Open edX - Assessment Content Setup Improvements

The following potential improvements have come up as a result of this research (not necessarily tied directly to this area of configuration):

• Research where OLX needs care and attention

  • OLX is a massive potential differentiator for DIY course authors who want to do insane things with courses, but I cannot say specifically where OLX needs to be improved, as my experience with it is purely in a supporting role, rather than as a principle creator of XML problems.

• Ensure that OLX is properly supported and documented

  • Most OLX documentation is as old as the platform, and there are very few resources designed to ensure that OLX can be used by advanced authors who are willing to build.

• Ensure that core supported OLX is robust and maintained

  • This is a trickier one, but we have completely custom tags set up for things like the Circuit Schematic Builder. By supporting those tags, we effectively declare those features to be supported by the core. By not supporting them, we’d blow up a decade of work using them. But most of these advanced problems that have been integrated into OLX are completely unmaintained. This line is far too fuzzy.

• Ensure all core components have higher quality editing experiences

  • There are frankly no components in Open edX that have a truly good editing experience for authors, and while I’d say a lot of our tools are still generally better than Moodle, more modern tools have made significantly higher investment in author UX.

• Maintain the flexibility of problem editing

  • The easy solution to problem editing is, as has already been done in the 2U editor, to strip the capabilities of editing back to a fully-opinionated, restrictive editing experience, similar to Canvas. But this would detract a huge amount of flexibility from our editor, and reduce us at best to parity with them, rather than delivering an exceptional, unique experience of our own. The existence of GIFT shows that syntax-based editors have historically been seen as the antidote to inflexible, slow, clicky UIs by power users (to the point where GIFT may be what our existing editor was based on to begin with), and OLX isn’t going to fill that need due to the inherent clunkiness of tag-based structure. While we need to provide a simple on-ramp for new authors, we need to be careful not to alienate experienced users.

• Improve the workflow and UX of content libraries

  • Content libraries can only be used to access content from existing libraries from a long dropdown list. There is no way to browse or search for content, and the only way to access library content is to use the randomised content block. This makes content libraries inflexible and generally not fit for purpose because it does not matter how content is stored in libraries if accessing that content is a poor, inflexible experience. The baseline is that it needs to be a huge amount better than the experience of copy-pasting content, which it is not currently.

• Improve the sharing of content in libraries

  • Our current system has very specific access rights associated with content libraries that mirror courses. A user must be given specific access to a library in order to use it in a course, which also gives them full control over the library.

• Provide the ability to add existing content to content libraries from within the course interface

  • The current model of centrally creating content for libraries is incomplete, there needs to be a good loop between creating content for libraries, and using it in courses and creating content in courses for use in libraries.

• Cultivate relationships with high-profile third-party LTI tool providers to ensure that Open edX is both supported and advertised as an option

  • Many tools require a few extra steps to make sure they work with Open edX on the tool provider’s side. Even those that do not often have a list of supported platforms that lend a lot of legitimacy to those platforms, and not being listed can make Open edX seem less respectable to users of those tools, or reject the platform as an option based on not being supported by their favourite tool (even if it does actually work).