Elm 0.19 Broke Us 💔

[nelsonic]

https://dev.to/kspeakman/elm-019-broke-us--khn

[rub1e]

I clicked this by accident but I was curious so I read the first (non-technical) part of the article

What are your thoughts re the need for an engaged and super-positive community (and the impact of not having this) for new technologies like this to succeed?

[nelsonic]

A welcoming community is the difference between success and failure.
Evan's recent talk "What is Success?” is really worth watching: https://youtu.be/uGlzRt-FYto

Kasey Speakman (the author of the "Elm 0.19 Broke Us") has some "justification" for being "heart-broken" about Elm "breaking" his love of the language/platform because the changes have "affected" him (and his team) personally and he has some grain of truth with regards to Evan being a "BDFL" who closes issues swiftly (often unceremoniously) but a fact that is not captured in this blog post is that building an Open Source package or (in the case of Elm) entire ecosystem is like building a City. Essentially Evan has "pedestrianised" one of the main roads in the centre of the city to make it more "walkable" but it means people who are used to "driving through" can no longer do it.
This will make existing "users" angry but it's an "investment in the long term".
The reason I mention City-building is because it's large and complex and changing things affects many people. And the city-planner is a busy person who needs to make difficult decisions.

Indeed the comments thread on HN balances out the "one-sidedness" of this post.
see: https://news.ycombinator.com/item?id=17842400

Evan works for No Red Ink which is by far the biggest user of Elm.
Everyone at No Red Ink got a chance to review/test Elm 0.19 pre-release and if they felt that it was a "bad idea" they would have informed Evan.
The most recent Elm Town Episode: https://elmtown.audio/upgrading-to-elm-019-luke-westby-richard-feldman features Luke Westby & Richard Feldman (both of No Red Ink) and they discuss the 0.19 release in more detail than the blog post.

Ultimately I feel that 0.19 is a good release that benefits both end-users of any Apps built with Elm (because compilation is faster and resulting apps smaller) and beginners have less to get confused by; custom operators are a "nice to have" but they make code "terse" and difficult to debug.

Interestingly, Elixir (our other programming language) of choice @dwyl is going in the opposite direction to Elm in that it seldom removes features from the language and has an incredibly powerful "meta programming" language (macros) built-in which allows users to extend the language to their heart's content.

I feel the pain of people who have "lost" a language feature they used/loved but I think it's better for the overall ecosystem and specifically for beginners joining an Elm project.
I'm still "bullish" on Elm and will continue using it for the foreseeable future.
Elm is "Touring Complete" and has all the features we need to build full applications that are fast, reliable and easy to maintain.
I actually love the fact that Evan is "pruning" features that are "stumbling blocks",
I think the language should be as minimal as possible much like the original lego.
And if people do not have the imagination to build using the "basic" blocks, they should "graduate" to using the more advanced PureScript/Haskel or ClojureScript so they can have all their "advanced" features.
People easily forget (or don't think about in the first place) that the original Google Mail (GMail), Google Docs and Google Maps were all built using ECMAScript 3! and it did not "stop" them from building "advanced" applications with good UI.

If the engineers at Google could build/ship entire/advanced web applications with ES3 which did not have built-in type-checking, advanced compilation and time-travelling debugging, then people whining about Elm 0.19 need to re-think how they are over-complicating their own lives!

[rub1e]

V v interesting - will have to read this properly and check out some of the links!

[nelsonic]

GOTO: dwyl/learn-elm#129 (comment)