[tialaramex]

Hylo is interesting in principle for exploring this particular notion (mutable value semantics) as a way to potentially write software without the lifetime annotations Rust needs.

But I don't find it promising that after apologising in 2023 for missing their self-imposed 2022 deadline to ship something that works and other people can use, in Q2 2024 it doesn't look like their new 2023 roadmap got done either. Maybe they're going to eventually deliver this amazing thing. Maybe they're just going to learn some lessons (probably for the Swift community) and never ship Hylo per se. Certainly 2025 "Take over the world" looks... ambitious with nine months left to do all the stuff left from 2023 and all the work described for 2024 on top.

[germandiago]

For me Hylo is the most elegant object model I have seen so far in the sense that it achieves dafety without all the anniyations and borrow checks that Rust performs. It just fits well the model for thinking yet it enables lots of optimizations. Of particular mention is that copies are lazy and can be passed freely and will only be done transparently on demand. That is what keeps the model simple.