[Cryptography] Subject: Re: Swift and cryptography

[Nemo]

An earlier reply of mine was rejected (go figure), so this time I will
try to make my points consisely and more-or-less impersonally. A couple
might even be on-topic.

Arnold Reinhold <agr at me.com> writes:

> Apple has 9 million registered developers (who pay $99/yr for the
> privilege),

Amazing what people will spend money on, isn't it?

Re: Dylan. The people, code, and ideas from Dylan have no relationship
to Swift. Browse to http://nondot.org/sabre/ and search for "drawing
ideas from" for initial evidence. Learn anything about the two languages
for additional evidence.

Re: Syntax. Programming language syntax is irrelevant and boring. (Also
the example I gave is quite easy to read for anyone with more than a
cursory knowledge of C++.) If your awesome new language's chief benefit
is its syntax, you are solving the wrong problem.

Re: Securely wiping data. This is a subtle problem, but also a solved
one. The solutions are platform-dependent -- e.g. OpenBSD has
"explicit_bzero" -- but they do exist. Examine the source for any decent
crypto library for evidence. Obviously, re-solving solved problems is a
waste of time. Indeed this is my basic point about Swift in general.

Swift appears to be little more than syntactic sugar over
C++/Objective-C, which is not surprising given its provenance. As far as
I can tell, every interesting feature is available in C++ -- especially
C++11 -- as a built-in feature (e.g. lambdas, shared_ptr) or an add-on
library (e.g. SafeInt).

Finally, and most on-topic: If the choice is between (1) a language with
an international standard and multiple competing world-class
implementations, and (2) a proprietary language with its associated
vendor lock-in (to a vendor known to collaborate with NSA), which should
we depend upon for our privacy?

 - Nemo
   https://self-evident.org/

P.S. I do admit to a secret fondness for Rust. Every cynic is a romantic
at heart.