The coming functionality mash-up in Personal Learning Environments

  • Published on
    17-Mar-2017

  • View
    213

  • Download
    1

Transcript

  • This article was downloaded by: [Temple University Libraries]On: 12 November 2014, At: 09:27Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

    Interactive Learning EnvironmentsPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/nile20

    The coming functionality mash-up inPersonal Learning EnvironmentsCharles Severance a , Joseph Hardin a & Anthony Whyte ba University of Michigan , USAb Sakai FoundationPublished online: 31 May 2008.

    To cite this article: Charles Severance , Joseph Hardin & Anthony Whyte (2008) The comingfunctionality mash-up in Personal Learning Environments, Interactive Learning Environments, 16:1,47-62, DOI: 10.1080/10494820701772694

    To link to this article: http://dx.doi.org/10.1080/10494820701772694

    PLEASE SCROLL DOWN FOR ARTICLE

    Taylor & Francis makes every effort to ensure the accuracy of all the information (theContent) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Contentshould not be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoever orhowsoever caused arising directly or indirectly in connection with, in relation to or arisingout of the use of the Content.

    This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

    http://www.tandfonline.com/loi/nile20http://www.tandfonline.com/action/showCitFormats?doi=10.1080/10494820701772694http://dx.doi.org/10.1080/10494820701772694http://www.tandfonline.com/page/terms-and-conditionshttp://www.tandfonline.com/page/terms-and-conditions

  • The coming functionality mash-up in Personal Learning Environments

    Charles Severancea*, Joseph Hardina and Anthony Whyteb

    aUniversity of Michigan, USA; bSakai Foundation

    Current Virtual Learning Environments (VLEs) are focused very much on meeting theneeds of the institution in providing a basic, common technology platform for teachingand learning. However monolithic VLEs are too hard to customize at the individualuser level, and evolve far too slowly to meet teaching and learning of users who wanttheir teaching and learning environments to be under their personal control. This paperexplores how the concept of the Personal Learning Environment has influenceddevelopments with learning technology, within the context of emerging social software,and examines a range of developments with existing VLEs that move them in thepersonalized direction. It contrasts the issues involved in bespoke extensions to VLEs asopposed to the incorporation of existing tools (mash-ups), and suggest that the latterapproach offers the best hope to escape the bonds of a single VLE product by allowingteachers and learners to simply aggregate whatever tools and capabilities they desirefrom the Internet to use in their learning. Real progress is being made on several fronts,including the provision of interfaces to social software systems that support the buildingof applications that can be organized around a personal or group context, and in thedevelopment of specifications for learning tool interoperability.

    Keywords: personal learning environments; mash-up; virtual learning environments

    Introduction

    The current crop of Virtual Learning Environments (VLEs) is focused very much onmeeting the needs of the institution in providing a basic, common technology platform forteaching and learning. While the creators of VLEs (indeed including Sakai) are interestedin bringing the best possible technology to bear on teaching and learning, generally theirfirst priority is to meet the needs of the enterprise in terms of identity integration, auto-population of courses, ease of institutional technology support, automatic gradeprocessing, and similar issues that impact the enterprises abilities to provide a consistentand reliable system. This consistency of experience and ease of support is often reinforced,for example, when an institution is trying to convince recalcitrant faculty to usetechnology in teaching, and the faculty demands that the VLE system be very simple andthat they are given training on precisely how to use the system. Here, the value of singlesign-ons for users, auto-creation of class sites, and auto-population of class participants,simplicity and consistency across the institution in training needs, and automation of basicprocesses are very apparent.

    *Corresponding author. Email: csev@umich.edu

    Interactive Learning Environments

    Vol. 16, No. 1, April 2008, 4762

    ISSN 1049-4820 print/ISSN 1744-5191 online

    2008 Taylor & FrancisDOI: 10.1080/10494820701772694

    http://www.informaworld.com

    Dow

    nloa

    ded

    by [

    Tem

    ple

    Uni

    vers

    ity L

    ibra

    ries

    ] at

    09:

    27 1

    2 N

    ovem

    ber

    2014

  • However, faculty and students who are familiar with the Internet and Web find thishomogeneity very limiting. They are often familiar with setting up accounts,joining various types of sites and adding themselves and friends to those sites, andfiguring out new systems by trial and error. These users are more interested in support fortheir need to experiment with new ideas in their teaching and learning environmentsand then evolve/improve those ideas as they go along, often switching directionsquickly and coming up with completely new approaches in the process. The monolithicVLE system is too hard to customize at the individual user level, and evolves far tooslowly to meet the teaching and learning needs of those who actually want to use theInternet and Web creatively in their teaching and learning. These users ultimately wanttheir teaching and learning environments to be under their personal controljust like therest of the web.

    Personal Learning Environments (PLEs) come from this desire to control ones ownteaching and learning environment (PLE, 2007). PLEs start with the current andexpanding capabilities of the World Wide Web, especially those referred to often as Web2.0 capabilities, those involving individual site customization of appearance, resourcefeeds, tools and tool placement, and increasingly group or social interactions, and addorganizing mechanisms and tools focused on educational efforts to produce anenvironment that can be optimized for learning. The PLE supports both completelyindividualized, personal life-long learning efforts and trajectories, and learning withinmore structured learning contexts (for example, courses at an institution) where there issome organized or facilitated activity.

    The evolution of the PLE

    The PLE concept is relatively new as it pertains to the creation of enabling technologiesthat foster learning exchanges or networks that privilege the individual over theinstitution. The UKs Joint Information Systems Committee (JISC)-sponsored Centrefor Educational Technology & Interoperability Standards (CETIS) traces its origin to anunpublished paper by Bill Olivier and Oleg Liber produced in 2001 and entitled Lifelonglearning: the need for portable personal learning environments and supportinginteroperability standards (Olivier & Liber, 2001).

    Olivier and Libers call for the integration of institutional learning contexts with apeer-to-peer model that emphasized personal, life-long learning gained concrete expressionin Colloquia (2007), a user-centric learning management system supporting group learningand working. By 2004 the concept had gained sufficient traction to feature as a sessiontopic at a JISC Interoperability Conference held in Oxford, UK (CETIS, 2007). Thefollowing year CETIS received funding from JISC in order to define a PLE specificationand reference model along with an open-source prototype implementation. This workresulted in the release of the PLEW (server) and PLEX (desktop) applications (CETIS,2007).

    The advent of PLEX is an indicator that the transition from existing, institution-centric teaching and learning systems toward those that focus on the context of theindividual learner, and provide easy and effective ways to engage the resources of a varietyof institutions is already underway. While there are many opinions as to how institutionsand individuals will make this inevitable transition from a VLE approach to a PLEapproach, or one that combines the benefits of both, in this paper we propose that thetransition can be natural and gradual, and discuss some the mechanisms and technologiesthat can make this happen.

    48 C. Severance et al.

    Dow

    nloa

    ded

    by [

    Tem

    ple

    Uni

    vers

    ity L

    ibra

    ries

    ] at

    09:

    27 1

    2 N

    ovem

    ber

    2014

  • Fundamentally, there are many areas of overlap between PLE systems and VLEsystems. VLE systems will be well served if they understand and provide PLE-styleapproaches as features in the next-generation VLE systems, allowing those systems tointeract and interoperate with others as a matter of course. The key component of a PLE-styled VLE is putting the individual in far more control of their environment.

    Looking at current VLEs that are successful in the marketplace, one commondifferentiator is the ability to flexibly expand these systems in order to add features.Sometimes a feature is initially desired by only a small group of teachers or learnerswidely adopted VLE systems provide a market place for such local features which allowcompatible extensions to be used experimentally, and thus to see which ideas will catchhold and flourish. The broader the market penetration of a particular VLE, the morevaluable it is to write an extension for that particular VLE, and the more likely it is thatsuch experimental ideas will find a home and end up being more widely distributedavailable.

    The VLE extension points in use today are an example of support for functionalitymash-ups that are similar to the ways portals make use of portlets. However theseextension points exhibit two major shortcomings: (1) they are proprietary in nature andthus can lead to vendor lock-in and (b) development and installation of the extension istypically restricted to VLE system developers and administrators. End users cannotthemselves extend these VLE systems. This is a major shortcoming of the current crop ofVLE offerings, and a roadblock to PLE functionality.

    We can look at the patterns of these VLE extension points for technical inspirationwhen designing PLE functionality mash-ups. However there is much work to do,particularly in terms of changing the patterns of permissions to give ultimate flexibility ofcontrol to the teachers and learners over the assembly and configuration of thefunctionality they choose to adopt.

    In the remainder of this paper we propose a vision for functionality mash-ups underthe control of both teacher and learner. We will explore and describe the evolving technicalstandards that may affect this area as well as explore some emerging technologies thatbegin to deliver on the vision for functionality mash-ups.

    Defining functionality mash-ups

    We have actually been using broadly-scoped functionality mash-ups for some time now tothe point where it is no longer a surprise. Google maps (2007) provides rich softwarefunctionality in millions of web sites using the Google maps servers. Google does notneed to know and is perhaps unaware of the myriad of places/contexts that utilize theirmap display and navigation technology, for Google provides a service accessible toanyone on the Internet and all are free to integrate it into any application or web site theychoose.

    The fundamental essence and elegance of functionality mash-ups can be seen in thefollowing snippet of code required to include a Google map in a web page:

    5script typetext/javascriptsrchttp://maps.google.com/maps?fileapi&v2&keyABQIAAaxtg4

    5/script4

    These lines of code effectively communicate two important pieces of information andnot much else: (1) where to find the software and (2) which context the software is

    Interactive Learning Environments 49

    Dow

    nloa

    ded

    by [

    Tem

    ple

    Uni

    vers

    ity L

    ibra

    ries

    ] at

    09:

    27 1

    2 N

    ovem

    ber

    2014

  • supposed to use when it is operating. The value of the key in the URL query stringdifferentiates this particular placement from the millions of other Google map placementsinstantiated around the web. All of these placements/contexts can be operatingsimultaneously the Google map software keeps each of the views completely separate.Within this context we see no evidence of other Google map usage. By design, the contextacts as a filter to give us our own highly localized view of this rich and powerfulapplication.

    The context and procedure to establish the context is also a way for Google to knowwho is responsible for this placementbecause the user must sign up for the key if aproblem happens later Google can track usage back to an individual if there are problemsor issues that arise with this context or its use of Google resources.

    Google has made the use of map software in web pages even simpler than using a RichSite Summary (RSS) feed within a web page. This kind of simplicity is necessary to movefunctionality mash-ups into the mainstream and indeed ultimately place them in the handsof the end-usersnot just website developers.

    In a sense, this functionality mash-up is applying service-oriented patterns to userinterface elements. Application capability is provided as a service and that service is usedin a variety of contexts. For the remainder of this paper, we will borrow terms from theWeb Services for Remote Portlets (WSRP) specification (Thompson & Kumar, 2007) todescribe the two ends of this so-called service-oriented user interface.

    In WSRP parlance

    Producer refers both to the application software and the servers that the software runson. In the Google maps example earlier, the Google software running on the many Googleservers is the producer of the Google map application service.

    Consumer refers to the placements/places where the software is used or consumed.The consumer is the client of the application service. The consumer is whatever web sitehappens to use the Google map application service.

    This terminology is effective because it makes clear that the organization that isproviding the service is investing time, effort and resources in producing the service. Inthe case of Google maps there is no charge to use the service, but it is reasonable to expect intime that there would be some set of services that are important enough to users that theywill pay for those services, or that Google would derive some other benefit from their use.

    It is important to note that this application delivery strategy is not limited to largemega-companies like Googlewe should begin to think about developing theseapplications for educational systems we work on. The MIT Simile project (Simile, 2007)has developed just such a general-purpose application to display timeline data. Similestimeline application is placed into a web site using the same techniques as Google maps.

    Figure 1 shows the increasing usage of the Simile Timeline application over time.Interestingly, this form of software distribution offers several advantages to producers:developers can track usage, fix bugs and distribute new versions of the application in aninstant, since it is running on ones own servers.

    Evolving VLE architectures

    Because of the continuous need to innovate and experiment, current VLE systems musthave an architecture that supports a rich facility for extension. Current enterprise-scopedVLE systems such as BlackBoard Building Blocks (2007), WebCT PowerLinks (2007),

    50 C. Severance et al.

    Dow

    nloa

    ded

    by [

    Tem

    ple

    Uni

    vers

    ity L

    ibra

    ries

    ] at

    09:

    27 1

    2 N

    ovem

    ber

    2014

  • Moodle Modules (2006), Sakai TPP Tools (Sakai Foundation, 2005), Bodingtons GuanXi(Bodington Identity Provider, 2006), and Avoirs Chisimba framework (Chisimba, 2005)come with very rich extension points. Yet each of these mechanisms whether explicitly orimplicitly remains closed in nature because functionality extensions are not shared acrosssystems. Each VLE vendor/community currently uses the suite of modules as a competitivetool against the other VLE systems in the market place.

    While mainstream VLEs offer competing sets of proprietary features plus extensions,there are signs of an emerging architectural trend toward defining the VLE as a limited setof core capabilities that can integrate tools together from many sources (Figure 2).According to Dr Andrew Booth of Leeds University, one of the principal designers of theBodington VLE:

    I certainly dont see the VLE or LMS continuing in the form that they currently are, where youhave a monolithic application with lots and lots of different tools presented to teachers andstudents. You very quickly come to the limits of any tool that you use. The answer is toabandon the tool you have been using and to swap in another tool. Now I can see a situationwhere you extend this from a single tool to all of the tools of an LMS. I can see a situationwhere the VLE or LMS actually shrinks to a container into which you plug these learningtools. (Booth, 2007)

    If the VLE shrinks to become a much simpler container which brings together andorganizes the capabilities of a wide range of tools around a particular learning context,

    Figure 1. The Timeline Traffic graph.

    Interactive Learning Environments 51

    Dow

    nloa

    ded

    by [

    Tem

    ple

    Uni

    vers

    ity L

    ibra

    ries

    ] at

    09:

    27 1

    2 N

    ovem

    ber

    2014

  • then it is much more likely that this new minimal VLE can be modified to become a PLEcontainer or, rather, a PLE container can be built which is completely compatible with theVLE containers standards and protocols for tool organization.

    Dr Booths team has explored this approach to VLE architecture in his SOCKET andWAFFLE projects (Clark & Booth, 2006). SOCKET stands for Service-OrientedConsumer Kit for ELF (E-Learning Framework) Tool and WAFFLE stands for WideArea Freely-Federated Learning Environment. This work opens a potential path forwardin flexible VLE system design, designs that can naturally evolve into PLE systems.

    Functionality mash-ups in a learning context

    The problem then is to take all these technical trends and elegant architectures and realizethem within a learning context. The learning context is what makes this so much morecomplex than a simple authentication/single sign-on integration. While it is a grossoversimplification, a learning context might be thought of as a particular instance of acourse. An example learning context would be SI 539Design of Complex Web SitesFall 2007. This learning context represents a particular cohort of learners examining aparticular topic at a particular time, and perhaps in a particular locale. The key is thatthere are many possible learning contextsteaching the same course in a later semester isa distinct and different learning context from the learning context of the course taught thissemester.

    For this VLE/PLE architecture to work, tools must operate and respect the learningcontext in which they operate. The tool must act as if it is dedicated to the learning contextand take all configuration and authorization from it. There are very few general-purposetools that understand this notion of context. Tools built as extensions to VLE systems,however, are forced to operate in such a context. The key is to reach the point where onecan develop tools that can operate outside of a VLE/PLE system and yet behave as if theywere deeply plugged into the VLE system.

    Many of our most popular tools are built to be sticky, with the hope they becomedestination sitesthey show a very global view of their content to incoming users. Takedel.icio.us as a simple example: there are many ways in which this site could be used byboth students and teachers in a given learning context. If you visit del.icio.us and search on

    Figure 2. A Federated Virtual Learning Environment.

    52 C. Severance et al.

    Dow

    nloa

    ded

    by [

    Tem

    ple

    Uni

    vers

    ity L

    ibra

    ries

    ] at

    09:

    27 1

    2 N

    ovem

    ber

    2014

  • si539 you will return a set of tagged course content that is not contextualized to aparticular learning context. The bookmarks offer few clues as to time, place, learning goalor authorial intent. Yet, within a particular learning context we must have control oversuch content and the tools that present it and foster interaction around it. We may or maynot want the general public to see our bookmarks, but we want the ability to make specificchoices about the use of bookmarks in our particular learning context.

    Faced with the fact that del.icio.us is not quite right for use in a learning context, wemight decide to write book marking tools as plug-ins for whichever VLE/LMS we happento be using. However, our book marking tools are not likely to ever keep up with all of theclever features the del.icio.us programmers will add to their product, as the two effortswould be moving forward against two very different sets of stakeholders at two verydifferent development rates, with two very different resource bases. So in order to gaincontrol over our learning context, we must choose to accept less-than ideal software forthe learning context because we value control more than we value perfect functionality fit.

    A compelling example of this architectural approach is Apples iTunes University(Apple, 2006). Apple has taken the iTunes store and made it possible for universities andinstructors to create their own view of the store scoped to a particular learning context.

    The architecture for iTunes University delegates all authentication and authorizationto the VLE system. Once trust is established between a VLE and iTunes University at ahigh level, all of the low-level security details are fully under the control of the VLEsystem. The VLE connects to iTunes University and configures it within the selectedlearning context using a simple web-service exchange which occurs when the learnerlaunches iTunes University from the VLE system. There may well be features of the iTunestool that we would not want in the context of our learning environment, like selectiveexposure of students to commercial offerings, and there may well be features we would liketo see added in, like the ability to easily upload students materials and make themavailable across courses, but with this level of security and placement control, we arewilling to accept the functionality trade-offs (Figure 3).

    Use cases and technical architecture for functionality mash-up

    Each of the functionality mash-up examples cited thus far have developed their owninterface specifications, since few well-developed specifications exist in this area. For thestandards that do currently exist, there is little or no support in the current market-leadingVLE systems. Adopting standards are crucial if we hope to scale functionality mash-ups tothe point where we use them as a matter of course in our VLE/PLE environments. Thereare a number of important areas that need standardization to implement general-purposefunctionality mash-ups in a learning context (Figure 4).

    The needed standards break down into three basic areas: (1) provisioning to initiallyestablish the learning context and the agreement between the consumer and producer (thisis often done by the instructor or owner of the learning context before the first learner canuse the learning context); (2) establishing a user session for an individual joining aparticular learning context; (3) run-time services needed by the producer application andprovided by the consumer application.

    When we are considering a PLE, provisioning, session and run-time services must bevery mature and very easy to use. Because individuals are building their own learningcontexts, it would be best if they did not have to download software, install databases ordo anything else of this kind. All of the activity should be reduced to a very simple series ofuser interactions.

    Interactive Learning Environments 53

    Dow

    nloa

    ded

    by [

    Tem

    ple

    Uni

    vers

    ity L

    ibra

    ries

    ] at

    09:

    27 1

    2 N

    ovem

    ber

    2014

  • DiscoveryI think that I could use a blog in my teaching at this point in my course.What blog software can I include in this learning context? This currently works wellwithin a VLE environment where a system administrator pre-integrates a set of tools in theVLE and the user simply picks tools from a pre-set menu. But in a PLE environment thereis no assumption of a system administrator; users must be able to discover capabilities ontheir own. We can look to content mash-ups and RSS for a model for simple discovery.When you visit a site capable of producing RSS you see a special icon in your browser.Discovery happens almost as a side effect of using the web.

    Figure 4. Functionality mashups in a learning context.

    Figure 3. Apples iTunes university.

    54 C. Severance et al.

    Dow

    nloa

    ded

    by [

    Tem

    ple

    Uni

    vers

    ity L

    ibra

    ries

    ] at

    09:

    27 1

    2 N

    ovem

    ber

    2014

  • PreferencesI would like to configure a blog in my learning context. A key here willbe ease and consistency in configuration interface design if the configuration step is toodifficult or if every tool has a completely different approach to configuration, it will provean impediment to users.

    Session launchI want to see the podcasts for this course. The launch should beseamless to users, typically implemented as a web-service exchange between the consumerand producer. It includes an exchange of identity and roles and a pointer to the learningcontext the incoming user will be using. It is important that the producing applicationhandle both new users and configuration changes for existing users, reconfiguring itself toproperly reflect the learning context on each launch.

    MarkupI want all my tools to look and function in a similar manner. This is avery challenging task because of the conflicting need for web tools to innovate to maintaintheir competitive edge. Ideally the entire UI of a tool should be expressed as an abstractionallowing flexible layout and skinning. So far about the best we can do is allowing re-skinning of markup using agreement on some basic CSS patterns. The Fluid Project (2007)is developing an approach to common user interface that applies broadly across manyapplications.

    Run-TimeI want all of the files I upload to be in one place and to have the sameuser interface whenever I am uploading files from any tool. I want all my files available tobe added as attachments in any tool I am using. As we begin to distribute functionality toa wide range of servers owned by many different organizations it becomes increasinglyimportant where the data is stored in the long term. Where are my files? is just one usecase where the user wants to see functionality from different sources appear as anintegrated single image.

    In addition to these use cases there is the issue of long-term persistence of the dataproduced through user interaction with the application within the learning context. Whatif the learning context is destroyed a week after the end of the course is completed? What ifthe entire blog server that was used by thousands of courses around the world isdecommissioned, without warning or deference to any of the institutions or users whodepended on it for accredited learning? What if a person had a reference to a particularlybrilliant blog post as the centerpiece of their third year portfolio and the blog server is shutdown?

    Moving into the PLE moves us toward a life-long view of our educational activities. Ifwe are federating functionality from many places, we need to make sure that ourinformation lasts as long as we need it to last rather than leaving the data lifetime decisionup to the system administrators or a myriad of system administrators of systems that arenot even focused on education; from a student or teacher perspective, there is a loss ofcontrol since they will not know of or be able to communicate with the administrators ofthe services where their valuable data lives.

    Existing standards for functionality mash-ups in a learning context

    There are no formal functionality mash-up standards in wide use in teaching and learning.Because the LMS/VLE marketplace is currently very product-centered and market-shareoriented, VLE product producers use proprietary extension points as a mechanism to gainand hold market share. This results in very low levels of participation in and support forstandards activity to develop portable standards around learning contexts. Even whenemerging standards such as IMS Tool Interoperability 1.0 (IMS, 2005) exist, support forthe standard in existing products is either weak or non-existent. And in cases where a

    Interactive Learning Environments 55

    Dow

    nloa

    ded

    by [

    Tem

    ple

    Uni

    vers

    ity L

    ibra

    ries

    ] at

    09:

    27 1

    2 N

    ovem

    ber

    2014

  • standard is flawed or needs improvement, the only way to improve the standard is to goahead and use the existing, possibly flawed, standard and then improve it as theweaknesses are identified and, hopefully, solved.

    What follows is a non-exhaustive list of standards that form the basis for movingtowards the vision outlined earlier.

    Language independent standards

    IMS Tool Interoperability 1.0

    This standard is loosely based on a subset of WebCTs PowerLinks. IMS TI 1.0 covers allthree areas (provisioning, launch, and run-time) but its support for discovery and run-timeoperations are very minimal.

    SAML (2006)/Shibboleth/GuanXi

    SAML are part of the provisioning process for many distributed systems. Bodington is aVLE that makes significant use of SAML for the exchange of role information. Whilethese technologies are increasingly in use as a form of coarse-grained authorization, whereSAML attributes are used to indicate institutional roles, there is still work to do before wecan agree on a set of attributes that operate at the learning context level which are undercontrol of the learning context owner.

    WSRP 1.0 (Thompson & Kumar, 2007)

    WSRP is a well-designed standard that provides a partial solution to discovery andprovides a basic mechanism for markup exchange over web services. WSRP has goodpotential but has never received the necessary resource investment to insure that there arewell-built, well-deployed, richly interoperable implementations of WSRP. Vendors whouse WSRP tend to make it work well with their own products but do not invest ininteroperating across vendor projects (Yang, Wang, & Allen, 2005).

    PowerLinks

    PowerLinks is a proprietary tool interoperability mechanism developed by WebCT andnow owned by BlackBoard, Inc.PowerLinks includes limited support for discovery,strong support for launch, and strong support for the run-time needs of learning contexts.The run-time support includes web services for: Users and Courses, Mail, Calendar,Assessments, Grade Book, and File Manager. PowerLinks was the inspiration for IMSTool Interoperability 1.0. PowerLinks is proprietary and is only available to BlackBoardcustomers.

    Java standards

    JSR-168: Portlet Specification

    JSR-168 (2007) addresses Preferences, Configuration, Identity, CSS conventions andlimited AUTHZ. The JSR-168 specification only goes so far, and further agreement isneeded to truly support learning contexts. Sakai is the only mainstream VLE that supports

    56 C. Severance et al.

    Dow

    nloa

    ded

    by [

    Tem

    ple

    Uni

    vers

    ity L

    ibra

    ries

    ] at

    09:

    27 1

    2 N

    ovem

    ber

    2014

  • JSR-168and this was just added in the Sakai 2.4 release. JSR-168 has the potential to beuseful for functionality mash-ups in a learning context if we could get together and agreeon common extensions points.

    JSR-170: Java Content Repository

    JSR-170 (2007) standard covers the storage of files, metadata, and properties. Thisstandard is only recently gaining interest in learning contexts. Sakai hopes to move to JSR-170 as an option for implementing its storage application programming interface (API) inits upcoming 2.5 release.

    Emerging standards

    WSRP 2.0 and JSR-286 (Portlet API)

    The updated versions of these specifications (JSR-286, 2007) contain a number ofimprovements: (1) improved alignment between the Java JSR-286 (2007) and webservices (WSRP 2.0) standards, (2) support for cross-portlet communication andevents, and (3) general improvement based on real-world experience and newdevelopments in web applications such as Ajax. These improvements should expand thenumber of applications that can use these standards without resorting to vendor-specificextensions.

    Open Knowledge Initiative (OKI): Open Service Interface Definitions (OSIDs)

    The OKI Project (OKI, 2007) is developing version 3.0 of the OKI APIs. OKI Version 3.0contains a number of significant improvements over OKI version 2.0 including: (1)improved support for out-of-band agreements, (2) improved Java binding for the OSIDs,and (3) improved processes for public involvement in the OSID process.

    IMS Learning Design (2007)

    This standard allows the creation of dynamic learning sequences with automated controlof the release and sequencing of learning activities. This standard is not widely adopted inmainstream VLE systemsthe CopperCore (2007) learning design engine provides animplementation of the IMS Learning Design SpecificationCopperCore does not includethe tool implementations. In a sense IMS Learning Design and CopperCore aretechnologies waiting for something like funtionality mash-up as a way to build learningcontexts to reach their full potential.

    Learning Activity Management (LAMS) tool contract

    The LAMS project (LAMS, 2007) has developed a REST-style interface that describes thecontract between the LAMS learning activity engine and a tool that is to be controlled aspart of LAMS. The LAMS Tool Contract covers the entire scope of functionality mash-ups described earlier except for Discovery. The LAMS tool contract also has importantfeatures that are not represented elsewhere. The tool contract includes provisions toimport and export configurations at a sufficient level of detail to move a placement fromone tool-producing server to another. The tool contract also provides a facility to export

    Interactive Learning Environments 57

    Dow

    nloa

    ded

    by [

    Tem

    ple

    Uni

    vers

    ity L

    ibra

    ries

    ] at

    09:

    27 1

    2 N

    ovem

    ber

    2014

  • the output of the tool, including all of the user activity, as a document. The exporteddocument could be stored in a repository for long-term archiving or included as part of astudents personal portfolio.

    IMS Learning Tool Interoperability (LTI 2.0)

    The effort to broaden and improve IMS Tool Interoperability 1.0 is just getting underway.The primary focus of the activity (so far) is to improve significantly the support in the areaof run-time services. The hope is that the scope of IMS LTI 2.0 will be expanded to becloser to the scope of WebCT PowerLinks. Once IMS LTI 2.0 is developed and available itwill enable a very wide range of applications to be developed using this patternIMS LTI2.0 producer applications can be written in any programming language, hosted on manydistributed servers, and will be ideal candidates for functionality mash-ups in a wide rangeof VLE/PLE systems.

    Emerging de facto solutions

    While developing and improving standards for functionality mash-ups in a learning contextare important efforts, general-purpose functionality mash-ups are likely to be promoted bymarket-leading web development companies such as Google, Yahoo and Facebook. Muchlike current VLE producers use their proprietary extensions to gain and hold market sharecompanies like Google and Facebook are trying to convince developers to write applicationextensions for their particular form of application functionality mash-up (Figure 5).

    The recently announced Facebook Platform (Facebook, 2007) is a significant steptowards functionality beyond Google Applications. Facebook applications begin to

    Figure 5. Facebook Platform.

    58 C. Severance et al.

    Dow

    nloa

    ded

    by [

    Tem

    ple

    Uni

    vers

    ity L

    ibra

    ries

    ] at

    09:

    27 1

    2 N

    ovem

    ber

    2014

  • explore the notion of context. Each Facebook application operates in the context of thegiven Facebook entry. The Facebook Platform includes a number of APIs:

    . Authentication

    . RSS support

    . Facebook query language (fql)

    . Friends (social networking)

    . Profile

    . Users

    . Events

    . Groups (a form of authorization)

    . Photos

    The initial set of APIs that Facebook Platform provides to developers is obviouslyinfluenced by the social networking aspects that form the core of Facebook software.Applications can (when permitted) access and modify Facebook information using theseAPIs.

    Using Facebook to produce a learning context might be as simple as building aFacebook entry for a class and then associating a number of Facebook applications withthe context. While initially Facebook will seem a poor substitute for a high-functionalityVLE system, all of the architectural elements are already in place to potentially replaceVLE and other collaborative systems. The only question is which direction Facebook willevolve their APIs, and what tools will be built to them.

    Open issues/challenges

    Even if we were to have a complete set of well-supported specifications in our PLE/VLEsystems there are still several challenging issues functionality mash-ups must solve.

    Two-way trust and security

    Security is very complex in learning systems. The relationship between a grade book and atesting system is a good example. In most traditional learning contexts, students are notallowed to use the grade book to enter their own grade. However, after they log in andtake a test, and the testing system determines what the students score is on the test, thescore is placed into the grade book. If this is done with a VLE using an external testingsystem, the score is determined by an external application and needs to be placed in thegrade book, perhaps via web services. Many students would find it useful to have a setyour grade service which bypassed the whole test-taking step and used the same webservice to set the grade. In this example, there is a need for careful trust management andvery fine-grained two-way authorization. IMS Tool Interoperability 1.0 has the beginningsof an approach to two-way trust that will need to have a significant expansion of scope inLTI 2.0 to meet requirements like this.

    Rogue applications

    Enterprise learning systems have access to information such as course rosters that areprotected by privacy laws. What if a mashed up application had access to the courseroster, either directly via an API call or indirectly by logging each user as they access the

    Interactive Learning Environments 59

    Dow

    nloa

    ded

    by [

    Tem

    ple

    Uni

    vers

    ity L

    ibra

    ries

    ] at

    09:

    27 1

    2 N

    ovem

    ber

    2014

  • learning context? Would this violate privacy? Perhaps rogue elements will produceseemingly attractive services for the sole purpose of data mining peoples e-mail addressesfor spam purposes.

    The problem becomes even more complex as we create intermediate federated contextsthat form when a group of users join a shared learning context. We must control therelease of information from our personal context into the group context, and perhaps wewill need to be involved with controlling the release of our personal information toapplications that are part of the group context.

    A trivial, but not likely to be satisfactory, solution is the classic sandbox approach toask the user all the time. It will be very intrusive if we are continuously getting SMStext messages such as The blog for SI 539Design of Complex Web Sites would liketo retrieve your e-mail addressreply yes/no as to whether you want to permit thisaccess.

    Now that Facebook has taken the step of giving Facebook applications access to apersonal context, these issues will begin to be explored. Facebook Platform supportsextensive sandboxing and the end user has complete control over which elements of theircontext can be accessed/modified by the Facebook application. We can hope that Facebookwill carefully explore rogue application issues and develop technical and policy solutionsthat work well and have reasonably simple end-user configuration characteristics.

    Conclusion

    Functionality mash-ups capabilities in a personal learning context have been a broad goalof teaching and learning technologists for some time now. They promise the ability toescape the bonds of a single VLE product and allow teachers and learners to simplyaggregate whatever tools and capabilities they desire from the Internet to use in theirlearning.

    While functionality mash-up has been discussed at some length, there are manydifficult challenges that have kept us from making much progress toward the functionalitymash-up ideal. We have noted a number of obstacles to innovation including a VLEmarket that uses proprietary extension points to gain and hold market share, a lack ofsuitable standards for building applications which enable functionality mash-ups, andsignificant embedded security and privacy issues.

    The good news is that there is real progress being made on several fronts. For example,Facebook now provides an API that supports the building of applications that can beorganized around a personal or group context. While this is not yet to the level of a learningcontext, it does explore important new areas in the integration of applications originatingfrom different sources yet sharing common and potentially sensitive centrally held data.

    There is also important progress on the standards front. The IMS Learning ToolInteroperability 2.0 effort plans to produce a broadly scoped set of interoperable servicesthat will allow standards-based functionality mash-up around a learning context. Newversions of WSRP (WSRP 2.0) and the Java Portlet API (JSR-286) will provide basetechnology to exchange increasingly powerful application functionality across Java Portletcontainers.

    Acknowledgements

    The authors would like to thank the reviewers including Sean Mehan of University of the Highlandsand Islands for their helpful reviews and commentary. The authors would also like to thankMacKenzie Smith, Aggie Booth, and Hans Masing for the use of their images in this paper.

    60 C. Severance et al.

    Dow

    nloa

    ded

    by [

    Tem

    ple

    Uni

    vers

    ity L

    ibra

    ries

    ] at

    09:

    27 1

    2 N

    ovem

    ber

    2014

  • Notes on contributors

    Charles Severance is a Clinical Assistant Professor at the University of Michigan School ofInformation. He was formerly the Chief Architect of the Sakai Project and the Executive Director ofthe Sakai Foundation. His current work is functionality mash up in learning environments. He canbe contacted at www.dr-chuck.com or csev@umich.edu.

    Joseph Hardin is a Clinical Assistant Professor at the University of Michigan School of Information.He was the Principle Investigator of the Sakai Project and is on the board of directors of the SakaiFoundation. His current work is on evolving approaches to Open Courseware. He can be contactedat hardin@umich.edu.

    Anthony Whyte is the Community Liaison and Security Lead for the Sakai Foundation. He wasinvolved in the IMS Tool Interoperability standards and is currently involved in the IMS LearningTool Interoperability 2.0 standards effort. He can be contacted at arwhyte@sakaifoundation.org.

    References

    Apple. (2006). AppleEducationiTunesU. Retrieved October 20, 2007, from http://www.apple.com/education/itunesu/.

    Blackboard Building Blocks. (2007). Blackboard Building Blocks. Retrieved October 20, 2007, from

    http://www.blackboard.com/extend/b2/.Bodington Identity Provider. (2006). Bodingtons GuanXi. Retrieved October 20, 2007, from http://

    www.guanxi.uhi.ac.uk/index.php/Bodington_Identity_Provider.

    Booth, A. (2007). Bodington Project. Sakai Video Report. Retrieved October 20, 2007, from http://www.dr-chuck.com/media.php?id80.

    Centre for Educational Technology & Interoperability Standards (CETIS). (2007). CETIS.Retrieved October 20, 2007, from http://www.cetis.ac.uk/members/pedagogy/articles/PLE/.

    Chisimba. (2005). Introduction to Chisiba. Retrieved October 20, 2007, from http://avoir.uwc.ac.za/.Clark, B., & Booth, A. (2006). SOCKET and WAFFLE Bus for Beginners. Retrieved October 20,

    2007, from http://www.elearning.ac.uk/features/socket.

    Colloquia. (2007). Retrieved October 20, 2007, from http://www.colloquia.net/.CopperCore. (2007). IMS Learning Design Engine. Retrieved October 20, 2007, from http://

    coppercore.sourceforge.net.

    Facebook. (2007). Facebook Platform. Retrieved October 20, 2007, from http://developers.facebook.com/.

    Fluid Project. (2007). What is Fluid? Retrieved October 20, 2007, from http://www.fluidproject.org.

    Google. (2007). Google Maps API. Retrieved October 20, 2007, from http://www.google.com/apis/maps/.

    IMS. (2005). Tools Interoperability Guidelines. Retrieved October 20, 2007, from http://www.imsglobal.org/ti/index.html and http://www.imsglobal.org/ti/tiv1p0pd/imsti_ guidev1p0pd.html.

    IMS Learning Design. (2007). IMS Learning Design. Retrieved October 20, 2007, from http://www.imsglobal.org/learningdesign/index.html.

    JSR 168. (2007). Portlet Specification. Retrieved October 20, 2007, from http://jcp.org/en/jsr/

    detail?id168.JSR 170. (2007). Content Repository for Java technology API. Retrieved October 20, 2007, from

    http://jcp.org/en/jsr/detail?id170.JSR 286. (2007). Portlet Specification 2.0. Retrieved October 20, 2007, from http://jcp.org/en/jsr/

    detail?id286.LAMS. (2007). The LAMS 2.0 Tool Contract. Retrieved October 20, 2007, from http://

    wiki.lamsfoundation.org/display/lams/ToolContract.Moodle Modules and Plugins. (2006). Retrieved October 20, 2007, from http://moodle.org/mod/

    data/view.php?id6009.Olivier, B., & Liber, O. (2001). Lifelong learning: The need for portable Personal Learning

    Environments and supporting interoperability standards. Retrieved October 20, 2007, from http://wiki.cetis.ac.uk/uploads/6/67/Olivierandliber2001.doc.

    Interactive Learning Environments 61

    Dow

    nloa

    ded

    by [

    Tem

    ple

    Uni

    vers

    ity L

    ibra

    ries

    ] at

    09:

    27 1

    2 N

    ovem

    ber

    2014

  • Open Knowledge Initiative (OKI). (2007). OKI Project. Retrieved October 20, 2007, from http://

    www.okiproject.org/.Personal Learning Environment (PLE). (2007). Personal Learning Environment. Retrieved October

    20, 2007, from http://en.wikipedia.org/wiki/Personal_Learning_Environment.

    Sakai Foundation. (2005). Sakai TPP Tools. Retrieved October 20, 2007, from http://www.sakaiproject.org/.

    Security Assertion Markup Language (SAML). (2006). SAML. Retrieved October 20, 2007, fromhttp://en.wikipedia.org/wiki/SAML.

    Simile. (2007). Semantic interoperability of Metadata and information in unLike Environments.Retrieved October 20, 2007, from http://simile.mit.edu/.

    Thompson, R., & Kumar, R. (2007). OASIS Web Services for Remote Portlets (WSRP).

    Retrieved October 20, 2007, from http://www.oasis-open.org/committees/tc_home.php?wg_abbrevwsrp.

    WebCT PowerLinks. (2007). Retrieved October 20, 2007, from http://www.webct.com/powerlinks.

    Yang, X., Wang, X., & Allen, R. (2005). How mature are portal standards? Retrieved October 20,2007, from http://epubs.cclrc.ac.uk/bitstream/1082/paper06A.pdf.

    62 C. Severance et al.

    Dow

    nloa

    ded

    by [

    Tem

    ple

    Uni

    vers

    ity L

    ibra

    ries

    ] at

    09:

    27 1

    2 N

    ovem

    ber

    2014

Recommended

View more >