next-gen maybe?

seems to be… otoh, if it’s a prefix, then probably Angular

(don’t wanna touch angular with a three foot pole)

bad experiences?

I've never got round to using it but often wondered...

Imagine spending lots of time figuring out the principles of Angular v1… and just when you finally figured it out, Angular v2 is out…

that's the story of my entire computing career

the next job always requires you to learn a new stack

that’s… sad 🤧

but then, JS ecosystem churn isn’t news

is it? I thought that was normal and healthy

definitely not at the rate of churn in the period 2010-2017

I’m also one of the unfortunate ones who has deal with angular. Not by choice, but we have to modify some software written using the (old) Angular that we run on our servers.

what are the problems with it?

we signed up to be developer, not superheros who doesn’t need to sleep

(aside: I latter learned that heroic programming is an anti-pattern)

the dark side of Ego driven programming

actually its pretty much all dark side isn't it

very true

caring about the quality of your work thats that upside of EDD

you mean problems with Angular? It may have improved since then. But the legacy of two-way bindings hurt a lots

(for v1) Also having its own dependency injection framework, mix in some module system before the JS community settled down on the modern solution

in short, it feels like they take complexity as a virtue and not a necessary evil

or they just don’t care because they can just throw an army of developer as if the reserve is infinite

or they started writing the project before a consensus had been reached, and did their own thing

ya, starting too early is part of the problem too

anyway, in JS no one is completely immune

> in short, it feels like they take complexity as a virtue and not a necessary evil you wonder if those bonkers Google interview question are selecting for this

wait… why are the interview’s “difficulty level” heating up like that in the first place? :thinking_face:

wow… that’s some SICP level material right there…

SickP 😃

I didn't know you could directly instantiate a MapEntry in cljs

I guess in Clojure you can (clojure.lang.MapEntry/create :key :val). Never looked into that

oh and MapEntry extends AMapEntry which extends APersitentVector. I knew map entries were very vector-like but I didn't know they literally are vectors.

(conj (clojure.lang.MapEntry/create :key :val) :x)
;; => [:key :val :x]

Just so I understand, a MapEntry is a tuple of [key val] implemented as a regular Clojure vector?

Yes, a mapentry is a value type containing a single key-value, map implantations store these mapentries. It's what you get back when you iterate over a map or when you use find. It's what allows you to use clojure.core/key and value. It behaves it many ways like a two element vector, hence why you can destructure it, eg (for [[k v] m] ...)

What I didn't realize is that the implementation shares the same base class as other vector implementations

So cool with these insights on the implementations you share, @plexus! Thanks! 🙏 I just watched your zippers talk and had several wow moments and almost physically felt things go click in my brain. I hope some of it sticks!

I read one description of Monads that gave me the impression that it’s really “just” like an interface like you know from the Java OO sense. I wonder why I haven’t seen that comparison before. Maybe I’m way off target, though. I’ve never gotten around to programming explicitly with Monads yet.

kind of, but there’s also the Monad law - any custom Monad you define is required to satisfy those law. Breaking them can have consequences as Haskell can auto-derive some other functions from the one you supply, but it assume that the laws hold up. In case you’re wondering, it won’t/can’t check the law through the type inference engine as those law are algebraic identity at the level of values and not types. Not sure if dependent typing can do this though.

I forgot where I read it from, but Monad is causing so much grief in Haskell for learner, that there’re ppl reflecting on why and what can be done to improve. A tentative conclusion is that it is an unlucky convergence of: 1. Monad being the first “sudden barrier” that requires the simultaneous mastery of several novel skillsets, each of which is already not easy in itself, and 2. Monad being a huge gateway/chokepoint - although Monad is “hard”, it is already “easy” from pov of advanced Haskell. Clearing Monad will open a whole world of intermediate+ topics in Haskell.

I think the “Monad law” that you mention is what I meant by interface equivalent, if not in implementation, then in spirit.

But I’m out of my depth 🙂

The bit I’ve heard sounds innocent enough. “We’ve established a convention that is exactly like this, and it affords us a lot of nice stuff”.

Of course, it’s more of a mouthful if you’re already busy learning the syntax and conventions of a new language, runtime and ecosystem (like Haskell)

the monad laws are constraints (in addition to those in the monad type-class) which can't be captured by (at least) haskell's type system i've got no idea whether it's even possible (or known to be impossible) to capture those constraints in any static type system, but "interface in spirit" works as a description for me

So a monad is, as you say, an interface and a set of laws. Or you could call them invariants. Or you could simply say that there could be a sort of test suite that certifies a thing as a monad. Then of course there is the mental overhead of the practicality of such a general abstraction.

tbh I'm surprised to learn that it's not enforced by Haskell's type system. I guess I just assumed they'd gel because of how I've heard the type system praised.

It’s interesting with the static typing people. Whatever invariants they cannot express in their type system they call laws.

Huh!

As an aside, the last time I tried coding in a static language, I had a terrible time. OTOH I often get bit by the open endedness of Clojure (bites sometimes self inflicted)

I feel like I'm sometimes missing an ergonomic language construct for closed enumerations, for one thing. But I could construct it myself; I just haven't arrived at the shape of what I'm dreaming about just yet.

'case' is often what I want.

All code, including macros, is equally suitable for deletion 😛

I wasn't talking about the macro, but about the expansion. A tiny macro can expand in a huge wall of code ;)

Yes yes, I know, I was being facetious. Sorry:sweat_smile:

this remind me of something like the zip file bomb (and in my high school’s computer club they teach us fork bomb)

a bit off-topic, there’s also that time I was tasked with updating some excel file. I didn’t know that excel can be used as a poor man’s database, and just updated a formula ~10 minutes before end of office hour… my workstation locked up… not fun 😖