I dunno a lot of migrations to X technology are pitched as valuable on one quality dimension. We improved performance by Y%! When software “quality” should really consider many criteria: correctness, performance, developer productivity, maintainability and others.
I’d rather hear about a set of tradeoffs in these sorts of articles (performance was critical so we traded Y for X) than just bragging about one dimension of improvement.
I wonder how long their build time is. The article from the “Things” folks said their system took 10 minutes to build despite being only 30k LoC, which seems really bad. I can’t help but think they must have done something off for it to take that long.
Hence, I wonder how long it take a to build what is probably a medium size system when done by a team at Apple.
Big discussion on the original source (243 points, 12 days ago, 228 points) https://news.ycombinator.com/item?id=44172166
TLDR: Native language wins over JVM optimisations.
In my recent but limited experience of working on iOS apps, they do need any performance boost they can get anywhere. Coming to the Apple ecosystem (appstore connect, testflight, xcode etc) it's quite a shock on how slow it all is (using xcode on a Mac Mini M4 for development, and as for web updates/approvals etc I now understand the pain others have mentioned here).
Is swift really appropriate for performance critical server side code?
Why on earth did everyone start using “uplift”
It’s a migration from Java to Swift. For some reason that key information was missing from the title.