Re: Heartbleed

[Simon Clubley <clubley@remove_me.eisner.decus.org-Earth.UFP>]

On 2014-04-17, Jeffrey Carter <spam.jrcarter.not@spam.not.acm.org> wrote:
> On 04/17/2014 03:35 PM, Simon Clubley wrote:
>>
>> If you think I am wrong, then, while telling me I am wrong :-), please
>> also tell me why Ada isn't in much more wider use than it currently is.
>
> Let's see, there's an Ada compiler that uses ANSI C as its intermediate 
> language. So there's an Ada compiler that targets every platform that C targets. 
> Clearly the number of targets that a language has a compiler for is not the issue.
>

Actually it is.

Two points:

(1) Will that compiler itself run on all the platforms C compilers will
run on ?

(2) You have to pay for that compiler and the compilers people can choose
for alternative languages are free of charge.

Yannick is right - people tend to go with what they think will give them
a job in the marketplace. That means if you want people to try out your
language you have to make it as easy as the current options to obtain
and use on your chosen platform.

It also means you have to make it easy to integrate libraries written in
this new language into code people have written in other languages which
they think will get them a job.

Don't forget how Linux took off. A lot of students had _free_ access to
this new thing called Linux and could play with it while still learning
the skills they needed to get a job. However, once in a job, those same
people, now familiar with Linux, started using it for small non-critical
jobs and as it proved itself, started using it more and more until we got
to where we are today with Linux.

If you want to replace C with something safer, that is exactly the same
path you are going to have to take.

I think there's a _big_ culture gap in this newsgroup between what you
know and what is required for a C replacement to really take off.

I think many of you are so used to working on highly critical systems
which simply cannot fail, along with all the support infrastructure and
support costs that implies, that you simply do not see the very different
world and mindset which is required to get a new language established.

There is a definite place for something which is not as rich as Ada but
is better than C and can be used for the jobs C is currently used for.

As I have said before, if you can get a direct C replacement established
(for example, my hypothetical Oberon-14) which can be used on a large
range of platforms, then once you have got people into a certain type-safe
mindset, you can start introducing them to Ada for the large projects.

And in the meantime, with this direct C replacement, we could get people
writing some critical libraries which don't have stupid buffer overflow
and other bugs.

Ada's a good language, but it's the whole ecosystem around Ada and the
difficulty of getting Ada compilers running on a large range of platforms
which has stopped it from being as widely used as it should be.

Simon.

-- 
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980s technology to a 21st century world