If I go with the dependency hell title I would start by defining it and giving some examples. I’d probably start with something like the wikipedia definition:
So for example one of the issues I’ve had is trying to test a PR for an MFE that targets Sumac, but you have tutor 17.x installed which targets Quince (tip: Q from Quince is the 17th letter of the alphabet), so something breaks and you either have to upgrade tutor or have multiple installations of it (I use TVM for that).
Maybe I could focus on these kind of Dependency Hell-ish situations and their particular solutions.
I’ve tweaked my proposal from last time and here it is:
Title: How we made deploying 50 microsites 10 times faster: A story of theming
Description: By adopting runtime theming based on upcoming capabilities in the frontend platform, we cut down a potential 25hr deployment time for 50 micro-sites to under 3hrs.
Proposal:
Runtime theming not only allows you to change themes at runtime, but it comes with some remarkable new features and enhancements that we will explore in this talk.
When faced with a multi-tenant setup with a large number of tenants each with a separate theme, we had a huge scaling challenge on our hands, of which a key was theming and MFEs. A single deployment of the core platform can serve as many microsites as needed, so why can’t MFEs?
We explore how runtime theming can bring the same capabilities of a single deployment for dozens, even hundreds of microsites if needed.
A lot of great ideas in the thread. However, I am struggling to come up with one on my own. The edx-enterprise related work that I was involved in last year has come to a stand still this year with client freezing development and collaboration with 2U not panning out. From the lack of any major discussions or even responses to the enterprise related posts in the Open edX Forum, I think there isn’t a big userbase for these set of services either. So, it is not going to be of interest to most Open edX users.
The best I could do at this point with my experiece is to propose something like The edx-enterprise stack - Why we need an alternate?. Where I can try to identify the different parts of “enterprise” requirements and propose an alternate way to built it. However, the topic is just too big and I lack enough experience. If we can find someone else who manages an edx-enterprise stack deployment, it would be a great talk to co-present.
@antoviaque Thanks for the feedback! About this, I have this new angles:
Micro-frontend state architecture strategies on Open edX: Talk about state architecture patterns in micro-frontend applications on Open edX. Cover decisions around using React Context and Redux (We can make the comparison here), and how they fit into a micro-frontend architecture. We could include frontend-app-authoring as a use case.
React Context and Redux in the evolution of frontend-app-authoring: Talk about how the Open edX frontend has evolved, and explain the design decisions that led to the implementation of React Context and Redux in frontend-app-authoring MFE (We can make the comparison here).
@mtyaka Good idea - see it wasn’t too hard to find one :) Even if you are not the one presenting it, you can write a talk proposal for it, which someone else could take on or submit (or copresent with you) - @kshitij what do you think?
@daniel.valenzuela Sounds like a good approach What would that give for the proposal title & description?
@tecoholic That’s a good option I think? You could go ahead and submit something like that imho - if it gets accepted you could pair with someone who has additional experience with it.
@ChrisChV It’s better - but it remains very technical, and to have some prior knowledge of what react & redux do to understand the description. I would try to make it a bit more accessible still.
I did not check the ticket until after the deadline and missed the forum ping.
My ideas:
An overview of the architecture of Grove, and a demonstration/explanation of how to use it.
Quite self-explanatory;
In the past conference I briefly talked with someone about how they manage deployments and talked about how we use Grove and he seemed to find it interesting;
I feel like explaining how Grove works in addition to what it does might bring interest to it;
Can be of enough interest to a non-technical audience.
The state of editing XBlocks settings/fields (need a better title).
The default configuration part of studio-editable XBlocks lacks flexibility: you either take the form as it is generated for the XBlock’s settings, or you make an entirely custom page - there’s no straightforward way to have a custom input for a specific field only;
The implementation of list and set inputs is currently incomplete, it gives the user a text field for JSON;
It is possible to achieve some customisation by using Javascript - this is the main topic because it explores how to solve the problem;
Not sure if there’s enough to be said about it for a talk, could be something shorter;
A proposal to allow per-field customisation? Possibly a pull request that will be ready to be demonstrated by the time of the conference - I might be able to do it in the context of some future XBlock that would benefit from it, or as part of talk preparation if it gets selected;
The current state of theming seems to still be in transition from branding packages to design tokens, and I think it might be a good idea to have a talk about that. Unlike last conference this direction now seems to be confirmed. We can maybe more generally talk about the current state of customisation, including:
basic styling of css using design tokens
deeper customisation of content and layout using plugin slots
customisation of translations using atlas
The second and third have already been covered to an extent in the last conference, so the focus can remain on changes since then.
@antoviaque Just returned from leave, so apologies for the last-minute ping! If it works for you, I’d like to propose this talk:
Mission Impossible: Building 7 High Quality XBlocks in 2 Months for WGU
Thanks for making the title so much more catchy! FYI: I changed the number of XBlocks from 12 to 7 as per @pooja’s response to my question here. Although the number is not as high, I think it still sounds like an impressive feat! Does it get your +1?
@artur It could potential be a good talk topic - but without being more direct about what it is that grove does (and how it does it) that is interesting? Remember, most people won’t know what it is, so that won’t be obvious to them. Why would people in the audience be interested in hearing that? What is it, in what/how it does things that is different? What can people achieve with it that is unique?
I think it’s better to leave React and Redux in the background, explaining the basics so that we don’t need such technical prior knowledge to understand the talk:
Building the frontend of Open edX: Patterns of State Organization in Micro-Frontends: Explore how Open edX organizes and manages data in micro-frontends, making it easier to create seamless and customizable user experiences. This talk will cover the key principles that guide decisions about how to share and handle information in different areas of the app. We will show practical examples of how these decisions are made in the frontend-app-authoring project, and how tools like React Context and Redux are used to build more modular and scalable applications.
It makes it easy to deploy and upgrade an Open edX instance, in a same repository where you would keep its configuration. The talk can include a demonstration of this functionality (not live as it takes quite a few minutes).
More interesting to a technical audience but a non-technical audience will be able to follow, and understand what is being automated and made easy.
It uses Tutor so it supports Tutor plugins.
More interesting to a technical audience. Can be a short mention/demonstration.
How using GitLab CI makes it convenient to perform deployment actions.
Can be made interesting to a non-technical audience as well by showing how it makes it easy to deploy and see progress and errors.
It supports AWS and DO (I think the architecture that enables multiple provider support, with provider-specific adaptations as Terraform files, is technically interesting).
Too technical, but can be a short part of the talk: short overview of the architecture, then of where the AWS or DO parts are used to provide specific funcionality.
Built-in support for monitoring tools.
Interesting to both technical and non-technical audiences.
@antoviaque Ah! shoot. I realise you said, I could “submit like that” and not submit that. I would like to propose:
Handling bulk enrolments for enterprise customers
A dive into edx-enterprise stack and how HMX uses it for custom catalog, bulk licensing and enrollment.
Enterprise features in Open edX - the present and the future
Taking stock of the features that make up edx-enterprise stack of applications. How are they implemented today and how can we make it more open and easy to adopt going forward.
@ChrisChV Your proposal looks good Kudos for the improvements!
@artur Good points - now how would you distill the essence of these points, into a single talk proposal?
@tecoholic Looks good! You could probably make the first topic (“Handling bulk enrolments for enterprise customers”) a subtopic of the second one ( Enterprise features in Open edX - the present and the future)? Then you have a great talk proposal I think
Since both @cassie and @antoviaque have proposals that hit this topic, I wasn’t going to submit another one, but I’d be happy to co-present or otherwise help prepare.
What would that give for the proposal title & description?
: