Lessons Learnt

During the ‘Developing Linked Data Approaches to OERs’ project we came across a few points that we thought it would be useful to comment on here:

  • Design. We go for substance first over style! Some people are good at looking past very plain designs to the functionality of the tool itself, but for others, the visual appearance of the tool is important to enable them to see the functions at all.
  • Attitudes to openness and licensing. Whilst the variety of Creative Commons licenses (and their respective logos) can still be confusing to some, we found positive attitudes to OERs in our design groups. Far from being reluctant to share their resources, academic staff spoke of a culture of sharing and handing on educational resources (particularly powerpoints and similar materials) in departments where collaborative teaching is the norm. Sharing high quality OERs outside departments and institutions, through services like iTunes U, was seen as a way of enhancing academic standing and profile – but attribution is still important for this to be effective.
  • HTML5. HTML5 is touted as having two roles: it enhances and extends previous HTML standards for creating web pages, and it provides a platform for creating web applications. We found that HTML5, when used for creating web pages, was more that adequate; but when used as an application platform, (so called ‘web apps’), HTML5 fell far short of its so called rivals on a number of counts.
  • JORUM. Providing support for the use of a wide variety of metadata standards when describing electronic resources is highly important: they provide appropriate ways of describing your resources so they can be indexed and discovered by numerous online services. While JORUM provides support for uploading bundles of resources using standard mechanisms such as IMS Content Packaging, it places a few limitations when it comes to supporting the use of multiple metadata standards. IMS specification is highly flexible in this respect since it supports combining not only IMS metadata standards but also other metadata schemas suitable for different types of resources and even the use of custom schemas. In contrast, JORUM only supports either using Dublin Core or using IMS metadata standards (IMS MD or LOM) – not a combination. We consider that in this area more flexibility is needed because, whilst the use of Dublin Core and LOM is very appropriate to describe educational resources, it might not be very useful when describing other types of resources.
  • eBooks.Outputting an eBook from a Word document is not easy. The standard route is to save the document as unfiltered HTML, then run it through an ebook converter. But there are a number of issues with this:
    • Unfortunately Word does not create composite images for diagrams that are not merely flat images (for example, anything using the diagram drawing tools). Although strangely, if the diagram is cut then pasted into a graphics program (GIMP, for example) they appear as composited images.
    • Word, even when outputting in its minimal HTML mode, writes all kinds of unwelcome formatting into the pages.  For example, bullet point lists are output with large amounts of physical spaces (  in HTML) injected into the text to replicate the indented margins. This causes problems with the converted book file.
    • Various converter tools (Calibre, Kinglegen, MobiPocket Creator) do a better conversion job that others, although none of them is perfect, and formatting is often lost or badly implemented. Ultimately we dropped Word in favour of Sigil, an open source tool for working with ePub documents directly.


Comments are closed.

Previous Posts