2022-06-01 Announce Django From React to htmx on a real-world SaaS product: we did it, and it’s awesome! ¶
TL;DR ¶
We took the plunge and replaced the 2-year-of-work React UI of our SaaS product with simple Django templates and htmx in a couple of months.
We’d like to share our experience with you, with concrete indicators on various aspects, and convince your CTO !
- 🔍 You’ve probably heard of htmx, maybe seen talks about it (maybe even
-
right here last year). Demos are great, potential seems enormous. Maybe you’ve heard it’s great for quick prototyping and tried it.
- ❓ But what about switching for your real-life project, against everything
-
you’ve heard since 2016 (”modern web interfaces need Javascript-driven webapps”, “the best SaaS products are made of an SPA”, etc.) ?
Does htmx keep its promises? What impact on your product, your team, your business?
- ✅ That’s what we just did at Contexte, by getting rid of the React SPA
-
of a SaaS product. It wasn’t a easy decision to make (because the UI holds some fairly complex interactions and we were told such interactions require client-side state management), but it’s now such a relief.
Maybe if we share our experience the decision would be easier to make for others.
💬 Draft outline of the talk:
-
What is Contexte? mainly about our Product/Tech team size and organization
-
What is our first SaaS product, Contexte Scan (click here for a Google-translated product presentation)? a quick demo, with an emphasis on the most complex UI elements (I’m not here to sell it to you 😉 )
-
Our React UI: bulky, buggy, constantly refactored, dependency hell (quick tour, some figures)
-
Our htmx UI: a bit of code here, with examples of how you replace client-side state management with just some htmx-loaded HTML fragments
-
Impact on end-users: same UX, better performances
-
Impact on the team: faster, fullstack, more agile
-
Impact on the planet: smaller footprint on end-users computers
-
Conclusion: go for it!