4 ms·

I find the "state of Elm" over the past couple of years absolutely fascinating. Some folks reasonably claim the language is dead and absolutely not suitable for

by bbkane 3y ago

I find the "state of Elm" over the past couple of years absolutely fascinating. Some folks reasonably claim the language is dead and absolutely not suitable for production due to scarcity of updates, public roadmap, or bug fixes. Others run Elm in production and reasonably claim it's perfect/stable already. There's a steady stream of blog and forum posts on this argument. I read them all.

xtian 3y ago
It’s because the only people left in the Elm community are those who are content with its current state and aren’t interested in having a stake in the future direction of core components or in even knowing whether there is a direction.
dartos 3y ago
There are only 2 kinds of tools. Tools people complain about And tools nobody use.
- pcstl 3y ago
  Yeah, I don't think this quote fits here. Elm had a massive shot at making the big time circa 2015-2016. If they hadn't blown that shot, yeah, they'd be the kind of tool a lot of people use and complain about. But what I see when it comes to Elm is a lot of people who don't use the language shaking their heads in frustration at what they see as missed potential, while a handful of people who do use the language insist that it's perfect and unchangeable and that Evan could never ever make a bad decision, all the people who disagree must just be short-sighted.
- satvikpendem 3y ago
  As the sibling comment implies, the fact that we see fewer and fewer complaints compared to a few years ago leads me to believe that Elm usage is dying.
noelwelsh 3y ago
Agreed. It's amazing how much history is determined by small decisions. As another commenter said, Elm completely blew their chance, which is essentially down to the actions of one guy making what seem, from the outside, obviously bad decisions. At a meta level, there has long been demand for a modern ML but all the contenders have fumbled the bag in various ways. E.g. OCaml was "nah, multi-core isn't important" for at least a decade.
- pharmakom 3y ago
  F# is the most popular ML derivative now. If it had a few more of the type system features in OCaml it would be perfect.
- truculent 3y ago
  > Elm completely blew their chance Blew their chance at what? And what trade-offs would have been necessary to take that chance? Those questions are at the heart of this talk, and I think make it clear that there is a much larger range of definitions for "success" than is often assumed. And that perhaps we should question the default definitions. Perhaps the decisions, if not the ones you would have taken, are not so obviously bad (or are at least understandable) with that in mind?

by bbkane 3y ago

xtian 3y ago

It’s because the only people left in the Elm community are those who are content with its current state and aren’t interested in having a stake in the future direction of core components or in even knowing whether there is a direction.

dartos 3y ago

There are only 2 kinds of tools. Tools people complain about And tools nobody use.

pcstl 3y ago
Yeah, I don't think this quote fits here. Elm had a massive shot at making the big time circa 2015-2016. If they hadn't blown that shot, yeah, they'd be the kind of tool a lot of people use and complain about. But what I see when it comes to Elm is a lot of people who don't use the language shaking their heads in frustration at what they see as missed potential, while a handful of people who do use the language insist that it's perfect and unchangeable and that Evan could never ever make a bad decision, all the people who disagree must just be short-sighted.
satvikpendem 3y ago
As the sibling comment implies, the fact that we see fewer and fewer complaints compared to a few years ago leads me to believe that Elm usage is dying.

noelwelsh 3y ago

Agreed. It's amazing how much history is determined by small decisions. As another commenter said, Elm completely blew their chance, which is essentially down to the actions of one guy making what seem, from the outside, obviously bad decisions. At a meta level, there has long been demand for a modern ML but all the contenders have fumbled the bag in various ways. E.g. OCaml was "nah, multi-core isn't important" for at least a decade.

pharmakom 3y ago
F# is the most popular ML derivative now. If it had a few more of the type system features in OCaml it would be perfect.
truculent 3y ago
> Elm completely blew their chance Blew their chance at what? And what trade-offs would have been necessary to take that chance? Those questions are at the heart of this talk, and I think make it clear that there is a much larger range of definitions for "success" than is often assumed. And that perhaps we should question the default definitions. Perhaps the decisions, if not the ones you would have taken, are not so obviously bad (or are at least understandable) with that in mind?