fix: logic

Author: wade-cheng

src/blog/optimizing-compile-times/index.typ

@@ -7,7 +7,7 @@
 
 #smallcaps[This blog] is generated with Typst using #link("https://wade-cheng.com/compile-typst-site/")[`compile-typst-site`]. It's written in Rust.
 
-Programming language culture is an interesting thing: while there's nothing technically stopping the average Python code from being duck typed and the average Rust code from preferring invariants via typing, this happens anyways. CLIs, like `cts`, and the simple case study of parsing command-line arguments gives some good examples. In Python, I might default to using stdlib's `argparse` and grab the arguments by getter properties: `if cli.verbose: foo()`---less explicit stuff going on. But in Rust, and in all the tutorial information around CLI building, we seem to prefer using proc-macros to generate code that validates the args into a CLI struct. Something something different philosophies of the average case. I'm sure there's derive validators in python, and you can forgo derive in Rust.
+Programming language culture is an interesting thing: while there's nothing technically forcing the average Python code to be duck typed and the average Rust code to prefer invariants via typing, this happens anyways. CLIs, like `cts`, and the simple case study of parsing command-line arguments gives some good examples. In Python, I might default to using stdlib's `argparse` and grab the arguments by getter properties: `if cli.verbose: foo()`---less explicit stuff going on. But in Rust, and in all the tutorial information around CLI building, we seem to prefer using proc-macros to generate code that validates the args into a CLI struct. Something something different philosophies of the average case. I'm sure there's derive validators in python, and you can forgo derive in Rust.
 
 But. One pet peeve I have about Rust, coming from interpreted languages like Python and simpler codebases from coursework, is the relatively long compile time compared to these prior experiences. In the more obnoxious cases, this even bleeds over `cargo check` efficiency, making the Rust LSP less effective.#footnote[Subframe LSP responses are great for coding. Stuff like bevy projects or #link("https://graphite.rs/") contributions can get a little absurd, again, relative to what I'm used to. Mumble mumble metaprogramming is cool but dangerous.] Now, this isn't happening by any means to `cts`, and there are tradeoffs to consider instead of wanton compile-time optimization: massive, slow, libraries are such because they're so feature-filled. It might be strictly easier for me and better for the users to use one library, except for compile times. So let's see what I managed to cull down.