Being safe on the internet (was Re: Here we go again - ISP DPI, but is it interception?)

[PlusNet]

Sent from my iPad

On 7 Aug 2010, at 17:15, Ian Mason <ukcrypto at sourcetagged.ian.co.uk> wrote:

> 
> On 7 Aug 2010, at 14:50, M J D Brown wrote:
> 
>> 
>> One method that has its adherents is to employ a widely-used language
>> and compiler (on the basis that compiler faults may well have been
>> exposed in the course of widespread use),
> 
> About 5 years ago I hit a bug in the gcc C compiler that mis-codes certain 64 bit arithmetic operations when compiling in 32 bit mode (i.e. 'long long int' handling where 'long int' is 32 bits), When compiling the latest Asterisk beta the other day I hit it again. I can't think of a more widely distributed and used C compiler.
> 
> Talking of safe languages no one has mentioned my old personal favourite Algol 68. it was even used as the systems programming language for a capability architecture processor and OS by Maurice Wilkes, the late Roger Needham, and others. Cue Dr. Gladman...
> 
Charles Lindsey could say a lot more than I could about Algol 68.  I did spend a fair amount of time while I was at the Royal Signals and Radar Establishment in trying to persuade Peter Gershon (when he was with ICL) to implement it on their then new range of machines. We also attempted to get the US Dept of Defense to adopt it but they decided that they needed a new language, which emerged as Ada.   

Aesthetically I much prefer Algol 68 over Ada but it's not a practical choice these days. In contrast, although few realise it, Ada is still widely used where either safety or security are critical.  

I have multiple precision libraries now written in C but I built them in Ada and then translated them into C.  I trust them a great deal more for my crypto work than earlier ones that i wrote from scratch in C.

       Brian 
>