clojure VS re-frame

I created a simple frontend for my SimpleServer exercise a couple of years ago: SimpleFrontend. In that exercise I used Reagent library which is a minimalistic React wrapper for ClojureScript - I have documented the exercise in Become a Full Stack Developer with Clojure and ClojureScript! blog post. At my company, Metosin, we also use quite a lot another ClojureScript framework that provides some additional utilities like state management - Re-Frame. In this blog post I describe my recent Clojure exercise when I re-implemented the SimpleFrontend using Re-Frame.

Reagent is nice and has been around for about a decade now, but I moved away from it towards very thin wrappers around React[1], because I felt like it was adding too much additional complexity on top of React, which is already quite complex on its own. I wanted a clearer view at what is going on and a simpler way to interop with native React components.

Although it seems to catch up with experimental support of React 18 now, Reagent has fallen behind the latest developments in React and may not benefit from all of its performance optimizations. It is still using class components instead of hooks and there have been concerns that the runtime conversion of Hiccup may drag down performance. I guess in most cases it is not really an issue or in any way noticable, so if you’re not doing any fancy stuff it should be fine. I may even come back to Reagent at some point, since I have to admit that I miss the UI-as-data model with Hiccup.

What I highly recommend, however, is using re-frame[2] for state management. It has also been around for a long time (2014, around the same time Reagent came along) and pioneered some popular ideas in that area. It may seem a bit overwhelming at first, but the docs provide a great introduction and I find the model very clear once you wrap your head around it. At the moment it depends on Reagent, but there are ways around that. [3]

[1]: see Helix (https://github.com/lilactown/helix) or UIx (https://github.com/pitch-io/uix)

[2]: https://day8.github.io/re-frame/

[3]: refx (https://github.com/ferdinand-beyer/refx) is an almost drop-in replacement without the Reagent dependency, but hasn’t been updated in a while. Alternatively, re-frame can be integrated with UIx/Helix by adding some interop code https://github.com/pitch-io/uix/blob/master/docs/interop-wit...

This article always reminds me of this excerpt from re-frame’s docs [0]:

> Now, you think and design abstractly for a living, and that repetition will feel uncomfortable. It will call to you like a Siren: "refaaaaactoooor meeeee". "Maaaake it DRYYYY". So here's my tip: tie yourself to the mast and sail on. That repetition is good. It is serving a purpose. Just sail on.

[0]: https://github.com/day8/re-frame/blob/master/docs/correcting...

* Single-Page App: shadow-cljs for the build concerns (https://github.com/thheller/shadow-cljs), Reagent with Re-frame for complex/large app (https://reagent-project.github.io and https://github.com/day8/re-frame). Even if we now prefer using HTMX (https://htmx.org) and server-side rendering (Hiccup way of manipulating HTML is just amazing, https://github.com/weavejester/hiccup).

Big and/or complete projects that use re-frame The main list: https://github.com/day8/re-frame/blob/master/docs/External-Resources.md

Reflet is a set of tools for building Re-frame + React based web apps with graph and non-graph data models. This includes:

https://github.com/day8/re-frame

Notably the author of re-frame has been weary of hooks.

I think that’s for a good reason. The approach in re frame feels like the best way to manage state so far for a react based app. Everything that changes state flows through an event. State can only be observed through subscriptions. Side effects are isolated to their own type of event. Debugging and testing are so straight forward with these concepts.

Redux got close but it has two problems in my mind. Like hooks, it encapsulates for no good reason. Put the state in one thing that you can observe holistically before and after pure events. It also has too much boiler plate.

The cljs stack I hear about a lot (and use) is ShadowCLJS with reagent (https://reagent-project.github.io/) and re-frame (https://day8.github.io/re-frame/). ShadowCLJS is more of a build tool, but is really well documented and easy to use. Reagent is basically react but a simpler API, and re-frame is a layer on top of that provides data subscriptions and event-handlers to manage app state. It's overkill for some apps but I find it's actually super easy to work with and not as much complexity as I thought.

For backend there is luminus (https://luminusweb.com/) or Kit (https://kit-clj.github.io/). They are basically project templates that wire together a ton of popular solutions for various things - database access, migrations, security, html templating, etc. Also includes frontend frameworks like re-frame if you want.

Reflet aims to be a natural progression on top of Re-frame to support complex, data driven requirements. In that sense, it is both easy to learn, but powerful. You could say it's sort of like Re-frame++ (or Fulcro for Re-frame). Its main design goals are:

At my current client we're working on having a frontend architecture for writing SPAs in JavaScript similar to re-frame's one: an event-driven bus with effects and coeffects for state management[1] (commands) and subscriptions using reselect's selectors (queries).

Some of us had worked with effects and coeffects before while developing SPAs with re-frame and had experienced how good it is. After working with re-frame, when you come to horizon, you realize how a good architecture can make a dramatic difference in clarity, testability, understandability and easiness of change.