[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: TOR in Java?



> I'm not a TOR developer; however, as an exit node server operator
> and system administrator I prefer C and am willing to accept C++.

Why it is preferrable over Java or C# on code that isn't that
performance-critical like TOR?


> C#, being a closed-source proprietary product, could be very
> unpopular, ..

Eh, there's Mono!

> I view Java as Sun's property; ...

Yes, but Sun is very open to proposals an architectural discussions
as you can see from the JSR-processes.

> it's an effort to get closed-source code
> to work across hardware platforms.

Java isn't necessarily closed-source code - look for example at JBoss.

> It's extremely inefficient in terms of computing resources.

Yes and no (you have a huge basic footprint, but large apps like app
-servers this allotment shrinks in comparison to the rest of the code
and data of the app), but that wouldn't matter for a rather small app
like a TOR-server.

> Licensing is such that a lot of OSs don't come with Java by default.

Yes, but there are two Java-Implementations which you can get for free
(the Sun- and IBM-implementations).

> Java does have some merit for applets on WWW pages,
> but that's a separate topic.

Java is very good for server-apps. It's more secure than C (because
you can't do any buffer-overflows which cause control-path changes)
and it's fast for server apps; Apache Tomcat f.e. is faster than
the Apache webserver even for static content!!!

> Personally, I don't believe in buffer overflows; ...

Ok, it's rather a myth like the Santa Claus.

> my feeling is this never happens to competent programmers.

Look at what a lot of prominent open-source C-projects have buffer
-overflows. There are buffer-overflow-exploits for the SSLeay-lib
used by TOR, Apache and other apps every few months.

> No malice intended, and I know I'm very opinionated.

By whom??? *g*