From zero to merge in record time: What PR authors can do to accelerate review timelines
Getting PRs merged quickly is a two-way street. Maintainers must respond to new PRs in a timely fashion and follow up on updates from their authors without too much delay. On the other hand, there’s a number of things that PR authors can do to maximize their chances of a smooth review process. This lightning talk would focus on the latter.
@DouglasDraper @kaustav @Agrendalath @artur @maxim @paulo @farhaan @tecoholic @mtyaka @daniel.valenzuela a quick reminder that today (Nov 1) is the deadline for posting your talk topic ideas 
I don’t have any solid ideas, I can co-present with someone.
One topic I don’t know much about, but want to learn and get improved in Open edX is integrations with other platforms, like SalesForce and/or other CRMs. Clients often want better features for managing their learners, collections of courses, subscriptions, etc, and Open edX doesn’t provide such functionality. The state of the edx-enterprise is sad, it’s basically unusable right now with Redwood on Tutor/Grove set up. Course discovery service is nominated to be deprecated. I remember there was a ticket on the roadmap about improving integration with SalesForce (@tikr do you remember what it was?), so may be between now and then I can learn more about that and present on such topic.
I really like this idea and am familiar with the bounds of the problem. I’d love to do this talk if I’m still around by then.
Retrospective: How has the community changed since the 2U acquisition of edX?
I wanted to do a follow up blog post to this post I made a few years back musing about the 2U acquisition. A lot of the things I predicted were spot on, though there were some curveballs. There was a lot of skepticism and concern at the time and I think it’s important to take some time reflecting on how the community has changed since then. In my opinion, things are a lot better, though not exclusively so.
I did touch on some of these on my lightning talk about React Query last year, but I’m sure there’s more to cover and I would appreciate your help getting the word out more :)
I think this re-worked talk will have a lot more appeal to people! And also be helpful for everyone voting in the election. Great idea.
Here are some ideas that come to my mind :
Open edX at Scale - How we scaled an Open edX instance to handle 30k users across 50 microsites.
We can talk about what an instance running at this scale looks like, the tools that we used to achive this scale and what challenges we faced along the way. cc.@paulo
Supercharge your PR reviews with automated PR sandboxes
This could be a lightning talk about the PR sandbox automation that we operate for Axim. We can talk about how many PR this has benefitted already, how to spin up and configure a sandbox for your own PRs.
Why can’t I just delete a user ? Demystifying User Retirement in Open edX
(This idea is from last year, which I never submitted because I went for a different proposal.)
I only came up with a single idea, but a ton of these above look great, and would love to help if I can. So here is my talk idea:
Exploring the OpenedX development and distribution tooling (alt title: Escaping dependency hell, the OpenedX Case)
An overview sprinkled with workflow suggestions and opinions on all the pieces required to install and run Openedx for development and production use cases. My first approach would be going from a bottom-up approach:
The challenge was to extract data from “millions” of xblocks and passing it to an external service (AI) to get meaningful tags and keeping these tags up to date as an when xblocks are updated. We also had to allow users to vote and validate these AI tags and finalize them. I don’t think it was really hard due to existing tools/helpers like openedx-events and openedx-filters, so the plan is to show it as an use case of hooks extension framework. Also include some information about the “new” event bus that uses Redis instead of Kafka.
@navin It might be worth mentioning the specific challenges early on, even within the title if possible. It’s not just labeling them, it’s categorizing them automatically (you can even say “with AI!”, which is fashionable currently 
) and through student contributions.
@daniel.valenzuela Good one I think, though it’s currently hard to tell exactly what it would be from its title, it might benefit from focusing on a specific aspect (the “dependency hell” sounds like a good one - but what is hellish precisely about it for Open edX?)
@maxim If you review your work over the past year (or even before), I am sure that you can find some topics to present? Everyone does the effort to come up with ideas, could you also try? It’s fine if it’s not the best idea immediately, that’s what the brainstorm is about 
@DouglasDraper @Agrendalath @artur @paulo @farhaan @tecoholic @mtyaka – same thing, the talks suggestions were due on Friday, could you post yours today?
@maxim I think it’s this one: platform-roadmap#267
Good point. I’d talk about the model-agnostic architecture that allows educators to choose the AI model best suited for their needs. Additionally, the integration with Judge0 for the “Coding XBlock” external code execution ensures a secure solution for coding exercises. I could also highlight the personalized feedback capabilities, which make these XBlocks suitable across a wide range of subjects.
We don’t exactly have any tangible results or numbers yet, but I could talk about how the “Timeline” and “Accordion” XBlocks can transform static content into manageable, interactive formats which should enhance focus and reduce cognitive load. I could also talk about potential use cases of these XBlocks.
I am sorry for the late update.
I resonated with @kshitij 's idea but I would word it differently:
Runtime Theme: How it works and how it saved 100s of development hours.
New CAPA Problem Editor: The journey to develop this editor and how it can be made better.
A very similar talk for Video editor
Or I can combine both of them into, Challenges in developing a new component editor
I would also be interested in giving or contributing to a talk about Meilisearch - why/how we’re using it and what it enables for us compared to ElasticSearch. (Suggested by @dave on Slack).
Sorry, I overlooked the deadline for this task.
I’d like to propose the “Migrating from Ansible-based deployment to Grove” talk. Since the configuration repo is officially no longer supported, it would be good to have a clear guide for operators on how to deploy their instances after upgrading to Redwood.
I’m struggling to find a way to transform my recent work into a talk. Much of it involves metawork for ASU, redwood bug fixes, monitoring and scaling up their production cluster (as Kaustav mentioned earlier, and this relates to my entry from last year), as well as DevOps infrastructure for their migration discovery. In short, work very specific to ASU.
The only thing I can think of is:
Thanks for the additional proposals - all looking good. 
@maxim Your proposals look great, thanks for coming up with them!
@Agrendalath This one can be a good one, but it would be good to make sure it’s not just a tutorial/workshop. Also there isn’t just grove installs - there is Harmony (which we should advertise first since that is part of the official project, and we want other providers to join it), but also people just using Tutor, or plain installs. Maybe a review of those approaches?
@paulo Maybe a talk about the challenges of large government-level projects would be something you have lots to say about? :p Joke aside, maybe after the Ethiopian project is completed, or has reached large milestones - it could be something to copresent with ASU and the Ethiopians? In the meantime, your other proposal look good.
I’m having a very hard time coming up with anything that hasn’t been mentioned yet. I am not familiar with current state of theming, but I would definitely like to see a talk about that.
+1 Ability to debug issues in production is one of the main issues I have with Tutor/Grove/docker. It was so much easeir back in the VM days.
Good point! If we go this way, it definitely won’t be a workshop.