The Auricle

Want some non trivial example of Web services for e-learning? There is certainly plenty of trivial examples of stock quotes or weather at airports but, otherwise, welcome to the dark zone. We're involved in a JISC project which requires us to study how to deliver and consume Web services so this article attempts to share some (low) illumination on this challenging but interesting area. Before moving on, let's consider what's Web services and how are they different from normal Web pages/sites?

Now I could expose you with intimidating technical definitions and really frighten us all by providing some examples of marketing hype. But no. Instead I offer the following; mainly because it's helped me get a grasp of the concepts.

The Web as we have know it has been built on a publishing model, i.e. 'get page from server'. Arguably, what we also need are systems which support distributed computing applications, i.e. different functional elements making up a client application may be remote from each other but can still communicate because they follow some standard protocols which allow such communication.

So a Web service receives a request message from a client application. If it recognizes the message the service sends back a response message. The system will only work because both the message and response is sent in a standard XML format. There are a number ways of packaging these XML messages but the important thing is that both the requester and responder know what's required. Simple Object Access Protocol (SOAP) is a common, but by no means the only, way of packaging these messages. Also, it's important to have a sense of humour when you consider the S in SOAP because simple doesn't mean trivial:)

It's important to realize that, all being well, a load of XML formatted data comes back from the Web service and it's up to the client application receiving this to utilize and present it. It's also important not feel intimidated by your lack of instant understanding of the XML data storm that comes back at you. XML is really meant for machines, not humans, to process and that once a client application script has been set up it will transform all that nasty XML into something humans can make use of.

But have Web services moved beyond the visionary to the concrete?

The answer seems to be an ever so tentative yes.

I found a link to an interesting case study contained in the JISC briefing document for JISC 07/04 Circular: Regional e-Learning Pilot Projects around Distributed e-Learning. The example cited by JISC is the LeAP Project Case Study on implementing Web services in an education environment from the Tasmanian Department of Education.

It's important to grasp that the LeAP case study is an architectural and systems comparison study and not an end user or institutional case study. Nevertheless, parts of it make for an interesting read. For example, on page 43 I found:

“As with the IMS Framework, the ELF document is intended as a starting point for discussion, with refinement of the detail anticipated over time.”

The ELF document is the JISC et al E-Learning Framework which has moved beyond being just 'a starting point for discussion' since it is the conceptual engine for the eTools projects now being undertaken under the distributed e-learning strand of the JISC E-Learning Programme.

And , of course there is the Isoph Blue LMS I mentioned in my recent article E-Learning Frameworks and Tools: Is it too late? - The Director's Cut along with mentions of the IBM Lotus Learning Management System and other major players, like Oracle iLearning, Saba, and Docent, with interests in emphasizing the Web services aspects of their products.

As part of my own orientation I thought it would be useful to try and track down some examples of non-trivial Web services which could be of direct or indirect relevance to e-learning.

Let's start with the indirect relevances.

As I mentioned in my ALT-C 2000 paper, a good example of an implementation which 'consumes' the output of a number of services is Erik Benson's All Consuming site. All Consuming uses Web services from Weblogs.com, Amazon.com, Technorati.com, and Alexa.com. All Consuming publishes aggregated information by literally 'consuming' RSS information about books from weblogs supplementing this with information about the book (acquired via Amazon's public Web services). To this mix All Consuming adds news information about each book via the Google public Web services. The screenshot below illustrates this aggregation.

All Consuming adds value to existing services in ways never evisaged by the original developers of the individual services. Are Google and Amazon unhappy about this? Not at all, the net effect is to steer more 'eyeballs' towards their 'products'.

Let's pose a related scenario (or use-case if you want:)

A member of Faculty publishes an online reading list of books for a module and wants to supplement this with additional information about the readings with some from Amazon.com. Why Amazon? Because the company is still, arguably, the leader in providing Web services and publishing information about them for consumption by third parties. Put more simply, they have a big database of information to which they allow access by independently developed computer/network applications so makes for a great test bed. Google Web services would also be a great alternative. However, if you are unhappy with Mammon providing the information, substitute Amazon for any similar non-commercial Web service (if they exist).

How could we do this? Amazon provides a number of techniques for getting at that information and it's not necessary to use SOAP/WSDL because the company's Web services options also includes XML over HTTP, e.g.

“http://xml.amazon.com/onca/xml3?t=[Your Associates ID]&dev-t=[Your developer token]&BrowseNodeSearch=[BrowseNode]&mode=books&sort=[sort type]&offer=All&type=lite&page=[page number]&f=xml”

Worked examples of the above technique can be found in Dan O'Brien's article Amazon Web Services: A Brief Introduction Using PHP. Further demonstrations and downloads of the XML approach can be also found at Jaap van Ganswijk's collection of free scripts. We happen to be PHP oriented here but similar functionality is available in most mainstream computer languages.

Other interesting sites built on services include Philipp Lenssen's FindForward site which does all things Google.

So from the above examples you can perhaps see that significant parts of the commercial world appear to believe that Web services are important aspects of their activity. Although they are commercially oriented, that doesn't mean that accessing the services needs to cost any more than your development effort. Creating some potentially useful client/Web applications from these 'free' commercial services should be possible.

But what about the non-commercial e-learning world?

It's been much more difficult to track down non-trivial examples of Web services development outside of the big commercial players. Nevertheless, here are a few promising initiatives.

Ariadne's Knowledge Pools looks promising (see Web Services to Achieve Interoperability Between eLearning Applications). However, there is a paucity of examples of how to connect to the Knowledge Pool. The existing documentation appears to make the assumption that a local system is being installed and that we are all Java enterprise oriented. Nevertheless, personal contact proved fruitful and the Aridane people have been very helpful and put me in touch with the person charged with the development of a PHP client for the Ariadne Web services. So it may be possible for us to try this out in the near future.

We also have the JISC Connect services but their SOAP interface information isn't quite ready, but I'm informed that this should be available in the near future.

And, of course, there is the joint JISC, DEST (Australia), Industry Canada, E-Learning Framework (ELF), and the service-oriented projects which will arise from this.

So from where I'm sitting and with the exception of Amazon and Google it's all a bit pending and works-in-progress at the moment, but yet All Consuming does illustrate the potential. What we need, therefore, is an FE/HE equivalent of All Consuming which at a stroke would sell the concept of Web services to an otherwise bemused community.

We also need, urgently, demonstrations of services which are not necessarily dependent on a portal/portlet model. That's not to say that I don't feel portals are important. They are. However, there's a lot more people around who can demonstrate services which aren't dependent on having access to a portal than there is with access to, and knowledge of, portal technologies and their associated implementations, e.g.WSRP.

So JISC Connect, JORUM, TOIA, Ariadne et al take a leaf from Amazon and Google's book and let's make it as simple as possible to send requests and get responses from your services and in that way viral marketing will do the rest!

Now if someone could just bring out say a demonstrator online discussion' service, or an assignment service or that reading list service … 🙂

MoodleMoot is the name given by the international Moodle community to the Moodle conference programme. MoodleMoot Ireland took place on the 10th September 2004. I couldn't make it to Ireland so over a period of days I've watched the movie instead. There are many interesting presentations and discussions in the full event but in this article, with one exception, I am going to concentrate on the contribution to the conference by Moodle's creator, Martin Dougiamas. It's worth visiting the MoodleMoot site and if prompted login as guest. The edited streaming video of MoodleMoot is nearly 4 hours long so I hope you find my highlighted timecodes useful.

Martin Dougiamas had a virtual presence at the conference through a video link to Australia. He had two slots during the day. In the first he gave a fairly detailed review of background, philosophy and technical infrastructure of Moodle. In the second slot be was one member of a panel answering questions. As it turned out most of the questions were directed at Martin! I have found this particularly helpful in giving me a more complete picture of Moodle. In the rest of this article I have distilled out some of the main points that were made by Martin. I have done this following a format first started by Derek Morrison in his UKeU: The Movie article, by giving you the read time markers from the video and a very brief statement of what you can see and hear at these points.

AT
2:01:40 Martin's overview of the development process and why Moodle is not like a traditional software company.
2:03:20 The hopes and dreams of Moodle. An ever growing number of users circa 1832 registered users in 92 countries. These range from the very large to very small, e.g. one student and one teacher
2:04:45 Moodle modular structure makes it possible to plugin different bits and pieces. Easy customisation.
2:09:00 Question raised about Moodle. I have oftened wondered. See the answer here!
2:09:45 What can we expect in the future from Moodle?
2:10:05 At this point Martin looks at the importance of the Moodle forums. This is one of the key support features designed into Moodle.
2:13:20 How do new features get into Moodle and what about the tracking of bugs? See 3:41:35 for an anecdotal acccount of how the bugtracker worked for one person.
2:15:35 There is a business arm to Moodle, called moodle.com. Hear how this fits in?
2:18:05 Moodle moves forward at speed with no little thanks to the external developers ie people giving back to the wider community
2:20:20 Who are the Moodle partners?
2:23:20 With so much development taking place with Moodle how is it all managed and controlled. At this point Martin explains the importance of CVS and how it helps protect…
2:26:25 against possible software patents being taken out.
2:27:45 Sourceforge is the place where you get the Moodle parts. An understanding of this is required if you are the system administrator of Moodle.
2:35:45 Martin's final part of his presentation in his first slot was to cover the very important aspect of the testing of Moodle.
Questions were then raised by the conference audience.
2:37:30 The presentation demonstrated how much energy that Martin has for Moodle. This prompted the question of What would happen if Martin gets bored (or even worse) with Moodle? There are some reassuring words in the answer.
The final event of the conference was the panel session where Martin featured a lot. Fast forward the video here.
3:24:05 Are WebCT and Blackboard quaking in their boots because of the emergence of Moodle? Martin gives a measured and reasonable response.
3:25:45 What features and further developments do you see in Moodle particularly to support the social/constructivist pedagogy? An interesting answer to this one.
3:29:15 Martin was asked to elaborate on how Moodle is dealing with issues of accessibility.
3:33:23 A conference member made the arresting statement 'Wiki is dead, long live markdown'. I have to admit Martin's answer did not mean too much to me until I carried out a Google search to find out what is markdown!
3:34:23 Explain how, in its present form Moodle supports the social/constructivist pedagogy. So, does Moodle cut the mustard. Listen to the answer.
A number of questions were raised about support issues, not directly to Martin and you can listen to the last part of the conference to hear the responses. However I will conclude with the anecdote that I referred to earlier in this article, from Dr Barry McMullin of Dublin City University.
3:41:35 You can find the anecdote at this time point.

Again, we think Moodle is worth watching and we think its takeup and influence is likey to grow a view reinforced Dr Howard Miller, the University of Glasgow delegate who announced in a 'news flash” at timecode 2:58:50

“The University of Glasgow adopts Moodle as its centrally supported VLE”.

Now Glasgow is a large, major league, research intensive institution which has been biding its time in the VLE stakes, so we should pay attention.

Enough said!

At this year's Association for Learning Technology conference (ALT-C 2004), held at the University of Exeter, there was the usual supporting vendor stands in the conference exhibition space, so in between sessions I took the opportunity to test VLE/LMS vendors' perspectives on a couple of pretty key issues. First, how possible and easy did they make it for licensees to extend/enrich their core product? Second, how well did their business models appear to be adapting to the emerging era of Web services. It's only my opinion but if you want to know more read on! To put this article into context Auricle readers may like to read the extended version of my ALT-C 2004 paper E-Learning Frameworks and Tools: Is it too late? - the Director's Cut.

So let's consider the concept of the product licensee or 3rd parties extending the vendor's core product.

Sure, some allow development and integration of supplementary 'blocks' or functional components and others will give access to a SDK and some APIs. Of course, first you've got to have bought into the eterprise class level of their products, i.e. you are well and truely locked in to their particular platforms. What was interesting here was the divergent views taken by the mainstream vendors, with one viewing their product's potential extensibility via 3rd party blocks as a business plus whereas a rival vendor viewed the blocks model as a sign of the lack of 'completeness' of their rival's product. An interesting divergence!

Let's consider this 'completeness' business case a little more because it will lead us nicely into a consideration of the potential impact of Web services upon current e-learning vendor's business models. To me, the 'completeness' model seems totally at odds with apparently what's going to be required of the next generation of e-learning tools and services. No system can be 'complete' for long and particuarly not one which swims in the dynamic world of teaching and learning. Nevertheless, the 'completeness' argument must be ever so seductive to the hard-pressed HEI executive member or Director of Information Services who just wants the solution.

'Completeness' of course could also mean that “we (the vendor) decide, after a process of consultation with you (the licensee), what we (the vendor) can afford to implement in this iteration of the product and no we don't want to make it easy for any old 3rd party to extend the product. But, of course, if you're also a major vendor of a complementary e-learning product in your own right, i.e. a trusted business partner we might consider it.”

So the classic business model at work here is to offer a 'package' of functionality and long-term support delivered via, and only via, the vendor's platform. Want only to use a one aspect of the system's functionality? Want to extend the functionality? Sure you can do that but you're still going to have to buy in to the full package and platform and have that long long long term relationship.

And who can blame the vendors? This is the tried and trusted business model of “get them on your platform and they will find it extremely difficult to get off again”.

What's really interesting is how like lambs to the slaughter we line up to be 'locked in' assisted in no small measure by an institution's systems people who view high levels of integration as their raison d'être and who view a VLE as just another system to be integrated.

We also seem to be slipping into a mindset which allows us to declare resounding e-learning success at a conference because we have 'n' hundreds/thousands of VLE vendor 'xyz' courses available via our institution's web site. Vendor xyz's marketing people must jump with joy at how their sales and marketing strategy is proving so cost effective due to these high profile, and perhaps unsuspecting, 'evangelists'.

But, as some of the speakers at the ALT-C conference pointed out, look carefully at these 'courses' and with some notable exceptions it's not unusual to find only a few announcement and perhaps some content uploaded. Hello! … here's some bad news … that's administration not e-learning. But here's the good news, particularly for those concerned with the bottom line … Simple content organisation/delivery and announcements does not require a recurring and expensive enterprise class licensing arrangement with a commercial vendor.

But what about that services model and how are the vendor's responding?

My impression is that the vendors I talked to were either playing their cards very close to their chest or they hadn't really considered the potential implications or opportunities of a Web services model. Read my paper and view the resources and you'll see that some of the other players in the e-learning marketplace, like IBM, are taking a very proactive approach to the design of services-based architectures.

Web services have got the potential to destablize current business models. Business models based on the complete 'package' or nothing are not going to be responsive to potential licensees who want to aggregate one or two aspects of a vendor's product into their own institution's portal, client or any indeed other service consuming container. As my paper suggests there are already some major commercial players working on, or offering, service-based e-learning solutions; so the current market penetration by VLE vendors is, arguably, going to be less stable than it has been.

One of the putative benefits of the services model is that the niche developer could create a 'best-of-breed' service, e.g. a discussion engine, which could replace a less satisfactory offering within a learning environment. Now let's assume that in five years time the e-learning services model has penetrated the marketplace. How are the big cats in the jungle going to attempt to make money? By offering 'packages' of services of course:) And what about that niche developer hoping to penetrate the marketplace with their 'best-of-breed' service? … Why they will have been bought out by the big cats … and so the world goes on.

Except!

There is that dratted 'stone in the shoe' called open source/open software. Suddenly individual developers, development teams, or communities of practice are producing service based solutions which they make available to their community at no or low cost. But that's surely killing the marketplace? So what to do?

Why there's always custom extensions or application profiles. Yes it may be a service and yes it will work via your portal but if you want feature 'xyz' then sorry you require … 🙂

The presentation I gave described the processes and outcomes of an e-learning initiative involving collaboration between the University of Bath's Centre for the Development of New Technologies in Learning (CDNTL), and the university's Department of Social and Policy Sciences. The collaboration tested one online solution to a very practical problem; the teaching of large cohorts (~200) of students, which leads to reduced opportunities for interaction and feedback, especially with regards to group work. GroupLog was designed to support a particular pedagogy rather than any one subject, i.e. a collaborative, activity based approach to learning, through the use of structured group work. Designed with this pedagogy in mind, the value of GroupLog can be considered to be cross disciplinary.

The tool uses content management and weblog principles to permit groups of students to obtain feedback from their lecturer that would otherwise be difficult with such large cohorts. Furthermore, it presented an opportunity for peer-learning by allowing student groups to review each other's work which they would otherwise not have experienced nor could have been provided using the institution's VLE.

The tool has now been used for over a year with two, separate cohorts of students. Evaluation shows that the tool has been favourably received by the students and lecturer alike and further development of the tool is now underway.

The demonstration can be downloaded as either a powerpoint file (215 KB) or hmtl.

Well, seeing as Derek's doing a bit of pre-publicity for his presentation for ALT-C, University of Exeter, here's a brief plug for the demonstration I'm giving. CDNTL has recently received funding from the JISC to develop a prototype web application for collaborative group work, called GroupLog as part of the eTools programme. I'm going to be demonstrating the present prototype, GroupLog, a Tool for Group Work with Large Cohorts (click link for the abstract), at ALT-C on Wednesday 15th September, at 11 am in the Library Seminar Room.

I'll be making the slides available on Auricle on Wednesday after the demonstration.

Busy week this week so there's limited time for writing long articles. However, I'm presenting a paper at next week's ALT-C conference at the University of Exeter which I'll make available on Auricle on Wednesday. The title of my paper is … E-Learning and Frameworks and Tools: Is it too late? - the Director's Cut. My paper extends the arguments I put forward in an Auricle article of the same name earlier in the year. So if you are going to ALT-C 2004, it would be good to see you in the University of Exeter's Newman Lecture Theatre (A) on Wednesday 15 September at 1330.

In my three previous Moodle Meanderings articles we've sort of 'poked a stick' at Moodle and verified that in fact it's alive and looks well, so this time we decided to see how easy it would be to extend one aspect of its functionality. Let's be clear … Moodle is still a work-in-progress. Nevertheless, I feel that it's an increasingly respectable work in progress which has got a relatively low barrier to entry and use. It may be all some institutions require.

I particularly like the alternative ways of viewing Moodle discussions. I also like the explicit learning activities - learning resources model which underpins course or topic organization.

What's appears a bit weak is the way that only some learning activities (implemented as Moodle modules) can be allocated to groups, e.g. Moodle wikis can be group allocated but 'lessons' don't appear to be. And, as best we can tell, resources are not group allocatable. Some of this will probably change in the near future with for example the primary developer of the Moodle 'workshop' indicating he has plan to 'groupize' the activity.

So what's been our particular contribution to the Moodle project?

The Moodle architecture supports the concept of a 'block' which is a functional element located along the sides of main Moodle frame. Blocks can optionally displayed on the home page of a Moodle site or for each 'course' page. Examples include the Moodle calendar and the 'Upcoming Events blocks. Some blocks can be configured to display or manage data relevant to a local context.

A contributor from the Moodle community (Paul De Jong) had recently added a block which allowed news syndicated from another site to be displayed within Moodle. Whilst this block was useful we felt it could be enhanced so that a course could be supported not by just one syndicated feed but by as many feeds as required by the course author.

We already had a basic model in Auricle which can display multiple syndicated feeds (see the drop down menus at the top of this site). We wanted to improve on this so that authors specify however many syndicated resource sites the wished. The result is that each Moodle course with an active 'Syndicated Feeds' block can now look something like below:

So it's now possible for a course to offer a multiplicity of distributed resources all accessed via RSS syndication. Shown below is the configuration interface or our enhanced syndication block which is presented as a series of multi-line web form fields.

N.B Describing each syndicated resource takes two lines. The first is the site name; the second the RSS address.

Although Moodle automates a large part of the block creation process the task is still not one for those lacking confidence in PHP coding, although undoubtedly the process would be much quicker the next time.

Whilst enhancing the work of others can be a quick way to achieve results sometimes there's some gotachas! in doing so. For example, we found that our new block worked fine on campus but wouldn't display external syndicated feeds. We found that in common with many open source and some commercial RSS developments no account had been taken of what happens when, for security reasons, some instituitons expect data to be accessed via a proxy. The result was we also had to adjust the basic script to allow for this contingency.

The other major gotacha! and one we can't resolve is that, currently, Moodle only allows administrators (not tutors, not students) to configure blocks. Our block would best be configured by a tutor not an administrator. This problem arises because 'fine tuning' of Moodle privileges is necessary; again this issue has been recognized by the Moodle developers.

Anyway, if anyone else is dabbling in the Moodle waters and is intersted in our syndication block we've posted on the Using Moodle site. I suggest you view the message chain as 'Display replies flat with newest first'. Enjoy:)