On Tue, Feb 18, 2014 at 05:44:36PM +0100, Tomasz Rola wrote: > On Mon, 10 Feb 2014, Yuhao Dong wrote: > > > Hi, > > > > I'm trying to decide between Racket and Go on writing my onion-routing > > system inspired by Tor. Basically, a network server, involving lots of > > long-lived connections that often pass large amounts of data. The thing > > needs to be super scalable; I often find that these servers, although > > network servers, often become CPU-bound doing encryption and > > encapsulation of protocols, so I do have experience that this is not > > "premature optimization"! > > > > [... many faults of Go-choice ...] > [... many faults of Racket-choice ...] > > > Are there solutions to these problems? These aren't showstoppers by any > > means, but could finally end my endless dilemma between the two langs :) > > For the reasons you wrote above, I wouldn't choose any of them. > > If you know C and or C++ very well, you could do this project in them. > However, since this is networked code, there is a lot of stuff to > remember, or you may later find many embarrasing security holes in your > code. Besides, using certain C++ constructs (incorrectly or maybe even > correctly sometimes) may lower your performance a lot. > > Of other languages mentioned by others, Erlang is perhaps a good choice > but I don't know it at all, so I cannot comment. Ada doesn't look bad but > I understand you may have either to go opensource or pay for > Windows-capable compiler. In exchange you can proudly use "military-grade" > language :-). > > Myself, I would also consider OCaml, mostly because I learn it myself and > it may be the right middle ground - good compiler, high level, can do > prototypes (AFAIK). I'm not sure how good it is wrt security holes in > generated code. > > Java is out of the picture for me, because I know it well enough to > dislike. Python is out, because while it looks good, I want to replace it > with OCaml, so I wouldn't have a reason to write excess code. Haskell is > out, because it feels too much like putting strangely shaped shoe on > straight foot (and is not easy to optimize). > > All of the above is just MHO, so if you feel otherwise, don't bother about > me, I am very subjective when it comes to choosing a language.
I'd stick one other language into the mix -- Modula 3. It's an elegant systems language its main flaw it its Pascal-like syntax. See its Wikipedia page for details and links. Ask questions on the developer's mailing list. The implementation team is currently getting it to run multicore; it already runs single-core with software-managed multithreading. But it does have a stop-the-world garbage collector. That's only a problem if you use the garbage-collected heap and have tught real-time constraints. There's also a non-collected heap (explicit allocatino and freeing) in case garbage collection is too rich. I think there are other non-default garbage collectors, but I'm really not sure. It has a new code generator that generates C code in case you need ultra-portability. And, just in case there's confusion, it is *not* *at* *all* the same language as Modula or Modula 2. By the way, the languages I use if I want reliability and early bug detection are Modula 3 and OCaml. I use Scheme occasionally when I want to play with coding notation a lot. -- hendrik > > Regards, > Tomasz Rola > > -- > ** A C programmer asked whether computer had Buddha's nature. ** > ** As the answer, master did "rm -rif" on the programmer's home ** > ** directory. And then the C programmer became enlightened... ** > ** ** > ** Tomasz Rola mailto:tomasz_r...@bigfoot.com ** > ____________________ > Racket Users list: > http://lists.racket-lang.org/users ____________________ Racket Users list: http://lists.racket-lang.org/users
