> In the context of the OS, multithreading (or something essentially equivalent) is the only way to exploit the abilities of "modern" machines.
I noticed that you qualified multithreading with "or something essentially equivalent". I guess you would agree that qualifying it this way is a good idea. Mr Blandy did not care to do that though: "[...] is the only way to exploit the abilities of modern machines."
I don't like that because it sounds like he's trying to popularize one approach to concurrency without even acknowledging that it is just one approach.
Note as well that Rust intends to have world-class support for a wide variety of approaches to concurrency and parallelism. Threading and channels are already in the stdlib, and fork/join and SIMD are in the works now. The goal is to enable the programmer to be able to choose the best tool for the job.
If you don't want to come across that way, give us some explanation of what you're thinking and why. Give us some evidence that your statement is true. Something besides just dogmatic assertions that you're right and the author of the article is stupid and/or biased.
Moving on...
In the context of the OS, multithreading (or something essentially equivalent) is the only way to exploit the abilities of "modern" machines. But "modern" doesn't mean very modern (at least, as I understand it). It means the difference between MSDOS and Windows: "Windows is multitasking, whereas DOS... DOS is serially multitasking." (I don't recall who said that, but it was brilliant.) Without this capability, you're only running one program at a time.
Now, I don't care if you implement this as "theads" or something else, but if you don't have it, your OS is pretty much worthless, and has been since about 1990.