Notes from Universities and Colleges in the Giant Global Graph

''The session outline and background info: Universities and Colleges in the Giant Global Graph.

Intro, Show and Tell, General Discussion
Adam Cooper gave a brief [[media:CETIS 2009 GGG Intro.ppt|introductory presentation (PPT)]] (also [[media:CETIS 2009 GGG Intro.odp|ODP]]) followed by a round of introductions and show-and-tell:
 * Nadeem Shabir talked about Talis Aspire as an example of Linked Data in an educational context and brought in an OER dimension
 * Tony Hirst talked about his work at the OU and Cambridge U in particular where he is thinking in graphs and searching for consistent identifiers in university systems.
 * Joss Winn showed a little of write-to-reply the semantic web capabilities that can be easily added into Wordpress via "triplify"

There was quite a lot of discussion among about 25% of the participants, sometimes quite esoteric, and as we moved about the room with introductions it became clear there was quite a split in the participants and those with senior institutional roles were forming the opinion that we were not talking about anything of immediate relevance to them. This seemed to be slightly moderated by the end of the session but it is clear that we need to move from talking about "potential" in a hand-waving way to having some more believable cases for investment before people holding budgets in H/FEI support units will be convinced. One of the aims of the session was to better understand how we can move along this path.

Discussion
Who are the stakeholders?
 * "user" (as opposed to "education system") stakeholders are public and business
 * funders, govt
 * legal people, data protection controllers
 * developers
 * prospective students, current students
 * managers, staff

Current and prospective students are not all passive consumers. Have agency and culture of open sharing. Given linked data, would they shortcut us and provide services to peers.

What data might universities expose?
 * courses/opions (currency of data)
 * expertise
 * resources/facilities
 * events
 * student/staff output (repositories)
 * alumni data

How would current students interact with "real" out-there data: sciences, social sciences?

Who are the data owners? Problems if this isn't clear? How will they view exposure?

Ideas
Proposition #1: Book-a-room:
 * service= availability of uni/college rooms, maps, capacities, facilities, price
 * data owner=estates?
 * public/business are clients

proposition #2: Marketing:
 * services= find-an-expert/consultant, find-a-course
 * data from professional bodies, staff lists, the REF, bloggers,
 * clients= prospective students, business

Blockers:
 * data protection
 * fear
 * sensitivity/ethics re privacy
 * "the website is enough"

Enablers:
 * student demand
 * public service ethic
 * perception of business opportunity, profile
 * handling FOI requests

What could JISC/CETIS do?
 * Taking a staff/student perspective:
 * workshop to explore data
 * play with tools
 * discuss value and brainstorm possibilities
 * Capacity building - already going on, enough?
 * Info -briefing paper?

Group 2


CO2 Counting is all about transparency on carbon footprint. Possibly a requirement or an aspect of student choice. Students may ask "what will my footprint be if I attend?"

Group 3


"There's more than one bottom" was a reference to the view that "bottom-up" doesn't necessarily mean developer-lead.

Thanks to Adrian Stevenson of UKOLN for contributing notes that have been edited and adpated

Discussion
It is possible to sell the idea of linked data to VC, snr mgt etc. You don’t need killer apps, you just need to prove that you have something and that its not vapour ware.

We noted the problem of sustainability of linked data services like DBPedia that are largely provided by enthusiasm and good will.

It could be that we need to provide some ‘enlightened self interest’ to make a case providing linked data, in a similar way to delicious. Organisations already engaged in making data available (esp as open data) are obvious candidates for lobbying for adoption of linked data principles.

What argument has been made to get BBC to adopt linked data approach? We could do with hearing these?

The serendipity argument is overstated – the coolest use of your data will be something you didn’t think of is clearly not true (acc to PW).

FOI could have been useful as a lever if the cost of servicing FOI requests had been punishing. Could a case still be made based in this? i.e. could the cost of providing FOI be offset against the case of providing open linked data?

CSV files have been put up by The Guardian, at a minimal cost. Some think the cost of true linked data would be much more.

Is there a case that providing LD makes the quality of the data better? Reconciling data sources exposes weaknesses and errors in intergrity. Could this be used as an argument? If the data is open then there is more incentive and need to make sure it is sound. This is also another reason for resistance to make data open of course. it was later noted that Chrysler used linked data principles internally and a number of other organisations seem to be heading there from a REST point of view

Associate the LD case with e.g the uptake of OER. A case has been made that people have enrolled on courses as a result of using OER materials. Can we tack linked data approach onto this – the linkability brings in new students.

A sell maybe that your LD source is on the LD graph and connected to other LD sources on the graph such as DBPedia. Being meshed in with these other data sources is an another exposure route that helps sell the institution. Your institution needs to be at the party and not left behind. Counter-argument Rupert Murdoch, Google news. These are questioning whether one does need to be part of the bigger picture – Take e.g. Ryanair, Easyjet, Jet2 don’t play ball with travel engines such as travelsupermarket.

What Could be Done?
JISC is already involved at the developer level with devCSI and JISC Call 03/09 from the Information Environment Team included mention of linked data. This is one aspect of a bottom up approach, in contrast to top-down semantic web (e.g. ontology-first) but developers are not the only group of actors at "the bottom".

JISC (and its support centres) could lead by example. e.g. open up their own databases. Could it be argued for JISC making its own projects data available as linked data? There would payoffs for in that benefits analysis and programme synthesis for example would be far more efficient. Maybe HEFCE and the Resarch Councils could/should do the same.

Some fear of being left behind in future scenarios suggests we sould develop some cases around institutional reputation, student choice, anticipating the channels propspective students (undergrad and postgrad, "traditional" and work-based) will use for form opinions and what the best ways are to feed information into these channels.