The State of Elm survey results have been released! You can view them here.
Thanks to everyone who participated and a special thanks to @wolfadex who helped me categorize a lot of the free text answers.
25 Likes
Thanks for the great work on State of Elm! I liked the way you described each section but couldn’t understand why the number of participants is lower than 2017 and 2018 results. I had a quick lookup on number of users in different communities for Elm:
- Elm Slack’s #beginners 20k+
- Elm Subreddit 10k+
Even if 10% of users in each of the communities participated in the survey, the number of participants would have been higher than previous years! Maybe I’m wrong, but quantities have the same importance of qualities in such surveys. What do you think?
1 Like
I was a bit surprised to se the us so high up in “Which country do you live in?” but clicking on per capita shows the real answer:
6 Likes
And as I remember answering " In which application domains, if any, have you used Elm" I chose “Gaming” since that is my main use of my free time but I had problems choosing the relevant one for my full time job. Maybe I answered “Communication” in the end which was the most relevant but not a perfect match. Maybe split this up in the next one(?) and only ask what you are using Elm for professionally etc…
Sorry for “spamming” but very happy to see “plain CSS” competing with “elm-ui” in “What tools or libraries do you use to style your Elm applications?”. Elm UI is amazing I can imagine (I have not tried it yet) but for us web-seniors nothing beats CSS. “Use the platform as they say”. 
2 Likes
I think that’s a common pattern in open communities. They tend to accumulate users who are no longer active and therefore the total user count is misleading.
I’m not sure what you mean by “quantities have the same importance of qualities in such surveys” though. Do you mean that quality is more important than quantity when it comes to having active users? (if yes, then I definitely agree)
That’s a good suggestion! I bet there will be more feedback coming soon so could you (and anyone else) who have suggestions write them as github issues here so they don’t get lost?
1 Like
I am also a bit confused about the apparent decrease in “interest” this could be perceived as. My “gut feeling” is that this year there has been more talk about Elm than ever before but that is kind of hard to see “objectively”. I tried to find a longer spanning chart than per week of installing Elm using npm like this:
Is there a way to see that through years? Would be interesting and in my mind represent good enough the rate of usage of Elm…
There’s elm | npm trends
Note that this graph shows less downloads than there really are, since some of us use direct downloads or elm-tooling (Announcing elm-tooling CLI) which doesn’t download Elm through nom.
2 Likes
A small update, I noticed I forgot to include the 2016 survey participant count. It looks like this now
I’m one of those people
2 Likes
As someone who has been a fan of functional languages since 2002, and someone who was part of a team that adopted Elm pretty early (before moving on to other jobs), I really think Elm could beat Typescript/React in market adoption if Elm and Rescript community joined forces under one unified language, ecosystem, and community. I know that’s nearly impossible and it would take a lot of compromises from each language to meet in the middle, but people can still dream.
Maybe the work within f# is the where elm and ReScript react meet in the middle.
1 Like
It is interesting that “functional programming” is not the top thing we like about Elm. I read this as we are pragmatic about it and the design of the type system is more valuable.
Also, the core team is on the top of pain points. Hm.
1 Like
Well to be fair the type system is pretty much the type-system you get with a ML-style functional language - personally I’d add this to “functional programming”
1 Like
I can certainly agree with that.
I considered putting these under the same group since they are closely related. The main reason I didn’t is because many aspects of the type system are available in non-FP langauges, e.g. Rust and TypeScript, and not all FP languages have static types. A lot of people love Lisps because of their FP aspects as well as dynamic types. Those same people might like Elm for it being FP, but dislike the static types.
Separately, I definitely felt like that section could have benefited from more elaboration. It feels good to see someone else liking Elm because of its type system when you also like it. There’s a sense of belonging there, of mutual understanding. Trying to separate myself a little from the language though, I’d like to know what about the type system people like, what about “FP” people like. Do they like it because it’s familiar? Do they like it because it’s novel to them? I feel like there’s a lot that could be learned from why people like these things. But I also might being going down the wrong path by trying to look for those things.
1 Like
I think the survey was/is fine.
I can only speak for myself but I certainly did not think as deeply about the questions and their implications while answering them.
And yes of course you are right - type-systems are not necessary for FP but I think the syntax and the style elm is putting itself firmly in a family of languages where Hindley-Milner is a defining factor and for me it’s a big part of what makes Elm “safe”.
1 Like
elm-ui falls short for me in aspects of Responsive design, animations, lack of support for grid like designs (based on css-grid in contrast with elm-ui being flex heavy) and it feels too much more opinionated on some things, so in the end it prevents me from making the design I was tasked with to do from designers. I’m waiting for these things to be addressed by elm-ui v2 maybe. Until that, CSS is a clear winner to me, especially with combination with Tailwind 
2 Likes