[SystemSafety] C for OSs

[Olwen Morgan]

Well, it depends on what you mean by hard real time. Erlang has been 
extensively used for telecomms applications at Ericsson and they have 
some highly time-critical constraints. But they're probably not what 
most people would mean by hard real-time. On the other hand, I can't see 
any essential technical impediment to an Erlang subset that would avoid 
the need for garbage collection (though possibly requiring an abstract 
machine other than BEAM)? What I'd really like to see is an Erlang 
abstract machine that includes a minimal hardware abstraction layer for 
small microcontroller targets (possibly a cut-down RTEMS? - I'm floating 
arguably fanciful ideas here). One could live with a cut-down HAL-like 
RTOS if it sat alone and didn't need to be Linux hosted.

Once again, I suggest, the availability of Linux is blinding systems 
designers to simpler solutions.

Olwen


On 09/09/2019 11:42, David Crocker wrote:
> I don't know Erlang, but I understand that it uses garbage collection
> (like most programming languages that are nice to program in). Doesn't
> that render it inappropriate for building hard real-time systems?
>
> David Crocker, Escher Technologies Ltd.
> http://www.eschertech.com
> Tel. +44 (0)20 8144 3265 or +44 (0)7977 211486
>
> On 08/09/2019 21:43, Olwen Morgan wrote:
>> All,
>>
>> Forgive me for multiple postings but ideas are occurring to me in fits
>> and starts on this one.
>>
>> Given a free hand to choose a language running on an x86_64 target for
>> implementing a highly concurrent critical system, I'd almost always go
>> for Erlang, whose track record at Ericsson is mind-bogglingly good. Of
>> course, the BEAM Erlang abstract machine needs a hosted environment in
>> which to run. One helpful development in this area is the GRiSP2,
>> single-board hardware abstract machine for Erlang/Elixir. This gets
>> rid of UNIX but still leaves dependency on an relatively
>> low-availability hardware.
>>
>> Now, if someone could implement a true bare-metal BEAM for x86_64
>> (possibly only a smallish step farther?), you wouldn't need
>> special-purpose boards ... and I could die happy ...
>>
>> ... but not yet ... (in case those of you who'd be glad to see the
>> back of me are rubbing their hands with glee) ... :-O
>>
>>
>> Olwen
>>
>>
>> _______________________________________________
>> The System Safety Mailing List
>> systemsafety at TechFak.Uni-Bielefeld.DE
>> Manage your subscription:
>> https://lists.techfak.uni-bielefeld.de/mailman/listinfo/systemsafety
> _______________________________________________
> The System Safety Mailing List
> systemsafety at TechFak.Uni-Bielefeld.DE
> Manage your subscription: https://lists.techfak.uni-bielefeld.de/mailman/listinfo/systemsafety
>