Alex Little

 I attended the recent Unfold CoP meeting in Glasgow which was very useful – all the presentations etc are available via the Unfold project website. The most useful parts for me were….

• We (well actually Patrick & Diane) ran a workshop on our Sled player, which seemed to go well, though we had a few problems with the responsiveness of our Sled demo site (more about that later!!)
• I presented some of the work I’ve been doing with the integration of external services into LD, eg how we did the QTI integration and the problems we had in using the same methodology to integrate ePortfolio.
• James Dalziel’s presentation on the updates they’ve been working on for LAMS, and the new release that is due out soon, with the ability to be able to write your own tools to plug into the LAMS interface. Would like to get chance to explore and have a look at what they’ve done in more detail
• Scott Wilson’s presentation which gave much more detail on the tools integration method used by LAMS – I think it’s a really good start at being able to connect up different service providers, but it’s not quite a ”plug and play” type of methodology just yet !
• Mark Barrett-Baxendale’s presentation on their real life use of learning design on actual students at Liverpool Hope University – where they are making use of CopperCore engine and the Sled player

One thing that did come up were some performance issues that we’ve not been able to resolve yet and we’re not yet sure of the actual cause, if it’s the Sled player, CopperCore engine or something else. When we were running the workshop we found that having 15 or more people connecting to our demo system at the same time slowed the system down to a crawl – making it almost unusable. I thought that this may be due to the fact that it’s just running on a desktop machine and we’re only using the default database, rather than upgrading to MSSQL or PostgreSQL and that upgrading the database or putting it on a bigger machine would solve the problem. However , when Mark presented their work at Liverpool Hope Uni, they are also having performance problems if they have a whole class of students accessing the system at the same time. There, they already have upgraded the database to PostgreSQL and are running it on a ‘proper’ server.

So, I’m not sure where the problem lies, whether its the Sled player, or CopperCore, but to help try and track down the problem I’m going to find a load testing tool and run it against an installation of Sled and the CopperCore player on one of our servers and see what happens. Finding what the problem is will probably be the easy part though – going to be more difficult to figure out what to do about it!

After I’ve spent a bit of time looking at this performance stuff, I’ll then get back to what I’m meant to be doing for the JISC toolkit and demonstrator projects as I still have quite a few bits to finish off for these, and need todecide what the priorities are for the last couple of weeks.

Have been away quite a lot for the past few weeks – hence why I’ve not been keeping this up to date. I attended the latest Unfold CoP meeting in Glasgow a week or so ago which seemed to go well, I’ve put more about this in my LD blog, so have a look there!
I’ve just been back in the office for a day now, so just catching up on things. On 10th Oct we released our new IET website; the actual design hasn’t changed (well, only a little) but we have:

• Created the site content from scratch – there were 100′s (or 1000′s) of old, unmaintained, out of date pages on the old site, so in starting afresh we have removed all of these. They are still available on backup if we do need to restore any sections/pages for any reason, but we hope we’ve got the main content areas covered
• Moved to a clustered servers – so the website should be much more robust. Though we have had some intermittent problems with one or other of the cluster servers becoming unresponsive overnight. I think we’ve got this fixed now – might have been due to one of the server timeout settings not being set, so if a long running page was hit (or caused an infinite loop) the server didn’t automatically terminate the request after a certain amount of time.
• Changed the publishing system for the site – now all IET staff can edit the site via their web browser (a bit like a wiki) and everynoe has access to edit any of the pages. Previously people needed to use Dreamweaver or other web editing software to actually go and edit files – and also have access to the files. This meant that updates to the site often didn’t get done because of the hassle invovled in finding the right person who has access etc etc. We’re not sure yet that this will definitely solve the problem – but it does mean there is no excuse if someone spots a typo or incorrect email address etc for them not to just go and update the page themselves. It also means we can track more easily pages which haven’t been updated in a certain length of time – I think we should be aiming for ensuring pages get updated at least every 6 months – or at least reviewed even if nothing actually needs to be changed.

The next thing I’ve got to work on is getting the LD JISC projects finished off and what should be the priorities to get done, and also some work finishing off bits on the new IET site.

 I’m just now beginning to install the ePET tool, so that I can explore what web services functionality it has and how I may be able to link this up to the Sled player. There seems to be a few bits and pieces to get installed (mysql, zope etc) in order to get ePET running, but I’ll log here any problems that I come across.

 I’ve now got the player running on Linux. It’s been a bit of a pain to figure this one out as we’ve not got a linux machine here that I can just use and test things out on, so I’ve been relying on ther people to test any changes I make and let me know about any errors that occur. In version 2 most of the player appeared to work ok on linux, but a few pages in some UoLs causes errors. This turned out to be pages with ‘global content’ (ie they ask the user to enter details or upload a file and stores the data in LD properties) and the problem was that the stylesheet which did the transformation referenced a function which was in another XSL file. Even though the include statement appeared fine (and worked fine on windows) it caused the player to fall over on linux. Once I moved the function to just within the main stylesheet, all seemed to work fine. I’m sure that it’s solvable in a better way than I’ve done it, but this fix will do for now

I’ve now got the Sled player running on Linux, more details are in the Learning Design section

I’ve now finally got round to putting up the first version of my Sudoku solver program and online demo.

I have released the latest version of the Sled player to the SourceForge website. For installation instructions and download links, please visit the ‘technical’ section on the Sled website.

Now that we have the Sled website up and running, I’ll no longer be maintaining the Sled installation instructions on this website, however I’ll continue here with this blog, so you can keep up to date (ish!) with developments.

 I have now released the latest version of the Sled player on the Sourceforge website. You can find the installation instructions and download links in the ‘technical’ section on the Sled website.

In some of my spare time at home, I’ve developed a sudoku solver program, which is nearly ready for me to make available. It uses jsp & servlet technology, and should run on tomcat or jboss. I’ve still got a few bits to finish off – the web interface is a bit basic at the moment, but so far the engine that actually solves the sudoku seems solid – well I’ve not managed to get it to fall over yet!! The actual engine isn’t particularly sophisticated, just keeps track of which numbers can go where and updates this everytime a new box is set to a particular number, it can then find which numbers can only go in a particular column, row, square,  or group of squares in order to solve the sudoku.
The web interface can:

• give you hints as to where the next numbers go
• show you the solution to the whole grid
• tell you whether or not the engine is able to solve the sudoku in it’s current state (if either it’s not solvable at all, or if not enough numbers are present to determine the solution)
• allow you to work on and solve 9×9 and 16×16 square sudokus – in fact the engine will cope with any size.

Anyway, hope to get this posted up in the next few days  

Today I’m having a go at installing the OSPI ePortfolio system (http://www.osportfolio.org), as we need to have an ePortfolio system set up to link to with the Learning Design Player. We need an ePortfolio system which has either a java or webserivces API to it,and it appears that OSPI does, however the documentation on this area seems quite ‘slim’, so I’m not going to know for sure that it is suitable for us until I get it installed and have had a play. The other problem I’m going to have with OSPI is that I don’t know very much about ePortfolio, OSPI appears to have a lot of features and I’ll need to try and get my head around it all. However I think in terms of this projet we do not have to cover every possibile function, we just need to show with a couple of functions that it is possible to link up and demonstrate how it could be done – leaving the hard work of full integration either to someone else (!) or the next project!