Some Jisc-funded work led by Oxford University refutes common misapprehension that this is something to do with age . It ’ s residents we ’ re mostly talking about here
the enterprise is having to learn how to adopt commodity IT BYOD is a disruptive phenomenon - it ’ s costing us money and forcing us to re-think our services.
Also being called Bring Your Own Cloud, or BYO EverythingThere ’ s an app for that For advanced users - Linode - can run everything themselves - for peanuts. Could add EasyBib perhaps...
much of our efforts to provide more targeted or personalised services are in pursuit of better filters. We ’ ve pursued two, contrasting approaches: targeting our users, and allowing our users to personalise
A (tentative) distinction between targeting and personalisation A scale from targeting to personalisation, with recommender systems being at the cutting edge of targeting, and true personalisation not yet in evidence
We ’ ve concentrated recently on activity data and recommendation In the UK, Huddersfield has been a pioneer in exploiting activity data: of course, this approach extends beyond the library
This approach is growing in importance - spilling over into analytics (what we used to call business intelligence) Everyone has a stake in this. At the business level, this is pretty much widely accepted
Activity data & analytics lead us to this kind of approach to personalisation Jerome project at Lincoln University I would call this targeting, rather than personalisation They plan to add ‘ Why ’ , ‘ When ’ , ‘ How ’ , and for things they can ’ t predict, ‘ Huh? ’ :-)
targeted recommendation is patchy at best “ filter bubble ” - where algorithms imposed on you separate you from information which disagrees with your perceived viewpoint - isolating you in a cultural or informational bubble BP example Deepwater Horizon spill or investment opportunity
Cookie legislation is struggling to find the balance between privacy and usability Cookies are not the issue - but we see here the start of some real concerns about the easy abuse of systems which exploit our relaxation of privacy
it would appear that we have entered into a Faustian pact (not about privacy per se, but about control) Friedrich Kessler - the dominant party can impose a "take it or leave it" demand on the weaker party (think of those ‘ terms of service ’ we agree to with a click)
attention - sooner or later users are going to want to start ‘ banking ’ it a controlling how it is spent
VRM coined by Mike Vizard and made popular by Doc Searles of Harvard Business School
the simple, loose definition of the personal cloud now
Example - comments on blogs. If this is to become a significant part of academic practice then users (academics) need control over the comments they make on other people ’ s work. Personal Cloud right now is barely beyond the Homebrew Computer Club stage, but things are happening.
Already here Privowny - french company - browser toolbar which allow complete control over what is submitted to sites + monitors what they are tracking.
Slightly further out - the technology exists but the demand is not yet there Philip Windley
Some way out yet but it pays to consider where this is all heading We think of APIs as something we offer to users - what if we ’ re required to consume them? OpenID Connect is an example of this.
Not necessarily for the web anymore
Did you hear about Google Reader ’ s demise? Implications for RSS
Eden Dahlstrom, Educause study
analytics will help with first point Does the library service offering force them into a visitor role?
Gardner Campbell, Educause study, (my emphasis) - he also writes about helping students to build a Personal Cyberinfrastructure It starts with a domain!
It's their cloud, not yours
Paul Walkpaul@paulwalk.net@paulwalkhttp://www.paulwalk.netIt’s their cloud, not yours!
contents• residents & visitors• bring your own everything• approaches to personalisation• emerging issues• the personal cloud: a glimpse at a future• how do we respond?
Web-users are either residents or visitors• a new way of framing types of users conceived by David White (OxfordUniversity)• “We found that our students could not be usefully categorised as Digital Nativesor Digital Immigrants. i.e. This distinction does not help guide the implementationof technologies it simply provides the excuse that “some people ‘just don’t get it’which is why your new approach has failed so badly…”• “In effect the Resident has a presence online which they are constantly developingwhile the Visitor logs on, performs a specific task and then logs off.”http://tallblog.conted.ox.ac.uk/index.php/2008/07/23/not-natives-immigrants-but-visitors-residents/
library activity data• The University of Huddersfield Library• “We have collected 3.9 million librarycirculation records over 15 years.”• “If you do not use the library, you areover seven times more likely to dropout of your degree. 7.19 to be precise."http://www.jisc.ac.uk/publications/reports/2012/activity-data-delivering-benefits.aspx
Local context, expressed as activity data• analytics are fashionable• evidence-based serviceprovision is the goal• highly responsive servicedelivery is something to aimfor• predictive analytics are theholy grailhttp://www.jisc.ac.uk/publications/reports/2012/activity-data-delivering-benefits.aspx
Jerome: axes of personalisation• Where?• which campus do you study on? Which library do you want to use? how farfrom the University do you live? Are you a distance learner/researcher?• Who?• are you a student? Undergraduate or postgrad? Or a member of staff?Teaching- or research-focused [or both]? Or maybe you’re one of ourAssociate Readers or a visitor to the Library?• What?• which subject(s) do you study/teach/research, within which of theUniversity’s faculties?• http://jerome.blogs.lincoln.ac.uk/
poor characterisations of individual users• current recommender systems do not work so well, especially when the contextis broad. Within a single, focussed application, they can be made to work, but notacross the internet• data is gathered anonymously and from poorly differentiated contexts• this adds up to what Eli Pariser, in The Filter Bubble, calls:•“a bad theory of you”
cookies"The one site that installed the most wasDictionary.com. A visit to Dictionary.com resulted in234 trackers being installed on our test computer[...] the vast majority of the trackers (200 out of234) were installed by companies that the personvisiting the site probably had never heard of."http://www.npr.org/templates/story/story.php?storyId=129298003From the Wall StreetJournal’s What dothey Know AboutYou?
privacy, control & the Facebook experiment• we gain personalised services at the expense of the possibility of having anycontrol over what we’re willing to reveal• the regular “mistakes” made by Facebook have all eroded the user’s controlover their privacy in the system by making it very, very hard to understand• contract of adhesion - “a contract between parties of greatly unequalbargaining power....”• http://en.wikipedia.org/wiki/Friedrich_Kessler• but - this is not the internet, it’s just one application. Facebook will fade....• the world is experimenting with privacy
uncoordinated personalisation everywhere• the only place this can really be coordinated in a future-proof way is by theclient• either acting directly as a user• or• through some proxy which is instructed and trusted by the user• attention (data) is a valuable currency
expectations are changing - VRM• from Customer Relationship Manager (CRM) to Vendor Relationship Management(VRM)• Principles of VRM• Customers must enter relationships with vendors as independent actors• Customers must be the points of integration for their own data• Customers must have control of data they generate and gather. This meansthey must be able to share data selectively and voluntarily• Customers must be able to assert their own terms of engagement• Customers must be free to express their demands and intentions outside ofany one companys control• http://cyber.law.harvard.edu/projectvrm/Main_Page
The Personal Cloud willreplace the PersonalComputer as the centre ofusers digital lives by 2014http://www.gartner.com/it/page.jsp?id=1947315
defining the personal cloud (1)• those remote, digital services used by you, personally• essentially, you have your own infrastructure, provided by a number ofsuppliers• you choose (within quite narrow constraints) which systems you use
defining the personal cloud (2)• as for definition 1, but with the following constraints:• data: you decide what data to store and control access to it• apps: you decide which apps to use from which vendors and what data theycan access• terms: you define your own terms of service for anybody interacting with thedata or the apps on your personal cloud. You can easily move your personalcloud to a competing hosting vendor if you so desirehttp://personal-clouds.org/wiki/Main_Page
VRM: fourth parties• a new type of business on the net• third parties who work for the user,rather than the service provider• the fourth party represents the user’sinterests• in other words, an agent, or broker, ormediator• a new breed of companies providingsuch services starting to appear
the live web• a very different way of looking at the Web• the Web is fundamentally based on a request-response paradigm• the requests can be enriched by applyingcontextual information supplied by the client -under the control of the user• mixing APIs, rules and events• when this event happens, send this message tothis service
• agents which can act as the user’s persona -presenting a constrained and focussedinterface to the world• filters which learn and adapt to changingpriorities, sources & rules in a chaotic world• a secure place for them to curate data aboutthemselves and their preferences• resulting in:• systems which use contextual informationcurated by the user or by their agent, and whichdeliver accurately personalised services andrecommendationsthe future - APIs for users?
responsive (Web?) design• designing for interaction with users through the same systems interface buton different:• devices (desktop & laptop computers, tablets, smart-phones, even not-so-smart phones)• applications (‘apps’)Image taken from the Kineo website: http://www.kineo.com/mobile-learning/responsive-e-learning-for-multi-devices.html
new patterns - notifications, trusted application...• users have expectations about the sorts of features theyexpect from online services, e.g.:• notifications - presented in a standard way (not somuch through RSS as through dedicated apps)• integration and ‘trust’ relationships between systemsthey are already happily using - e.g. OAuth
changing attitude of institutional IT support• BYOE is welcome opportunity for customers, unwelcome problem for staff• What excites IT leaders in higher education most about BYOE are opportunities todiversify and expand the teaching and learning environment, while the greatestchallenges are issues that pertain to faculty and staff use of their own devices forwork-related purposes.• IT infrastructure is middle-ware between institution and users’ infrastructure• Think of IT infrastructure as BYOE "middleware" — the commodities that bridgeusers, their devices, and their consumer-level applications to the institutions data,services, systems, and enterprise-level applications. IT middleware should berobust, yet nimble.• Eden Dahlstrom, BYOD and Consumerization of IT in Higher Education Research, 2013,• http://www.educause.edu/ero/article/executive-summary-byod-and-consumerization-it-higher-education-research-2013
serve residents & visitors• build a picture of which of your users are residents, and which are visitors• be mindful that users who are visitors in the library context, may be residentselsewhere• consider how you might reach out to the residents in their wider ‘residency’• consider how library services (will) appear in each user’s personal ‘cloud’
turn the problem into an opportunity• new literacies• Those of us who work with students must guide them to build their ownpersonal cyberinfrastructures [...] And yes, we must be ready to receive theirguidance as well.• Gardner Campbell, A Personal Cyberinfrastructure, http://www.educause.edu/ero/article/personal-cyberinfrastructure• we should surely embrace this empowerment of the user?• Very few faculty or administrators are curious enough about the Internet, oreager enough to learn about the participatory culture it empowers, to even beginto imagine how to use or empower personal, interactive, networked computingin meaningful, effective ways in teaching and learning.• Gardner Campbell http://www.educause.edu/ero/article/wild-card-character-bring-your-own-panel-discussion
implications for the library• we need to:• be ready to anticipate a growing demand from our users that they controlthe relationship more than we• be ready to respond to pressure to reform how user’s activities aretracked (c.f. new ‘Cookie legislation)• consider how our services (will) fit into each user’s personal cloud• the library system:• the notion of the user ‘visiting’ the library system to find resources willbecome increasingly anachronistic• browsing as a human activity will fall away, search is king for now• over time, search will gradually become less apparent to the user too• the ratio of software to human ‘agents’ interfacing with the LMS will shiftaway from the human