Today’s JavaScript, from an outsider’s perspective (2020)

I think the problem here is that there's so many tutorials, articles etc telling people that if they want to do data analysis then the best way is to use Python without making it clear that this advice is predicated on you already having a programming education. If you start with a CSV file and want to do some analysis, without knowing programming, then you need to using a spreadsheet app, but this context isn't being reliably given to people. Instead they absorb a background assumption that "real" data analysis is based on programming and immediately drown when they try it.

> Programming is a profession requiring learning from experience - if a novice doesn't have the patience to overcome very small speedbumps like this on a daily basis while learning a new domain, this might not be a good path for them.

There's a lot of assumption packed into this statement.

Those speedbumps used to also not be there—they were put in recently[1].

Also, having read through your comment first and then read through the article (fully expecting to find what you describe), I'm compelled to say that your summaries grossly mischaracterize the substance of the complaints actually mentioned by the author.

> a painful but necessary migration incompatibility is an unfortunate reality, the only alternative of which is to stagnate and never introduce/deprecate breaking features

Well, no. The module fiasco is 100% a consequence of really, really bad decisions on the part of the NodeJS folks. They created the preconditions that led to the demand for their solution. This problem/solution cycle is rife in the NodeJS/NPM world. The only way to deal with it is to opt out[1][2].

1. https://www.colbyrussell.com/2019/03/06/how-to-displace-java...>

2. https://news.ycombinator.com/item?id=24495646>

Is node.js giving the author bullshit or did they knowingly step into a field they are unfamiliar with, getting themselves involved with a package manager, a javascript runtime, a browser, and a local fileserver?

As if you were able to instantly run whatever code you found on the internet back in the C++ days. It's bizarre. Not everything needs to have such a low barrier to entry that anyone can work with it. Who said _that_ is where we collectively set the bar? It's a false premise, and frankly it feels like toxic false entitlement open source maintainers are typically showered with.

The problems are all eminently avoidable, though. It's all death by a thousand cuts stuff.

1. Churn in something as basic as how modules work isn't common to every programming language, actually that's pretty specific to JavaScript. Java modules remained the same for 20 years, and when the 'new' module system came it's opt-in (you can pretty much ignore them and many people do because a 'new' module is forwards compatible and can be treated as an 'old' module as well).

2. They were only trying to use the browser because they had problems with Node and thought it'd be easier. The fact that there are security issues loading stuff written by the owner of the computer itself from disk is more an artifact of how browsers work and how they evolved than anything else.

3. The usability issue wasn't really with VS Code. It's that the default behavior of browsers on encountering stuff they don't understand is to ignore it and plough on regardless, so they got no error message and at that point were stuck. If they hadn't had Lee on hand their adventure would have ended here because Lee simply intuited what was wrong based on past experience. The fact that browsers routinely leave you looking at a broken page with either no error messages to help you, or a uselessly vague one, is a problem caused by (again) how browsers evolved, it isn't fundamental. More robust development environments don't have this issue. If something goes wrong you get an error, always, and it's usually helpful. You never just get nothing and are left wondering what happened.

4. The "bug" here is that you can't use TypeScript directly in a browser (why not? again, mostly complicated browser related tech politics that no beginner would understand), and the library in question doesn't advertise that it's using TypeScript properly. Even if the bug was fixed the user would still be stuck with TypeScript when what they need is JavaScript.

"Programming is a profession requiring learning from experience - if a novice doesn't have the patience to overcome very small speedbumps like this on a daily basis while learning a new domain, this might not be a good path for them."

The guy wasn't a novice. He's introduced as " a great computer scientist, but not a JS person".

I think we're, as an industry, exceptionally blasé about usability problems in our tooling. These problems don't just waste the time of novices, they waste the time of people with 30 years experience, people who have been coding since childhood. The problems encountered in the article aren't the classic beginner problems like forgetting to add a semicolon at the end of a line. They're the result of accumulated lack of care for the overall developer experience. The root cause is a rampant culture of abusing tools for things they weren't ever actually designed for. JS modules are weird because JS was never intended to be a programming language for large scale projects, so it took decades for modules to emerge and then they were bolted on. Browsers refuse to do something as basic as load code from local disk because they were originally designed to render documents, TypeScript exists and requires extra work to execute because of the spec-driven approach browsers etc which in turn is due to their origin in rendering documents, etc etc.

The blog post is specific to JavaScript because these problems are specific to JavaScript and the way we (ab)use web tech.

X is not hard, you just have to Y.

Be careful of these statements, you have to understand why a person says X is hard. The OP stated he did extensive work trying to get it to work and found it difficult.

You posted one link out of thousands. When talking about these things - its not like there's a problem with one solution.

There's one problem with thousands of answers that are frustratingly close to correct, but not correct.

That's why X is hard.