Open App Ecosystem

Building interactive websites in a post-framework web

Danyl Strype Wed 4 Mar 2026 10:44PMPublicSeen by 108

I recently stumbled across Daz's guide to building websites using HDA (Hypermedia-Driven Architecture), as an replacement for heavy and regularly obsoleted JavaScript frameworks like React. The finer details of his code examples go over my head, but what I think I understand from his description of the architecture, and the reasoning behind it, seems groundbreaking.

The examples in Daz's guide are based around using HTMX. A comment on SH mentioned DataStar, which is maybe an alternative to HTMX? Certainly seems to have similar goals from what I read on their homepage.

From what I read in Daz's guide, it seems like adopting this kind of approach to interactive web interfaces like Loomio, and fediverse web apps, could be a game-changer. Massively reducing the complexity of feature development and maintenance, and avoiding the cycle of having to massively refactor every time the JS framework in use becomes obsolete.

Thoughts?

Arnold SchrijverMon 13 Apr 2026 11:14PM

@Lynn Foster that is fabulous. Thanks.

Form an OAE subgroup to experiment with reimplementing Loomio using HDA

proposal by Danyl Strype Closed Mon 4 May 2026 3:01AM

What is the decision you need to make?

I propose that a subgroup of OAE members is formed to experiment with reimplementing Loomio using a Hypermedia-Driven Architecture approach. With the stretch goals of a) implementing a plug-in system, allowing Loomio to be easily extended like WordPress, and b) to build a plugin for ActivityPub federation.

Please note: It's not the expectation that any code generated in this experiment would be adopted by Loomio, or used in production at all. Although it would be under AGPLv3, as a derivative work of Loomio, so they would have that option. The objectives are simply to;

Test an HDA approach by applying it to an established Free Code app
Provide a practical project a group of people could participate in to learn about how HDA works

Why is this important?

The dependence of web app developers like the Loomio team on JavaScript frameworks like React keeps them on a treadmill, constantly tweaking the software to keep up with changes to the framework they're using, and forcing major refactors every time one reaches End of Life. In the case of Loomio, this sucks development time and energy away from their most important work; engaging with the people using it to make decisions, figuring out how to serve their needs better, and building new features and integrations that do that. It also sucks time and energy away from a project like rebuilding Loomio to escape the dependence on JS frameworks, creating a vicious cycle.

Forming a team to experiment with using HDA to build a web app based on Loomio is a way to learn about and test the value of HDA in practice. If it succeeds, the code we write could inform the next major version of Loomio. Or failing that, it could become a fork that some people hosting Loomio instances for their own communities might decide to switch to and support going forward. It could also be a good way to revitalise the OAE group, and obtain a yield from the amazing network of people who've been attracted here over the years.

What are you asking of people in this proposal?

If you have software engineering or web design skills, or you'd like to learn by doing, in collaboration with a group, I'm asking you to volunteer at least 2-3 hours each week for the next 3 months. If we can get at least 2 people committed to this, I think it's worth setting up a subgroup and proceeding with the experiment. If you don't want to be involved, that's fine, there's no reason to respond unless you strongly believe this would be an inappropriate use of the OAE group space.

We're not aiming for perfection — give your consent if the proposal is good enough and safe to try. If you have an objection, say what needs to change to make it workable.