EILE

Yesterday we presented a five minute overview if our project to around fifteen groups of people. It was exhausting and for every bit we improved through repetition, it was levelled by tiredness! It’s also really difficult to fit all of the things we’ve been doing in to a such a small amount of time, so I really hope some of the people we spoke to are reading this post. I did say to a few groups that I could be contacted through the blog, which it doesn’t appear I can be, so here are a couple of ways to get in touch:

Email: s.coppin@kent.ac.uk
Twitter: steve_coppin

Report post

James and I arrived in California yesterday for the IMS Learning Impact conference, where our nomination has been accepted for a Learning Impact Award in the research category. The conference is giving us a great opportunity to promote our work and the LTI standard, and also to introduce our new logo! We will be manning a stand and presenting a 5 minute overview to attendees and a judging panel. We are trying a new presentation so when we’re back we will produce a new video/screencast so we can share it with everyone.

We will let you know how we did on Wednesday! Wish us luck!

Report post

Friday April 29th marked not only the Royal wedding, but also the official end of the EILE project. We have thoroughly enjoyed working with JISC, IMS and Canterbury Christ Church University on this project and would like to thank them all for their support and cooperation. The EILE project are very proud of their achievements and look forward to presenting them at the upcoming Learning Impact Awards in a couple of weeks. Although the project has officially ended we are not finished with LTI yet! We are currently working with the JISC funded Cartoon Archive Rapid Digitisation (CARD) project on LTI enabling the British Cartoon Archive and will be seeking out other projects to continue developing with the standard. The EILE project team have enjoyed working with LTI and being a part of the impact we believe it will have in the production of elearning tools. We will continue to update this blogs with our progress and thoughts beyond this project and hope you’ll keep tuning in.

My sincere thanks to everyone who has been involved in EILE.

Report post

When our team initially discussed the benefits of LTI with out elearning team, it became clear that we had to be very careful to ensure the integrity of our VLE. Security concerns are well mitigated by the standard and reliability becomes primarily the responsibility of the service provider, although the consumer plugin must also be maintained to ensure these risks are kept to a minimum. The integrity of the VLE, including keeping it learning centric, is quite rightly still the responsibility of the institution. We decided the concerns may be alleviated by introducing a mechanism whereby integration candidates can be properly assessed and out of the process we have created our LTI Evaluation Criteria.

We have tried to keep the questions non-institution specific in hope that these could be useful elsewhere. If you’re in the business of assessing integration candidates and would like to use, or better still contribute to, these criteria please do get in touch with the project!

Report post

The University of Kent currently maintains a Moodle instance to deliver course based learning and a web portal, powered by uPortal. The role of a web portal is to provide a single point of access to disparate institutional systems. The situation we find ourselves in now is that as the learning environment becomes richer and more embedded in the every day life of a student, they often spend more time on our open-source Moodle instance than our portal counterpart. For this reason we are often asked to provide links to other systems, however there is a concern, internally and externally, that Moodle remains focused on delivering learning objects and services. The Full LTI standard both lowers the barrier to entry of integration and puts control in the owner’s hand, which could be a technical team or an individual academic, and this could add to the community’s concern.

The EILE project has noticed a change in the perception of using Moodle as a portal with more institutions opting to unite the two offerings. This could be largely accounted to the introduction of new themes, including the popular Aardvark theme, which offer the ability to use different tabs to jump in to different areas or pages of the system. It is also possible to use these tabs to link to external sources and the layer of abstraction it creates (from the usual course list form of browsing) gives a portal feel to the VLE.

As discussed in previous posts, we believe the Full LTI standard will provide VLE owners with much wider opportunities for tight integrations with learning tools. In lowering the barrier to entry and significantly shortening the development cycle, we hope to see many more tools becoming available. The single sign on ability and context awareness of the standard also allows intelligent and seamless linking to external sources, which fulfils a similar purpose to that of a portlet in a portal system. These features of Full LTI, in addition to the popularity of more layered themes, could make it much easier for Moodle to act as a portal.

Full LTI also makes it easier (or at least more tempting) to shift the focus of a Moodle instance away from e-learning. For example, if the VLE is providing information (or links to information) on conference events and campus parking fines, it could be argued that this has little to do with e-learning and therefore affects the perception of the tool. It is the desire of our e-learning team to maintain this perception to preserve student focus in the learning environment, as would be expected in a classroom or lecture theatre.

It is possible that the creators of the standard have already created a solution to this problem. Context-level enablement allows a tool to be integrated at course and category level, rather than site wide. Once integrated, the tool would only be available under this context and would not affect any other course outside of it. Together with a theme such as Aardvark and the category based theme ability of Moodle, context enabled LTI providers would be a very easy way to create a portal area of the VLE without affecting the learning focus of the core service.

There could, however, be limitations to this approach. It is important to note that the LTI standard is designed for the integration of learning tools that understand the learning vocabulary. It will directly reference teachers, students, courses, etc. and it is up to the tool provider to interpret these. If the tool provider doesn’t understand (or need to understand) these terms they could begin to make risky assumptions that are difficult to maintain. The EILE project also believe that one of the reasons that LTI will be successful is this limitation in scope and would not like to see the standard expand outside of the learning context. Too many standards have come and gone having tried to provide a generic way to integrate all tools.

In the current economic climate an institution is challenged to produce more value from a single offering. When a popular tool becomes capable of providing another service it should be considered, but not without considering the impact on the core service. Full LTI, layered and context specific themes certainly enable Moodle to act as a portal and arguably do enough to protect the integrity of the VLE. The notion of integrating and aggregating an institution’s offerings also reflects a change in approach in the physical delivering of learning. There is an increasing acknowledgement at all levels of education that learning is more than lessons and resources – perhaps it is time for e-learning to reflect this?

Report post

When constructing the initial bid for the EILE project, we were hoping to be able to prove the real cost savings achieved by becoming Full LTI compliant. It became clear that the deliverables of this project alone were perhaps more expensive than the alternative of creating custom integrations, though as discussed they would be more difficult to maintain and with a higher risk attached. It is also apparent that the cost of integration varies greatly with the type of tool and the level of integration that is desired. With this in mind, the nature of this post has changed and will instead be addressing the future cost savings for learning providers and the VLE owner. The EILE project believes these savings will be real and carry a great impact for the future of the learning environment.

Learning tool provider

Shortened development process
As discussed in our learning impact video and screen cast the Full LTI standard significantly reduces the development process for learning providers. The consumer side of the specification will be provided by the VLE creator or supporting community, which considerably shortens the development effort. It also negates the need for the author of a learning provider to understand the inner workings of the VLE, which are often complex and ever changing, especially those that are community supported. The EILE project has also discovered that large parts of the integrations we have built are common across learning providers and have started creating a common library to further speed up the development of new providers.

Maintenance costs
The Full LTI standard has been carefully crafted to support the existing integration needs of the leading VLEs and to be relevant to future demand. As a result, the LTI specification is less likely to change than a learning provider or a VLE. If your integrations are using a central library, such as the library the EILE project is developing, a rare change to the specification could result in a change to a single point in the library, rather than every provider. This is much more unlikely to be the case with custom integrations.

Another maintenance cost we have discovered is the separation of concern between the learning provider and the VLE owner. As the entire service is being consumed by the VLE, it is very easy for the learning provider to implement and deploy changes, with little or no conversation needed between the two parties, especially if no change in service is expected.

One integration for multiple VLEs
Possibly the most obvious benefit for a learning provider is that they only need to provide and maintain a single integration to be able to integrate with a compliant VLE, of which all of the current market leaders should be. Combined with the removed responsibility for understanding the inner workings of the VLE, as mentioned above, this will result in significant cost savings for learning providers.

VLE owner

Cost of own integrations
Often a VLE owner, such as ourselves, will be required to write their own integrations, especially for internal developments and learning providers requiring other types of integration such as web services. Not only does the owner have to absorb the initial cost of development, they are also responsible for the ongoing maintenance, which is usually as expensive, if not more.

Open source tight integrations
Currently we find that the only tools that can provide a tight integration with the learning environment are the proprietary license based service providers. The LTI standard significantly shortens the development cycle and lowers the barrier of entry for learning providers, inevitably stimulating the open source community. This will leave VLE owners with a wider choice of free and licensed tools and the increased competition in this market will lead to price reductions.

Moving VLE
It is clear with the introduction of Moodle 2.0 that institutions are concerned about the effort and associated costs of moving VLE versions, let alone VLE. The Open University have recently embarked on a large program of work and coordination to move Moodle versions. This effort could have been greatly reduced if the OU were using LTI providers, because the tool would be independent of version/VLE and only needs to be compliant to the specification. If a tool is to provide the functionality to re-associate a resource, it could also make the migration of content easier too.

Cost of failure
Without the sandbox environment that LTI provides, every integration has the potential to bring significant risk to the security, reliability and integrity of the learning environment. It only takes a single integration to accidentally remove records from a database table without a suitable check for the VLE to be rendered useless. The VLE has become an integral part of an institution’s program of learning and with the increased demand in a rich environment, coupled with a reduction in the funding available to provide it, accidents can and will happen. By maintaining a single point of integration the risks are considerably reduced and the VLE owner is free to seek out the best compliant tools available to them with confidence.

Conclusion

The EILE project firmly believes that the LTI standard is going to drastically reduce the effort and associated costs of providing an integrated learning environment. Both the learning providers and consumers are set to benefit from LTI and this mutually beneficial relationship will serve to support the longevity of the standard.

Exactly how much financial benefit an integration could gain becoming LTI compliant is still difficult to estimate, but the EILE project team hope to build more tools and provide case studies to support our predictions.

Report post

The EILE project team recently presented our findings at MoodleMoot UK 2011 and received a mixed reception. Our workshop presentation was very well received and sparked some interesting debate, which suggests the community are ready for the introduction of Full LTI, however, when I posed a direct question to Martin Dougiamas (creator of Moodle) over the growing importance of an interoperable VLE we received a different response.

My question was in direct reference to the core team’s focus on building a new survey tool for the new Moodle 2 release. I asked Martin whether the core developers’ time would be better spent working on interoperability standards so that we could instead adopt a market leading survey tool, such as Bristol Online or EvaSys. This would also leave the core developers to focus on the aspects Moodle is trusted with most, like course management and assessment. Martin’s response was that the survey tool they are building would have some good features. I would be surprised – although I could be wrong – if the tool they build within the Moodle infrastructure could rival some of the independent market leaders but, even if it could, one could argue that their time could have been better spent on the vehicle. Moodle 2 does include a Basic LTI plugin, which perhaps makes this response even more surprising.

Luckily for us, Moodle is a community supported open source tool, so even if the core developers do not want to actively support interoperability standards, the community can still make it happen. The EILE project are particularly interested in the companies offering hosted solutions, as we believe the introduction of learning interoperability standards has the potential to really improve their offering. In the summer of 2009, the University of Kent had a choice of hosting externally but relinquishing some control, or host internally and acquiring full control while accepting the associated risks. We chose the latter because the luxury of being able to deploy what you want, when you want, was decided to be worth the risk.

The current hosted solution model of providing an approved list of blocks and modules from which you can select protects the VLE well but rather limits the ability of the owner and academic. It is unlikely that a plug-in developed by the institution or selected from the wider community would not be integrated as it would require a full code review to maintain the reliability and security of the instance. The ongoing maintenance agreement could also prove a headache for the hosted service provider. This leaves the VLE owner with a rather vanilla version of Moodle, from which it is difficult to gain an advantage in an increasingly competitive environment.

The standard VLE is based upon an old way of teaching through resources and assessment. Academics need to be allowed to find new ways of teaching, to be liberated – Paraphrased comment from a JISC workshop

A hosted service provider opting for an interoperable VLE could not only liberate the VLE owners to find their own plug-ins, but even individual academics. The Full LTI standard offers a very low risk solution for this type of offering with its sandbox environment and loose coupling of tools. If this option had been available in summer 2009, perhaps Kent would have opted for a hosted solution (and I wouldn’t have a job!).

I think the clear benefits for hosted service providers will ensure the future of Full LTI in Moodle, even if Moodle themselves are not behind the change.

Report post

As mentioned in a previous post, the EILE project’s Learning Impact Award nomination has been accepted under the research category. We will be invited to set up a booth at the upcoming IMS Quarterly in Long Beach, California to present our project to a panel of judges and attendees. In preparation for the event, we have been asked to produce a short video outlining the impact our project has on the e-learning environment.

I would like to thank Matthew Wilson for his assistance in making the video and Nick Hiley, Head of the British Cartoon Archive, for accepting our invite to feature.

Report post

In the original project plan, EILE had discussed providing tutorials on how to become LTI compliant. We have learned through the project that a single tutorial would be difficult to produce, especially on a changing specification. It has become apparent, however, that our presentation/screen cast and example providers are providing enough support for the creation of compliant providers.

We have already seen the creation of two great providers for PMWiki and Drupal. These were created in a matter of days using only the WordPress example we have provided. Our community is also starting to notice reusable code and have extracted it in to a central library, which we will make available shortly. This should also provide valuable support to potential providers.

Although the providers we have made available are menu-linking, we would like to see if the same pattern emerges for resource-linking providers when we release our more comprehensive streaming media server solution. Instead of providing tutorials, we will invest our time in the support of our growing community and the provision of new LTI compliant tools.

Report post

We have recently created a screen cast based on the presentation we gave at the IMS quarterly in Utrecht, Netherlands. In the screen cast James and I provide a brief overview of the problems with integration, our introduction to LTI and the EILE project, before showing a demo of our LTI Moodle plug-in in and providers in action.

The video can be found at this URL: http://tinyurl.com/eilescreencast

Report post