This thread will be used to brainstorm on talk topics for the 2023 Open edX conference. Here we should share all potential ideas we can think of, whether we would present them ourselves or not. It’s meant to help the team as a whole, so we have a big pool of talks to pick from. Other community members who read this forum are also welcome to chime in
All other aspects of the conference (attendance, planning, etc.) are being discussed in this other thread.
I will soon be scheduling timeboxed tasks for each of us to submit talk topic ideas here in the next sprint. You can also browse previous, similar threads (2022, 2020) for ideas. Feel free to make suggestions anytime starting now, if your schedule allows it!
Once we have a list of ideas, we can start working on scheduling tasks for submitting proposals.
I’d love to do a talk on providence for frontend state management. People were pretty excited about it when I teased it as part of the MFE workshop.
A talk on how Open edX compares to other LMSes like Moodle and Canvas would be helpful, I think.
I’m not a core committer, but a talk on the program from the perspective of a participant would be cool.
A talk on effective principles for running all-remote teams would be a topic not DIRECTLY related to Open edX, but useful to many who are using it. It’d be a great opportunity to talk about Listaflow, the core committer program, and lessons we’ve learned both before and after remote work became widespread.
"An inside look at the governing body of the Open edX project: the TOC"
What is it, what can it do for me? Who does what?
This talk could be a collaboration between TOC members, presented as a team - I’ll suggest it to the group during the next TOC meeting
And these two proposals from last year could still be quite relevant:
"Learning from open source communities" Open edX is one of many open source communities. What can we learn from others? Through a review of other large open source projects and of the literature on the topic, we identify where the Open edX project differs compared to the established practices, and what we can all do to improve our community.
"Building collaborative classes - Borrowing from open source practices" For educators willing to build a more collaborative environment in their online courses, we present a few of the recognized best practices in open source. Many open source communities have successfully mastered many of the challenges related to building collaborative online projects. Their experience can provide a useful perspective to those looking to develop a different kind of engagement with their students. To not have students just consuming a course to pass it, but contributing to building and improving it too, for the benefit of everyone’s knowledge.
Also, it could make sense to suggest some core contributor workshoping / synchronous sprint sessions, where we get together to work on things we need to work on, and could benefit from a boost, like:
Brian Mesick and I have discussed presenting together on this Will be great to showcase this part of what the data-wg has been doing.
Yes! I also think @kshitij 's MFE tech talk would make a great workshop, and if it could include demonstrations for how to use Providence, even better.
Given the discussions that came out of ^, a talk on “Customizing/extending MFEs” would be really useful too… even if we can only really customise “themes” right now, it would be nice to hear about the proposed architectural direction to make customisations more maintainable in the near future.
I did the initial work in that project. Course metadata is now moved from MongoDB to MySQL, but most course content is still in MongoDB because nobody has yet developed a working alternative. Blockstore could be made into that alternative, but it still needs a lot of investment and it seems that the approach of storing data on S3 is actually too slow in practice (as I had feared). So I think the replacement/next steps for the “remove mongodb” project will be either (1) Write a new ModuleStore backend that stores data in MySQL, or (2) Something new based on learning core, or (3) Fix blockstore’s latency problem and implement blockstore-based courses.
I would be happy to work on any of these if someone wants to fund it :)
This is difficult to read! I had some suspicion this might happen as well. Do we have any ideas how we might make it faster? We at least have the advantage now of a well-established API, so presumably we’d only need to focus on how we’re doing the storage. If S3 is out, what might a suitable alternative be? Or do you suppose there’s a way to speed up how we’re using S3?
The fact I find myself asking these questions and being curious does indicate to me that this would be a good talk idea-- if there were evidence of a good path forward, the talk might spur action. And I would figure that much of Modular Learning’s long-term goals are going to require we find a solution.
yes, I think Dave and I could give a good talk along those lines.
It’s true that most of blockstore can still be useful it we find an alternative backing store besides S3. I’m not sure that’s the best approach, but it’s worth considering. Fundamentally, Blockstore groups OLX and Assets together and treats them all the same. But some assets are so big (e.g. videos, PDFs, etc.) that they really have to be stored on S3. So Blockstore can’t just use a new backing store; a different approach is needed.
A relatively simple and sane way forward would be to store OLX in MySQL but with a strict size limit, and require that OLX above the limit needs to use an include or reference to an external .xml or .html file for the rest of its data; such assets (along with PDFs etc.) would be stored in blockstore on S3. That would solve the latency problem, eliminate MongoDB, and be an easier migration for courses than migrating to the current blockstore approach of storing OLX on S3 and trying to make up for the latency with several layers of caching.
@Agrendalath Kudos for getting it accepted! That said in the current context, the expense would probably not be worth it for just a lightning talk - even though that would be a nice excuse to see you! Unless the talk could be passed on to someone who comes for another talk? Not sure if we have someone with the right context for this.