Stream: ideas

Topic: Unison


view this post on Zulip Johannes Maas (Dec 02 2021 at 18:56):

Are you aware of the Unison language? (A search of Zulip on shows that key word only in an introduction post.)

Functions are internally addressed by their content hash. You can attach metadata such as readable function names and actually even program textually, but that's a whole other thing. One of the most interesting consequences of the content-based function names is that when you change the function, its hash changes. That means you get a really interesting refactoring experience: https://www.unisonweb.org/docs/refactoring

I don't actually have anything specific from it that I would suggest for Roc. I just have a hunch that it might spark some interesting discussion here.

view this post on Zulip Lucas Rosa (Dec 02 2021 at 20:59):

yea I've played with unison

view this post on Zulip Lucas Rosa (Dec 02 2021 at 21:00):

it's super cool, their decision around that greatly affects existing workflows, which is cool but also means they need to create new tooling around code review and pull requests

view this post on Zulip Lucas Rosa (Dec 02 2021 at 21:03):

I would be worried that between apps/platforms and trying to convince people to use our editor that this kind of thing in roc would be too much friction

view this post on Zulip Lucas Rosa (Dec 02 2021 at 21:03):

but because we have our own editor we would be well positioned to do this exact thing with our editor, including having UI for collaboration workflows with git

view this post on Zulip Takeshi Sato (Dec 03 2021 at 13:40):

Similar opinion.
I knew unison before roc and I am interested with both.
They shares some points but their goals probably may be different.

Unison devs might be focusing on distributed programming and introduction of brand-new ways of programming, which are not only interesting but also unfamiliar. Very different from current developers' heuristics (sqlite3 codebase etc).
Considering there are highly sophisticated things behind the scene, unison's devs may be very careful that their documentations look friendly. docs are not typical but intriguing, too.
I am wondering if people accept unison's ways but, at least, I think unison will have some influence on future languages, as Elm gave some inspirations to modern js frameworks.

Compared to unison, roc's position may be similar to goalng's position. (though it is of course imperative)
Go is not fastest or featureful... but it is not bad. simple and reasonably fast. Some devs demanded such a lang.
It seems to me that roc devs are focusing on compilation speed, language perfs, or tooling (editor).
These priorities look similar to go to me.
(Honestly, I have not played with roc's editor yet)

view this post on Zulip Zeljko Nesic (Dec 03 2021 at 13:58):

Roc Editor lives in everybody's heads much more than in actual code. So the comparison in not fruitful, we are totally open of taking in ideas and inspiration. We have some bold ideas what the final version of the editor might look like, but please if you have share your ideas of what you think the editor might do and how it might work.


Last updated: Jun 16 2026 at 16:19 UTC