miso

e many C programs that are sort of correct but not easily written with standard, total, and sound types, plus the required assumptions are not expressed in the source code). Also, another angle I think might help to convey what I'm trying to say is to look at how much jargon is needed to explain things like for-comprehensions execute the monad steps in sequence, but if you use mapN and an applicative then you can execute independent steps in paralell. Whereas it's just "race the futures" (or Promise.all, or whatever) in common programming parlance. > Perhaps you've had this experience from experimenting with functional languages [...] > What language are you thinking of? Sounds like Scala maybe. Yes, I don't know Haskell, I have a few years of Scala with some Scalaz experience. And maybe you're familiar with mypy and TypeScript, and it's obvious how problematic is to build the bridge between existing semantics and a correct/sound type system. (And of course, it's very much impossible, because you either have soundness or not, and so TS is unsound to be useful, and mypy is useless unless you rewrite many things.) > [...] and I find functional languages to be models of elegance by comparison I haven'got the hang of any pure functional language (yet?). I like to think it's not because of lack of trying. I see where concepts can be useful (eg. how TS and Rust would benefit from HKT, and how many people are waiting for a more expressive type system to create better abstractions, better libraries, better APIs), but to me all Haskell code looks like something like a dadaist poem with strange symbols sprinkled here and there. (Eg I'm looking at the sample app for miso https://github.com/dmjio/miso#sample-application and I have no idea what is what, no idea about shapes, what's <#, what's >>, why div_ and button_ have underscores after them, why div_ takes an empty list, what's {..}? And sure, most of these can be found, but still a lot harder than even the most damned FactoryFactoryBean in Java. And sure, I appreciate that lenses and the Transition interface helps with stuff, but that's simply a very small incremental ergonomic enhancment compared to what I wish. Similarly, when I look at servant's documentation I wonder how come a such powerful language, but to model 'GET /user/:userid' it can't - or doesn't - use /user/:userid, but "user" :> Capture "userid" Integer. The type is nice, and whatever :> is, I'm sure that's nice too, but I'm not finding the elegance that TSed has - https://github.com/TypedProject/tsed#create-your-first-contr... ) So, all in all, I'm just a lazy party-pooper when it comes to Haskell (or FP in general). And mostly just because the pain of debugging TypeScript and the myriad "any"-s everywhere still seems the lesser one than trying to grok the the cryptic symbols.

rface with third-party libs. Also, Haskell supports typeclasses, and Generic programming, a rich lens library, etc. • Community involvement. You’ll be working directly at the intersection of a lot of interesting developments in GHC, specifically around cross compilation, syntax, lenses. Any new advancements to GHC has an impact on your frontend code. Also, many PhDs use, love and contribute to Miso. These people tend to be very smart, empathetic and intellectually curious as opposed to dogmatic and contentious that can be found in the software engineering industry at large today. • Nix, a one-tool-to-rule them all for dev, build, and deploy. While not easy, deterministic builds are the future and help one to sleep at night (I’ve found). It can be used solely for personal dev, or for an entire organization, and anything in between. • Virtual DOM. If your Haskell web framework does not handle the tedious task of DOM manipulation for you, you have been failed as a library consumer IMO. Miso has a very well-tested, simple recursive micro virtual DOM diffing library that can exist standalone and can detect regressions in performance via a benchmark suite. • Simplicity. Unlike many other ivory tower libraries in Haskell, Miso maintains a simple interface. Many different common abstractions were explored in the early stages of development (free monads, FRP, etc.) and found lacking, unnecessary or replaceable by a simpler solution without sacrificing features. Miso does not attempt to hide behind complex type-level abstractions and blame end-users for their lack of understanding. As opposed to a top-down approach that shoe-horns an abstraction into an environment where it does not belong, Miso takes a bottom-up approach to fully understand its environment, the browser, it’s APIs, the DOM and its short-comings and provide a simple commonly-used abstraction to the end-user that can scale to large applications. This also makes it easier for newcomers to Haskell to get up and running quickly. • Industry adoption. Several companies are using miso in production for various things in their company, from full product to internal tools / data visualization. • Isomorphic. Due to the rose tree structure of both the DOM and HTML, pre-rendering increases load-time and overall UX along with correct SEO. """ ^ from the author of Miso

face with third-party libs. Also, Haskell supports typeclasses, and Generic programming, a rich lens library, etc.) • Community involvement (You’ll be working directly at the intersection of a lot of interesting developments in GHC, specifically around cross compilation, syntax, lenses. Any new advancements to GHC has an impact on your frontend code. Also, many PhDs use, love and contribute to Miso. These people tend to be very smart, empathetic and intellectually curious as opposed to dogmatic and contentious that can be found in the software engineering industry at large today). • Nix, a one-tool-to-rule them all for dev, build, and deploy. While not easy, deterministic builds are the future and help one to sleep at night (I’ve found). It can be used solely for personal dev, or for an entire organization, and anything in between. • Virtual DOM. (If your Haskell web framework does not handle the tedious task of DOM manipulation for you, you have been failed as a library consumer IMO. Miso has a very well-tested, simple recursive micro virtual DOM diffing library that can exist standalone and can detect regressions in performance via a benchmark suite). • Simplicity. (Unlike many other ivory tower libraries in Haskell, Miso maintains a simple interface. Many different common abstractions were explored in the early stages of development (free monads, FRP, etc.) and found lacking, unnecessary or replaceable by a simpler solution without sacrificing features. Miso does not attempt to hide behind complex type-level abstractions and blame end-users for their lack of understanding. As opposed to a top-down approach that shoe-horns an abstraction into an environment where it does not belong, Miso takes a bottom-up approach to fully understand its environment, the browser, it’s APIs, the DOM and its short-comings and provide a simple commonly-used abstraction to the end-user that can scale to large applications. This also makes it easier for newcomers to Haskell to get up and running quickly). • Industry adoption. (Several companies are using miso in production for various things in their company, from full product to internal tools / data visualization). • Isomorphic. Due to the rose tree structure of both the DOM and HTML, pre-rendering increases load-time and overall UX along with correct SEO.