3 ms·
> Rust's safety w/out performance hit affects those two, specific areas. I think they might be interested. :) Let me put it this way: if for some reason Rust e
by pron 10y ago
> Rust's safety w/out performance hit affects those two, specific areas. I think they might be interested.
:) Let me put it this way: if for some reason Rust ever takes a serious market share from Java (or other similar languages) in the enterprise space, I will surely be well into my retirement by then, so my interest isn't financial. I was a C++ developer for many years, working on very large soft (and some hard) realtime systems, and I was a Java holdout, but once most of the defense industry switched, virtually nobody (including us) ever complained about a performance decrease. So even if the claim that GCs adversely affect performance in large, complex programs (where RAM overhead isn't an issue) were true, the number of organizations that build software of that kind and where this performance would matter more than in defense, is very, very small (not to mention that most of the current performance deficiencies in Java are not related to GC). If that's the kind of user that would be a significant portion of Rust's userbase, then Rust is in trouble. Companies that sell ultra-low latency GCs for a fraction of the cost it would take for enterprise shops to adopt a language like Rust, are, well not exactly on their way to the Fortune 500. The systems software space -- drivers, kernels, filesystems, and the entire embedded space -- is at least one, if not two, orders of magnitude bigger.
- nickpsecurity 10y agoI wasn't saying you were shilling to dodge competition so much as better performance is in your marketing material like many others. ;) I agree that a lot of the space they should be targeting won't have a humongous difference between a well-tuned GC and a typical, native implementation. You're right on that. "Companies that sell ultra-low latency GCs for a fraction of the cost it would take for enterprise shops to adopt a language like Rust, are, well not exactly on their way to the Fortune 500" That's a good point. Switching costs would be huge in many of these organizations. They'll make better inroads with something within their existing stack (typical) or doing new projects in the improved language (also typical). They're not rewriting all that stuff, though.
- jjnoakes 10y agoWith rust you don't need to rewrite though. Just integrate it a little at a time.
- nickpsecurity 10y agoHopefully. With the legacy codebases or preferences, that means hiring people that really know C or C++, training them for Rust, and maintaining two codebases in one. It becomes trickier. Hopefully the incremental option works well for Rust but it didn't with most languages & platforms.