipads and other mobile devices. Finally – they discussed what it meant to provide a complex service over the network – network, capacity, connectivity issues become very important.

I followed up with a session on talent management. Thomas discussed TM as an approach to employee growth and investment that helps organizations identify and gain talents. The talent management approach not only focuses on core skills and job match but also traits, behaviors, and talent. He differentiated between baseline and role specific talents. His talk reminded me of a presentation on motivation that I watched recently link. He recommended book called Best practices for talent management. Here is a link to the full presentation.

Here is a screenshot of my Ubuntu desktop via vnc swell as one from osx.

General impressions:

• Vnc had memory issues running full resolution on osx and screen refresh issues with gnome
• Using the mouse takes some getting used to – sometimes you control the iPad and other times you control the mouse
• Video steaming was ok. Good enough to keep track of what was happening.
• Keyboard overlay disrupted screen display
• Usable? Barely but there is a lot of promise there.

Update – a good comparative review was done by tuaw http://www.tuaw.com/2010/05/03/hands-on-with-ipad-vnc-clients/

No matter though – after a few minutes of recovery in the lake everyone was smiling.

Fun race and a good warmup for while lake in a few weeks.

At the lump (pictured) I came across someone flying model glyders. He showed me how he could toss a glider over the hill and find a thermal to climb with.

In all I got in 75 miles – not a bad start to the spring. From the looks of signs on the window, Parkway facilities are going to open up in late april so until then all rides are self-supported

Most of the race was on the American tobacco trail which provided a nice shaded and quiet route. The out and back format did hurt a bit though and although I had a good pace going until mile 21 I could feel the wall coming. I did a slow last 5 miles but managed to hold 3:44. Next race is the belewa lake tri and then white lake half on may 1.

People, their digital stuff, and time: Opportunities, challenges, and life-long challenges

Code4Lib 2010 kicked off with a keynote by Cathy Marshall.  She mentioned a few interesting stats (a webpage weighs 80 micrograms – Brewster Kale), 4.5 Billion personal photos in flickr, and continued to talk about how people preserve, use, and manage their digital artifacts.  BTW – her talk was titled – People, their digital stuff, and time:  Opportunities, challenges, and life-long challenges.

For me a compelling question she asked was – “digital lets us keep everything – should we?”  She had a good point – you never get rewarded for deletion, but sometimes you get rewarded for saving.”(ala Gordon Bell).

She talked about a study she did on massed data where she typed the metadata (e.g. place, artifact, context) that was assigned to digital objects.  The interesting aspect of typing metadata assigned to digital objects is in thinking about metadata assignment within the context of teh idea that people don’t approach digital object description/archiving in a deliberate and systematic way.   Her final message was that “new opportunities lie in the aggregation of individual archives and efforts.”

cloud4lib – Jeremy Frumkin, Univ Ariz., Terry Reese, OSU – An open library platform.

They are working on building this at http://cloud4lib.sourceforge.net. They have setup a collaborative workspace which they are running out of OSU libraries using a research account on Amazon EC2. They are having a breakout session at code4lib today to talk about how to structure this idea.

Ross Singer – Linked data

Ross talked about how he took MARC data & built a linked data service. Too complicated to talk about in depth here but you can find his presentation on the code4lib site. The really neat thing he did was take a vufind instance & embedded some RDFa into the template to make the linked data discoverable.

Harrison Decker – DIY cloud computing with Apache & R (rapache)

Data librarian, UCBerkeley. Like many things – data analysis in libraries is getting more complicated – libraries have an opportunity to help people learn how to analyze data & how we can use data analysis in our own work. His definition of the cloud “a replacement for the desktop, when it makes us work smarter, extendable.” He asserted the idea that decision makers are not statisticians but rather need processed charts & data. Rapache is an apache module with the R interpreter compiled into it, lets you embed an R script in web pages, provides interface to get/post (baseball example). Some ideas he had for use – interactive we/e-journal use visualizations, real-time survey results, data visualization for instruction, network analysis. . .

Public datasets in the cloud – Rosalyn Metz, Michael B. Klein

Their cloud definition ( on demand self-service, broad network, resource pooling, rapid elasticity, IAAS, PAAS, SAAS). She demoed launching an EC2 instance & mounting datasets hosted on EC2 for analysis. She mentioned socrata, Google fusion tables in which you can create tables and visualizations of your data. Michael talked about data access & issue and wondered where/how libraries might get data & asked what the role of IRB might be in data analysis.

Karen Coombs – OCLC – 7++ ways to improve library UIs with OCLC web services

Karen had a laundry list of ideas for using OCLC web services to extend & embed data in library discovery services. They are represented briefly – mostly because she flew through them!

• Cross listing print & electronic stuff – Use WC search api to see if lib ha print version & link it to the full record of ebook
• Item availability localization – Take zip code, OCLC number, bring up a google map, link to library catalog
• Bring Journal TOC to feed to MARC record – XISSN to see if feed of recent content is available
• Peer Review indicator – Use data from XISSN to add peer reviewed data to appropriate screens
• Information about Authors – Use Worldcat Identities, wikipedia API to get author information, link together
• linking to free full text – use XOCLCNum to check for free full text from OCA, Hathi Trust, etc
• Similar items – Use dewey decimal classification, WC search api to retrieve similar items
• Bonus – build a mobile catalog – use WorldCat Search api to get to your library catalog data, build in images from open library, ratings from librarything
• links – http://worldcat.org/devnet, http://www.oclc.org/applicationgallery

Extensible catalog – Jennifer Bowen

The afternoon started off with an overview of the eXtensible Catalog. They have developed a suite of tools including user interface, metadata, and connectivity (NCIP, OAI) to serve as a holistic replacement for discovery & management services. XC includes frbrized metadata, “why you got that record” in the index display, works as a drupal module, includes some nice complex staff metadata management tools so that staff have flexibility in defining how the metadata is displayed. The toolkit (Bowen indicated that there were lots) allow you to automate data loading and processing! It appears that right now XC is still in a semi-release phase but all software is available.

Conference fatigue has set in – The rest of the sessions are discussed in abstract here

Ok, I take it back. Jeff Sherwood talked about using Levenshtein string distance as a method for matching records using a mathematical algorithm – very neat. They also usd the Jaro-Winler algorithm for string comparison. code is at http://pypi.python.org/pypi/editdis/0.1, but.ly/ZGSmF, secondstring for Java, MARCXimiL – a MARC deduplication package, http://snurl.com/uggtn.

During the break we had an opportunity to chat with Evergreen & Koha users & found that the Acquisitions and serials implementations are in process (evergreen) or functioning (Koha). We had an interesting discussion about the utility of serials check-in.

During the dream-the-dream segment some ideas for best world scanrios included lots of talk of system integration – integration of digital library stuff with the ILS< integration of ERM stuff with the ILS, the integration of Koha/Evergreen.

After a lunch at Salsas JP & I split up. He hit the Blacklight session & I went to Dan Chudnov’s tools session in which we learned some python programming, talked about Django, Kohana, and the overall value of testing your code

She talked about Duke, Northwestern, International Children’s digital library, Van Gogh Musueum, and some prototypes (UNLV digtial library, Europeana). Several of the applications focused on delivering digital material (images, metadata, copyright info, locations) via an app. Of note, the ICDL focused on delivering ebook content (she mentioned another app called StoryKit that looked very neat). One of Wang’s assertions was that a valuable feature of the ICDL apps were that they could be used without network access.

Her summary of ‘killer features’ from her survey of apps included

• Image sharing
• Ebooks
• Audio and video
• Search – keyword & location
• Insights: added value
• User contributed data – tags, comments, sharing, creating

Wang talked about the growing trend to create web applications as opposed to ‘native iphone apps’ & mentioned a number of frameworks including iUI, iWebkit, JQTouch, and Dashcode.

Kevin had recently shown me JQtouch, a framework for writing iphone centric web applications using safari as a primary platform & it simplifies the process of developing apps that benefit from the iphone hardware & interface features.

Wang quickly demoed Dashcode & PhoneGap (which includes development platforms for multiple platforms & allows you to migrate a dashcode app to an objective c app) In 10 minutes she built a neat little app based off on an RSS feed. I was so excited I went home & wrote a ZSR library app off our of new-videos RSS feed (Please use your iPhone to view). . .Thanks Minglu!

Some services each aspect of the system include:

UCPubS – sales, fulfilment, distribution of print, online marketing, print on demand

Departments & UC partners – selection of content, peer review, manuscript management/composition

eScholarship – Open access publishing, peer review, and preservation services.

The focus of UCPubS is on working with partners to build a sustainable aggregate/archive/publish/market model. An example of a partner is the Global, Area and international Archive. They indicated that GAIA used UCPubS to help re-frame scholarly publishing in their field.

Other comments from other attendees?