I’ve used CUE a bit and I’m underwhelmed by it. In most cases it’s just another unneeded level of abstraction, it also doesn’t go far enough, most API based tools need to support imperative methods
You don't want computability in your config, and CUE is useful beyond APIs.
You can already use CUE within other programs when you need imperative, there is also the scripting layer where this is possible. There is also a plan to support a WASM runtime so you can have imperative subroutines in the scripting layer, written in any other language.
My problem is generally the config is a part of a larger data model, which often needs to define imperative methods on resources. Having stuff in Cue means yet another layer and language I have to learn and use for this part of my API, and then copy that into a full API language.
How are you defining that model and methods today? Are they already in the tech / lang you use or are you using some translation like protobuf or openapi?
(curiosity as I'm building a tool on top of CUE that combines it with notions of data models and code gen, and I'm always keen to learn more about the things people do today)
yes, while it is gaining adoption, there are still a couple of major tasks before it really takes off, namely dependency management and structural sharing.
Many heavy hitting tech companies have developers using and exploring it, mainly because it augments rather than replaces. It's very easy to add validation while not changing anything else. Lots of users around Kubernetes applications.
Didn't realize people still used Wikipedia for things besides for "facts" since its descent into the political / opinion battlegrounds. It's been years since I used it
Again, who's actually using it? All I see is Salesforce and two no name companies. It's supported by Netlify which is something, but I wouldn't say "hard hitting companies". Netflix and Amazon are hard hitting.
If it doesn't have a Wikipedia article, there's a very good chance it isn't very popular in the industry. At least in terms of open source standard technology, which this appears to be reaching for.
Google, Amazon, Tesla, Tencent, Alibaba, ByteDance are in the list, not all are on a website. It's not like a company mandate, mainly individual contributors checking it out and building things.
It's not widespread because it is just starting to get going, but it is definitely popular (well liked) among its users.
It's about as necessary as your comment stating facts about the future. :)
In any case, I apologise - my comment was indeed unnecessary and didn't bring any value to the discussion. Had it been made to me I might not have been as good a sport as you have so kudos to you!
