It's a question of where to focus your attention. Consider the explore/exploit trade-off. Boring tech may be a little, well, boring, but that means you get to spend your limited decision-making capacity on solving your actual problem. If your actual problem is betting on a niche language early, then go for it! Otherwise, use the thing that balances your existing knowledge, interest, existing 3rd-party resources and has the right outlook over the expected timeline to get your outcome.
The web has killed desktop apps. I use linux, but I don't buy software on any platform besides games.
I think it makes sense for professional productivity apps, like CAD stuff.
I don't think basic desktop apps are worth a fee, since no one will buy them with all the free alternatives, but maybe a case can be made with something that ties into managed cloud compute services (like superhuman's use of AI).
There is an alternative that I have found quite useful.
Create a Dualboot situation for normal linux work but set up a VM in your windows host that points to your bootloader and linux partition.
Now you have the best of both worlds, more memory and speed when you want it with native boot, but access to the dev environment from windows.
This was most effective in the case where a client required proprietary windows-only VPN software, but I wanted my dev environment. Configuring a vpn'd bridge adapter between Virtualbox+linux-guest and Windows host was a piece of cake.
I've been looking into this setup as well, having dual booted all my machines previously. Though rebooting takes only a few seconds it feels like more work than it actually is, complicate that with having to enter encryption passwords, reopen all your software and load up files, etc, I think keeping everything under one OS and running a VM and then being able to boot into that VM natively when necessary seems like the best modern solution for my needs
Can you provide any guidance on how to do this? Can it be done when the Linux partition is on a non-NTFS filesystem? It sounds extremely convenient, especially as I've already got a dual-boot setup.
I'd skip Objective-C and learn Swift to actually make a thing.
If you want general language knowledge, a really hairy production language and toolset isn't the place to look. You'll be fighting with lots of incidental stuff along the way.
Learning different kinds of languages will help you learn more languages.
I really like this comment. But I bet some folks gloss over a particular point of importance.
For me, there's a huge advantage in literally using tree-paper, either for code-reading, designing, taking notes, what have you.
It involves more of your senses, it's more tangible, every letter is written with more care and the medium is incredibly flexible (maybe I feel like drawing a circle on a whim, can't do that in emacs.)
It's better for scratch-pads than a text-buffer, imo.
Insights come more consistently and easily after iteratively marking up a printout. It's just fully-intuitive.
Better stacktraces would be better, sure, but there are good arguments on both sides whether it's better to separate the language from the platform or not.
I prefer a thin, readable compiler and core library over such abstractions. I don't mind the stacktraces, but I could see how some people might.
The fact that clojure is host-centric makes clojure and clojurescript compelling to java or javascript programmers at the expense of some aesthetics, and it also has performance implications.
I think this site leans pretty niche.