zaksingh

I've been working on a new programming language and I've started to think that the design of a language's standard library (in terms of error handling, consistent naming, consistent argument ordering, types, etc) has a greater impact on the learning curve and ergonomics of a language than the more commonly discussed syntax differences and unique features.

For example, I can _never_ remember if a function in JS or Python mutates a list/object or returns a new copy. Without editor hints this was an annoyance, as I'd have to try it out, check the source, or search the docs. Such issues never happen with a functional standard library (as in Clojure) where you can always assume immutability.

Meanwhile in Clojure, if I've been gone from the language for a while I forget which list functions return a lazy sequence, a seq, or a vector, and I find myself making frustrating errors as a result. Or how some functions (like map and filter) take the collection as the last argument, and a few others take it as the first.

It seems like a poor standard library design requires acquiring much more 'trivia' about the language to be successful, to write performant code, and (sadly) to pass many technical interviews.

What do you think makes a good standard library? A bad one? Do you have any favorite features you'd like to see more of?

I'm working on a tool that requires running my customers' APIs inside of self-contained environments. They can create/destroy these environments at any time. Is there a 'best practice' method for this situation? Some googling turned up very little (maybe I don't know the term for this type of thing), so I thought I'd raise the question here instead.

My gut instinct is to run each inside of its own Docker container, perhaps in a Kubernetes cluster. Is this safe?

Our small startup has managed to gain traction in the US, UK, and India. All 3 regions are desirable. Currently we are hosting a GraphQL API on Heroku and our DB and some other services (an Ejabberd server) on GCP. These are all in the US.

Users in the UK and India complain that our app is very slow. I can replicate this by running through a VPN in those regions. In the US the performance is fine.

So, my question is: how do people handle this? I could host API server instances in Europe / India, but shared resources like the database can only live in one place. All of our app features are dynamic and therefore a CDN wouldn't do much.

This feels like it would be a common problem to have these days, especially in the mobile app world. It's such an obvious problem that I feel as if I'm asking a stupid question.

How do you guys approach scaling your backend globally?

It's been "the future" for 40-ish years now, with dozens (if not hundreds) of failed attempts. At some point, one has to wonder: are we there yet? If not, why aren't we?

To me it seems logical that designers should one day be building frontends through visual interfaces that are almost as malleable as tools like Sketch. It seems to me that many features from design programs (like Framer) seek to make design closer to development, and we've even seen a Git for design workflows. So why hasn't it happened yet, and will it ever?

I've bounced between paper notebooks, Apple's notes app, and have recently begun looking at apps designed specifically for journalling and/or mood tracking. Is there a better alternative? Is it even worth the time investment?