Previous tool making

Before I get into making new tools I want to go through what I’ve done so far. There’s a line between each of these areas of work below that has shaped my current thinking and where I am today, and each of these previous projects has informed my thining as well as given me valuable lessons to apply to the next project.

It’s now almost 10 years since my last foray into tool making. Back then it was The Adaptive Digital Publishing Engine or as we came to know it – TADPOLE.

TADPOLE came out of our work trying to transform content for delivery on mobile devices. We sketched out our work in the form of an Ascilite paper and at the conference we were able to run through a demo of the Proof of Concept we had developed.

TADPOLE was built around the concept of single-source publishing, something that has become a cornerstone of many of the new tools available for publishing — like PressBooks, Manifold and the Coko Foundation. Single-source publishing is where you develop, create and edit one content source, but from that publish to different media and channels. This would allow a university to create one source for the content of a course but allow it to publish to the web, for print and for offline download. Our main innovation though was the development of the Adaptive Media Element which allowed us to adapt different media to suit the final output formats we were working towards – PDF for print, HTML for web and ePub for mobile.

Translating a Video AME for print, eBook and the Web.

Each of these formats had their own unique abilities and ways of handling media. For example if we had a video – well you can’t have a video in print, but you can have a still from the video and a link to where it is available online (say on YouTube). That same video can simply be embedded in a webpage and ePub actually allows you to embed the video as part of the file making it available offline, so we could choose to package the video in a suitable compressed state inside the ePub file itself.

During our time we managed to build our a proof of concept, which is what we scoped for the project, but then when we went to ask for resourcing to continue development we were rebuffed. For the next year I tried a few different avenues to get funding but with no luck. I thought perhaps putting it in for Shuttleworth funding might be helpful, but again no such luck. The video I did for that application was perhaps the best and most succinct description of TADPOLE that I created.

After the failure to get TADPOLE up it was not the end of making tools for learning though.

The next big project was the development of the Online Learning Model at CSU. The aim with the model was to develop a conceptual framework of best practice for online learning and from that build out ways and methods to embed these practices into courses across the whole institution. This was a group effort, with myself and Dr Lindy Croft-Piggins putting in the bulk of the work to pull together the Model itself and the associated resources to go with it.

My input into the model was on trying to aid the adoption and embedding of good practice into existing courses. The way we did this was to develop “strategies”, single pieces or examples of the broad elements of the model that could easily be incorporated into an existing course without the requirement for a fundamental redesign. In fact, the many of the strategies could be adopted into an existing course with minimal effort to enhance student engagement. The “strategies” themselves were based on a pattern language, a way of trying to simplify complex ideas into building blocks to construct learning experiences with. The aim here was to empower both learning designers and academics themselves and provide a stable of “quick wins” that could be deployed to aid the adoption of the model.

This work had a great effect and became a useful tool especially in the way that it provided a “lingua franca” between learning designers and academics. It was powerful to see over two years the conversations that we were able to have as a team and with swathes of the university change because we now had a way of speaking.

The site that we housed all of that information in, the Online Learning Exchange, was built by me using a Static Site Generator (SSG) – Jekyll in this case – and was a platform for development that I could see as quite powerful. Rather than requiring a whole backend, server and database, the site was actually run and stored on GitHub, with the files being finally uploaded to a simple server we had access to. Jekyll allowed me to build a complex website with the same functions as as a dynamic site but just using basic data formats, YAML and Markdown, and some fancy templates. I was able to get the site up and running in a couple of days and was able to quickly modify and update it, add new features without having to change the data or worry about running a server or database. The reason for this, and a key feature of SSGs is that “content” is treated differently when it is authored vs when it is published. What YAML provides is a way of storing metadata that allows code to be written to manipulate the display of that content. In the same way that metadata allowed us to transform a video into an Adaptive Media Element, metadata allowed me to build pages and collections from a single source of data — each of these pages is based on metadata contained in each strategy. and the code does all of the heavy lifting!

I then used this knowledge for the next project, CSU Learning Materials. This was an effort to provide a consistent style for learning materials produced by the various teams at CSU. The biggest impact that it had though was to spawn a reskin of the institutions Blackboard theme. In hindsight this was not a simple project and I loathed it, mainly because it revealed just how poorly Blackboard (and I assume by association most LMSs) are put together from a web perspective. Sure they function, but having to sift through the HTML structures it is not done well or efficiently. I swear I can get the same functionality with Blackboard with a about 50% less code if you just used HTML and CSS conventions correctly. What that work aimed to do though was start to pull together learning design with web design, to try and consolidate them and match them up so that good design flowed from one to the other.

I’ve carried on some of this work at the University of Adelaide. I built the style system used in the MathTrackX MOOCs, and that has since flowed into a reusable set of styles we can now deploy across multiple platforms via our Online Design System. I also played with trying to automate tasks and the how to potentially leverage SSGs to automate tasks and speed up development.

So what’s next? In a broad sense it’s about coming back to the idea of Tools for Learning – specifically for the design and development of courses. At the moment my team is having to scale up the production of courses significantly and over a long period of time. We are shifting from a very bespoke development unit into a larger production studio where for the foreseeable future we will need to be sustainably producing new courses. While we have been able to make the best of what is available, what has worked in the past is actually becoming a huge impediment as we move forward. It feels very much like we are working without tools and it’s time to change that.

Advertisement