XCRI Blog

If you follow XCRI you may also be aware of our work on harmonizing with other initiatives across Europe through the CEN MLO work. Here’s an important update I posted from my other blog.

Upshot: If you’ve already implemented XCRI 1.1, then you’re already 90% of the way towards complying with a future EU standard. In a future revision of XCRI we’ll need to add a couple of properties to the model for full 100% compliance, but they aren’t mandatory for providers anyway, so chances are you won’t need to change anything if you produce an XCRI feed, and aggregators will need a very minor update.

Its been very gratifying to see so many XCRI mentions cropping up - at the JISC Innovation Forum there should have been bingo cards for keynote mentions as we’d have won top prize!

However especially good to see this slide from a presentation at IWMW by Alison Wildish - formerly of Edge Hill and now at the University of Bath:

XCRI becomes just one of the things you do to make a sensible e-Prospectus, along with tagging and sorting out the database.

Well done, Alison! Hopefully we’ll see a Bath XCRI feed soon! 

Link

Seems like a long time now since the CEN workshop on learning technologies met in Athens last October, but I thought it might be interesting to publish the “Athens Declaration” made at that meeting:

There is a considerable interest in many countries in Europe in creating specifications for the exchange of information about courses and other learning and training opportunities.

There is a clear scope for greater harmonization of these efforts within a European context.

All existing national initiatives will benefit from contributing towards harmonization at a European level.

There are sufficient clear commonalities across existing national initiatives for future European standards to be developed.

Harmonization should balance the benefits of common standardization with the necessity for meeting local contextual needs and infrastructure.

Harmonization efforts should focus on small, simple models based upon existing commonalities that can be expanded upon at national or regional level, rather than all-inclusive monolithic standards.

CEN WS-LT signatories: XCRI, CDM, CDM-FR, EMIL, PAS1068, Italy & Greece

Since Athens we’ve made a lot of progress on developing the Metadata For Learning Opportunities (MLO) standard, with a draft nearly ready for submission to CEN. The MLO standard harmonizes the core models of the various european specifications, including XCRI, into a single core standard that would become a “european norm” (de jure standard) for advertising courses. The MLO working group has also actively engaged with the R3SG vendor community to make sure there is good buy-in from suppliers, and coordination with other key areas such as Europass.

These things take a lot of time, as text needs to be ratified in each EU language, but potentially this could be on the books by 2010. Along with all the other existing EU specs, XCRI is going to have to change slightly in that time in order to become MLO-compliant; however I think its fair to say that if you implement XCRI 1.1 now, you are already 90% there.

The next step is an open workshop at the EUNIS conference in Denmark on the 24th of June to open out the consultation to more suppliers, then a meeting of the CEN workshop in Lyons on the 27th of June.

For us at the XCRI project this really helps take us forward in our objectives for an open standard, and all the input the community has made to developing XCRI has given us some very important input into the development of MLO.

Look out for the reports from Denmark and Lyons later this month!

Sebastian Rahtz did us proud last year by winning the Innovation Competition at IWMW 2007. Can you make it two in a row for XCRI?

Brian Kelly writes:

This year we will be repeating the Innovation Competition. This time, rather than relying on a commercial sponsor, the Universities of Aberdeen and Bath and Edge Hill University are the sponsors. These three institutions have recognised the potential benefits of opening up their data and APIs to the community, and invite members of the community to demonstrate what can be done with their RSS and Atom feeds, their XCRI data, their microformats, their OpenSearch APIs and other data on their Web site.

More details are available on the IWMW website.

Spotted at the CETIS Metadata and Digital Repositories SIG meeting - see the bottom of slide 11!

| View | Upload your own

Yesterday I attended the CETIS Portfolios and Enterprise SIG meeting in Manchester, where XCRI had around half the afternoon for a range of presentations from people working with XCRI.

Our very own Mark Stubbs presented an overview of where we are and what’s been happening, and Alan Paull gave a presentation on his XCRI implementation model work, which is focussed on the business process end of XCRI - how organisations need to work across business units to prepare their systems and data to be capable of offering an XCRI electronic prospectus.

This was followed by presentations from two organisations implementing XCRI - West Cheshire College and the University of Bolton. We also had Selwyn Lloyd from Phosphorix presenting current work including their new CourseExchange XCRI-based product based on their open-source XCRI tools.

There were some great insights from the different presentations, in particular the Bolton developers gave a very candid and informative view into the internal processes used for getting the information together - which I think underlined the importance of Alan’s work.

In addition to the presentations, as an exercise I went through the delegate list to identify who was working on an XCRI-related activity or project, and ended up “X”ing about two-thirds of the people in the room! While on some levels things may have seemed quiet on the XCRI front, there is a LOT of development going on right now!

The presentations from the meeting will appear on the CETIS website.

As well as Sebastian winning the competition, we also had a presentation on XCRI at IWMW this year. Below you can see the slides for it.

Congratulations go to Sebastian Rahtz from Oxford for getting the top honours in the IWMW Innovation Competition - with a set of mashups based on XCRI course data.

The winning entry used Google Maps and Google Calendar as alternative entry points into course information. As has been discovered by the XCRI support team in developing the demo aggregator, geocoding services using UK postcodes are not available (except commercially), so internal lookup files are used rather than web services. Hmm, doesn’t JISC fund a load of educational GIS services..?

I really like this feature, so I thought I’d write a quick post to draw your attention to it.

One of the issues for developers of course aggregators is how to enable users to search effectively. Typically the approach is to either use an unstructured search using a keyword index, or to have an ‘advanced search form’ using criteria such as subject, mode of attendance and so on.

The approach I’ve demonstrated in the aggregator we’ve developed is to do a form of live filtering instead of iterative searching. This pre-fetches the count of results as you select various criteria - so you don’t have the frustration of having to second-guess whether a particular combination of options will return you zero results or too many to choose from. Why not give it a try?

The live filter is implemented using AJAX and Ruby, and uses the LiveFilter technology concept originally presented by unspace. The nice thing is you can add or modify the criteria to match user concerns - so we can start adding other options like part/full-time, evenings, online vs blended etc., as needed.

The “near postcode” option uses the outcode latitude and longitude to calculate proximity.

Feedback welcome either here or on the forum

The Demonstration Aggregator shows how its possible to deploy an aggregator setup that pulls in courses from several providers, and offers a user interface with basic features such as searching, browsing, bookmarking, tags and so on. It also demonstrates some value-added aspects such as geocoding the course venues and displaying them on Google Maps.

The aggregator has two major components: the aggregator daemon, and the aggregator website. As you can see in the diagram below (click for a larger picture):

The daemon is a threaded Java application launched using a scheduler (cron) that reads the target schedule to see which providers to poll, and then launches harvesters to pull in their data. The data is combined, addresses are geocoded, and the data written into a MySQL database. The implementation uses Apache XMLBeans, but otherwise has minimal dependencies and is very lightweight.

At the other end, we have a Ruby on Rails web application that operates on the same database, mainly in a read-only capacity, but also we have the interface here for adding more targets to the schedule, bookmarking and other user-specific actions. The web application’s main function is to allow providers to see how their feeds look for users, how searchable they are, how their category tags look alongside those of other providers and so on - basically, human-usability validation of the feed rather than technical validation.

Implementation-wise, the web application is deployed on Mongrel, provides search features via Ferret (the Ruby version of Lucene) and uses a lot of Ajax via Prototype and other standard libraries. As with the rest of the XCRI sites, we use OpenID for login.

The current setup is managed using Capistrano automated deployment, and is placed on a proxy setup behind Apache, which serves all static files. As demand grows we can use the load balancing proxy and mongrel_cluster to increase the number of Mongrel servers and keep things nice and snappy.

What we don’t have yet, and would be nice, is an external geocoding web service. UK postcode-geocode maps are all copyrighted (!) and so we can’t just use Yahoo or Google. Apparently UKOLN and EDINA are on the case…

You can see the aggregator web in action and browse the source code.

If you come across any issues, report them on the forum.