Oren’s Weblog

Brad Wheeler from Indiana is talking about the Twin Peaks Navigator which is intended to connect course management systems with repository content.

One of the issues this hopes to solve is the issue of students having to buy course packs containing photocopies of data the institution already has online.

They wanted to be able to store pointers to all sorts of digital content, even if it’s in a protected source such as a library database.

The idea is for a faculty member to be able to easily embed links to data sources within the online course content. In the implementation there’s a button that invokes a library search within the WYSIWYG editor. Very cool!

http://www.twinpeaks.libraryinfotech.org/
https://twinpeaks.dev.java.net/

The hope is to have this integrated into the next version of Sakai.

Brad Wheeler from Indiana is talking about the Twin Peaks Navigator which is intended to connect course management systems with repository content.

One of the issues this hopes to solve is the issue of students having to buy course packs containing photocopies of data the institution already has online.

They wanted to be able to store pointers to all sorts of digital content, even if it’s in a protected source such as a library database.

The idea is for a faculty member to be able to easily embed links to data sources within the online course content. In the implementation there’s a button that invokes a library search within the WYSIWYG editor. Very cool!

http://www.twinpeaks.libraryinfotech.org/
https://twinpeaks.dev.java.net/

The hope is to have this integrated into the next version of Sakai.

Mimi Yin from OSAF has posted a great set of slides titled Chandler In A Nutshell. These slides show the real innovative work the the OSAF folks are doing on re-envisioning the ways in which people actually interact with their personal data. It’s worth looking at this.

David Millman from Columbia is giving an update on MIT’s DSpace repository. David notes that the first year of DSpace was a focus on how to approach the idea of an institutional repository, not particularly on the technology.

There are 103 live DSpace sites, 38 in the US and 65 outside the US. Those numbers are from a CNI survey.

Tim Sigmon from UVa is leading off with an update on the Fedora repository. Tim’s slides are here.

The repository functionality exposes a number of APIs through web services: management, access, search, index search, and an OAI provider. Authn and Authz modules will be coming out in release 2.1, at the end of this month.

Different diseminators “vend out” different views of an object – for instance, a gif and a tif of the same image.

Version 2.0 is the current version, came out in Jan 2005. It changed the underlying description of storage to an XML description called FOXML. They also added representation of object-to-object relationships with an RDF schema – that’s the “resource index”. For instance you can store that objects are members of a collection. The RI search allows querying of those relationships.

In 2.1 they’re converting from a Mozilla license to an ECL license. It will have Authentication Plugins for Tomcat user/password, LDAP, or Radius, and an Authorization model using XML-based policies written in XACML. Policies can be repository-wide or on specific objects. There will be a Policy Builder Client with a simple UI to create access policies that will generate the XACML. They’re also adding a Directory Ingest Service and client tool to facilitate ingest of hierarchical directories and zip or jar files. The tools will allow automatic representation of that hierarchy in Fedora.

Fedora was not built with an easy UI in mind – it was built as a repository that applications could be built on top of. But development will begin this fall on an web-based Institutional Repository client.

They’re planning a lot of work on federating repositories. Shibboleth implementation is almost done.

VTLS is offering commercial support for Fedora, and have done a bunch of work on enhancements, including making Fedora repositories visible to Google searching.

Usable design expert Donald Norman has written an interesting article that discusses the shortcomings of the currently popular approach to designing systems based around human input, usually called Human Centered Design or User Centered Design. I think he articulates far better than I can some of the factors that cause me to be uneasy with a blind commitment to UCD methodologies. Definitely worth a read.

One basic philosophy of HCD is to listen to users, to take their complaints and critiques seriously. Yes, listening to customers is always wise, but acceding to their requests can lead to overly complex designs. Several major software companies, proud of their human-centered philosophy, suffer from this problem. Their software gets more complex and less understandable with each revision. Activity-Centered philosophy tends to guard against this error because the focus is upon the Activity, not the Human. As a result, there is a cohesive, well-articulated design model. If a user suggestion fails to fit within this design model, it should be discarded. Alas, all too many companies, proud of listening to their users, would put it in.

Here, what is needed is a strong, authoritative designer who can examine the suggestions and evaluate them in terms of the requirements of the activity. When necessary, it is essential to be able to ignore the requests. This is the goal to cohesion and understandability. Paradoxically, the best way to satisfy users is sometimes to ignore them.

Richard Bendis is prsident & CEO of Innovation Philaselphia.

Wireless Philadelphia is the local effort to be one of the first totally connected wireless cities in the US. The rollout effort will start this month. Interestingly enough, one of the objectives in the vision statement is to reduce the cost of city government. There will be 2,500 WiFil access points on light poles connected to WiMax installations on rooftops.

The City is viewing this as a way of encouraging and enabling the growth of small and mid-sized businesses throughout the city, as well as enabling learning for students across the city.

Wireless Philadelphia is a separate 501(c)3 nonprofit. It issued an RFP for the network in April. There were two finalists, HP and Earthlink, and decision will be made this month. The entire city network is expected to be available in Summer/Fall 2006. Is this cool, or what?

I came in late (east coast early morning sessions are hard the first day) to a session on identity management to hear Bob Morgan talking about how we’re likely to see two different kinds of federated identity management in the future: one that is driven by compliance and legislation where there are a small number of large identity providers at a national level; the other being lots of small loosely-federated identity providers in an ever-shifting tapestry as needs require.

Now Kirk Brown from Sun is talking about their concept of federated identity management. While there’s not a lot of detail here, it strikes me that their solution is a broker between multiple identities and identity providers. The example he gave of this in use is at Wells Fargo in their bill-paying service, which brokers lots of identities behind the scenes.

Mike Jones from Microsoft describes himself as a “protocol evangelist” within Microsoft. He’s talking about how identities work within contexts – coffee cards work at a given coffee stand, bank cards work at an ATM, etc. He gives the old example of using a SSN as a student ID as an example of misuse of a context. The lessons MS learned from Passport. Posport was designed to be an identity provider for Microsoft’s online properties – where it’s been a huge success (250+ million users). But it was also hoped to be a global Internet id provder – where it was a complete failure, for social and political reasons, not technical.

Mike is going through Kim Cameron’s Laws of Identity. The conclusion they came to from those laws is to define a Metasystem of identities – something that will do for identity what IP did for defining a common layer of internetworking. Like IP did not replace ethernet, this metasystem will not replace SAML or Kerberos or X.509. This is the basis of WS-Trust and WS-MetadataExchange web services. He showed a mockup of how this might look to a user, where information about the identity of a site is presented to the user for their perusal so they can decide to accept it or not, and then the user can choose which of many of their identities they want to present back to the site. They’re talking to Apple, Mozilla, Sun and others about these protocols.