burdges
commented
7 years ago I think rewriting in Rust adds an important selling points when you already have important new ideas and advantages, but not enough advantages to build a strong community around a whole new project.
Pijul is a good example : Very good ideas, but darcs and git already mostly covered the DVCS spectrum. I think pijul's hybrid approach sounds good for applications that do DVCS functionality under the hood, so being a library that can be called anywhere helps.
There is far more reason to join a community around pijul now that it can be called form anything, and is both fast and memory safe.
Imho, you should conversely accumulate good ideas before doing a major project in Rust. It'd be wasteful to simply rewrite say glib and gtk in rust. If you want to write a GUI in Rust then you need a stable of good ideas about how to do it better.
Arguably, Servo has inadvertently done this because they do not support aspects of HTML that they cannot render in parallel. We should probably consider deprecating HTML features that Servo cannot render well. Any HTML that we cannot parallelize makes deploying consumer systems with more cheaper CPUs harder, which makes computers more expensive, makes power management harder, etc.
In this vein, I'd love to see CloudABI inspired capability-base alternative to rust's std because writing for capabilities makes your code much more secure.
It should not be "Rewrite it in Rust" but "Write it much better in Rust".
Profpatsch
https://pijul.org/2016/02/02/pijul-0.2.html
Not to say it has been rewritten because the Scala version never actually went anywhere (it was too slow in fact). So maybe an actual good use case for the language for once and not just hype-of-the-year.