Removing React meant relearning where an interface should live
A 2023 account of replacing React with htmx resurfaced on HN, where the practical appeal was not an anti-React slogan but reducing the amount of client-side machinery a forum had to own.
The migration moved interactivity toward server-rendered HTML with htmx rather than a large client application. That trade changes the boundary of the system more than it changes a library name.
The useful question from the thread is not whether React is bad. It is whether the product really needs a client-side state architecture before it needs simpler operations.
A framework removal is an ownership decision
The Misago migration account describes removing React and adapting htmx for UI interactivity. The attraction is concrete: return HTML from the server for many interactions instead of shipping and coordinating a separate application state in the browser.
Less JavaScript is not the whole story
For a server-rendered product such as a forum, the move can shrink build complexity and make more of the page visible in ordinary HTML. But it also relocates responsibility. The server now owns more interaction flow, and the team must be deliberate about latency, partial updates, accessibility, and browser history.
That is why “replace React with htmx” is not a universal recipe. Rich offline behavior, heavily collaborative editing, and local-first interfaces may still need substantial client state. A content-oriented application may not.
Start from the interaction, not the stack
The enduring lesson is architectural rather than tribal. List the interactions a product actually needs, then decide which of them require a durable browser-side model. Everything else can remain a request for HTML.
That approach produces a smaller system only when the team resists rebuilding a client framework in fragments after removing one.
Get breakdowns like this every day
The top Hacker News stories, unpacked — by email, Telegram, or RSS.
Solar-panel washing has a surprisingly strict burden of proof
A practical post asked whether washing solar panels is worthwhile. HN turned it into a maintenance calculation: a visible layer of dirt is not yet evidence that cleaning recovers enough output to justify water, labor, and risk.