Oren’s Weblog

Kay Pilcher pointed out this CNN article on how phishing schemes are getting more sophisticated.

Apparently these crooks are now installing entries in host files on compromised desktop systems so that even if the unsuspecting person types in their own URL in the browser they can be directed off to the scam site.

And there are even reports of compromised DNS servers being used to redirect people to the wrong place.

If you can’t trust DNS, the Internet is effectively over.

Cliff gave a terrific talk on how many places are missing the boat on several fronts with Learning Management Systems. My notes:

Acutely different cultural viewpoints that are now colliding in the realm of so-called ‘learning management systems’.

What are the connections between learning management systems and the broader educational landscape? LMS got on the radar of librarians about two years ago, when it was noticed that LMS were getting installed all over the place and being managed in a “policy-free” environment, getting information into them from library-operated systems is difficult, and, as LMS vendors get into licensing content to go into their LMS, raised the spectre of institutions may be licensing the same materials multiple times.

Things become “learning objects” by context. The articles of faith here are highly granular objects with large amounts of metadata and that if we do it right these objects will be highly reused and a whole ecology of these things will evolve. Somehow this hasn’t happened – this area stubbornly resists scale-up, and people are constantly inventing new objects instead of reusing old ones.

On the other hand, LMS systems have come from nowhere in six or seven years and been installed as infrastructure. But if you look at what’s in an LMS it’s mostly not learning objects- exercises, syllabi, copies of lecture notes, threaded discussions, etc. Plus there’s no relation between these and the kinds of collections that are used for recording and passing on knowledge to support inquiry – books, articles, photographs, images, scientific data sets, etc.

One of the interesting things in the learning objects space has been the struggle to understand what metadata to apply to objects. In the broader world of digital content the whole metadata paradigm is becoming unglued.

In libraries there is elaborate metadata applied to items. The only reason we’ve been able to afford that over the past thirty years has been because of shared “copy” cataloging, where only one person catalogs the item and other libraries make use of that.

As libraries and museums begin to digitize their special collections, these are unique, not shared. It’s not uncommon to discover that an institution is spending more to describe the collection than to digitize it. Applying metadata is basically unaffordable at this point except in very special circumstances. One of the fundamental problems we’re facing right now is a flood of material that is growing much faster than our capacity to describe it.

So we’re seeing the evolution of methods to cope – cataloging collections of objects at a time, letting the computer do it (works well for text, but not for images so far).

What’s in LMS? The stock in trade is an artificial but large entity called a course. It’s not always clear where courses start or end or who participates in them. It’s starting to become clear that an LMS is really a collaborative environment with some extra special things glued on to it. That leads all sorts of projects to have to masquerade as “courses” – student projects, administrative efforts, etc. Could it be that this entity of a “course” could be a better unit of object than the highly granular learning objects?

Then there’s the content that’s generated by the interaction of people over the time period of the course.

Cliff brings up the example of faculty who don’t want the course to start again each quarter but to accrete over time, building on the previous record.

This leads to another set of policy issues. The observation was made that LMS collect all sorts of wonderful statistics that will be a boon to faculty in their teaching. The library has never sent this kind of data to faculty – instead has operated on very strict policies about privacy of the data on what users are doing, including students. Libraries now design systems to minimize the amount of data they hold, so they can’t be forced to share it by law enforcement. One of the things that’s actually happening now is that people are realizing that the web has a long memory. It has amassed enough knowledge of people’s history to provide for a wealth of embarrassment. As people move from being college students to political candidates it’s often awkward to have this material around.

What does this mean for LMS systems? Are we going to have closed systems where the records live and they’ll be destroyed after a year? Or are we going to have very open environments? In many graduate courses the expectation is that student work will get posted to the web for global inspection and review. How long will they stay there? Where does this sit in terms of student policy? As we start to regard LMS as student publishing environments, these issues will become real.

Do we really want to keep everyone’s first year calculus problem set for a hundred years?

The trend is towards more openness in scholarly communication. Those will inform the evolution of policies in this area.

Use environments are not necessarily the same as management environments. As institutions want to manage digital collections over the long haul they’re looking at repositories – keeping authoritative copies safe, accessible, etc. These are different from use environments. It’s possible that the LMS is the use environment, and things will roll out over time into the institutional repository for preservation.

Persistent references have been a problem for a long time in the networked world. We have not dealt with this very much in the LMS environment. As we want to be able to cite these works we’ll need that and it’s important to deal with this sooner rather than later.

There’s a belief that if you design systems right you can pick up all sorts of useful metadata as part of the process of creation, but if you have to go back to add it later it’s very expensive. Like in images, the camera can tell you when it was taken, what the light settings were, etc. If we were simply able to collect provenance information at creation (where it came from) we would be way ahead of the game in understanding how to obtain rights. Wouldn’t it be nice to tag items in the LMS when they’re added? eg. this is a student paper, this is a faculty contribution, this is a third party piece. Coming up with the right taxonomies instead of sending people back to tag things would be better.

I had a brief conversation this morning with Richard Katz about the state of Web Services and Service Oriented Architectures in higher education.

My going-in opinion was that lots of web services activity is happening in all sorts of places within institutions.

Richard’s take is that there’s a lack of progress on, or even discussion of, component based architectures for essential institutional services, and that the new large open source efforts like Sakai and Kuali are, to a large extent, rebuilding the same monolithic approach to business systems that we’ve seen from the commercial world.

These aren’t necessarily mutually exclusive conclusions, but I think there’s some work to be done to really ascertain what’s up with web services and component architectures in higher ed.

Any opinions out there?