Go vs. Rust: Productivity vs. Performance (2014)

Taking an average of wildly varying microbenchmark scores doesn&#x27;t seem particularly useful. Furthermore Rust has never had particularly optimal implementations for the benchmarks game (I can&#x27;t speak as to the quality of the Go implementations). Furtherfurthermore this blog post is so old that the Rust results cannot be considered representative. Furtherfurtherfurthermore I would like it if we could just all collectively stop treating the benchmarks game as though any of its microbenchmarks measure anything other than &quot;is this language capable of binding to libpcre&quot; or &quot;is this language capable of binding to libgmp&quot;. I swear, the benchmarks game is the SunSpider of programming language comparisons.<p>(Furtherfurtherfutherfurthermore, why are we still comparing Rust to Go? I thought we had gotten over this!)

&gt; Go seems to be for the Python, Ruby and Java developers and will be used for enterprise applications, mobile apps and application servers.<p>No thanks, Go offers no benefit over the Java eco-system, neither in terms of deployment[0][1][2], not in terms of libraries and dependency management, nor in language expressiveness (Java 8, Scala, Clojure, Kotlin, Ceylon, Groovy)<p>[0] <a href="http:&#x2F;&#x2F;www.excelsiorjet.com&#x2F;" rel="nofollow">http:&#x2F;&#x2F;www.excelsiorjet.com&#x2F;</a><p>[1] <a href="http:&#x2F;&#x2F;robovm.com&#x2F;" rel="nofollow">http:&#x2F;&#x2F;robovm.com&#x2F;</a><p>[2] <a href="http:&#x2F;&#x2F;www-01.ibm.com&#x2F;support&#x2F;knowledgecenter&#x2F;api&#x2F;content&#x2F;SSYKE2_6.0.0&#x2F;com.ibm.java.doc.60_26&#x2F;vm626&#x2F;GenericWrapper&#x2F;commandlineoptions_jit.html?locale=en" rel="nofollow">http:&#x2F;&#x2F;www-01.ibm.com&#x2F;support&#x2F;knowledgecenter&#x2F;api&#x2F;content&#x2F;SS...</a><p>EDIT: Also forgot to add that we enterprise devs also have C#, F#, Swift, C++14, C++&#x2F;CX at our disposal

This article doesn&#x27;t really match my experiences.<p>I have found Rust to be considerably more expressive than Go. Traits and proper generic types, iterators and pattern matching all contribute to this. I tend to write less lines of Rust code to accomplish the same tasks.<p>As for runtime performance, in my experience they are pretty close right now for my use cases but I do expect Rust to grow faster than Go over time.

The article has a major flaw: it confuses productivity with LoC.

I think Rust is a fledgling this time. But I love the clearly orientation that they focus on. It is a safety system language, especially, the ownership&#x2F;borrowing.

&quot;Go seems to be for the Python, Ruby and Java developers and will be used for enterprise applications, mobile apps and application servers.&quot;<p>Out of interest, what would I gain by learning GO? I already know Python pretty well. I have done Java in the past and don&#x27;t see any problem jumping back into it. Both languages have tons of libraries and tooling available. Are mature, and I can&#x27;t see much reason to opt for a new language.

&gt; <i>I think Rust will attract developers from C, C++, Fortran and will be used for developing high performance systems like gaming, browsers, telco servers, distributed computing systems as well as low level, cpu efficient embedded&#x2F;micro computers.</i><p>People choose platforms, stacks, languages, environments, not languages: Rails not Ruby, browser not JavaScript, Unix not C, host not Cobol, ...

they don&#x27;t exist in the same space. Go is not a systems programming language and Rust is not a scripting language.

Despite being always interested in seeing benchmarks comparing new languages I don&#x27;t feel like basing my choices on such benchmarks. From my experience it&#x27;s usually a matter of how comprehensive is the ecosystem surrounding the language. Anyway, great article.

Great article, and I have to say I completely agree with your point on using go for enterprise level development.

Go is a simpler language and as such it is a better language.

Rust&#x27;s not there yet and I wonder if it can compete. The thing is it is too specialized and low level (not that it is a bad thing for what it is intended for). The problem with all of the low-level things is that they are seriously undervalued.<p>I am glad that there is servo but we need something like an operating system (IoT might work for the start). I hope Samsung changes its plans with Tizen and eventually will develop something with Rust in mind.