Re: Need help with PowerPC/Ada and realtime tasking

[JP Thornley <jpt@diphi.demon.co.uk>]

In article: <EACHUS.96May28192552@spectre.mitre.org>  
eachus@spectre.mitre.org (Robert I. Eachus) writes:
>                        .........  There are failure rates for the
> sensors, some tied to the geometry of the radar or the signatures of
> the target, and some failures due to the algorithms (and software)
> used. 

As I read this, the failure rate is connected with the choice of 
algorithm.  If this is not specified in the requirements then these 
failure rates (resulting from the algorithm chosen by the software 
engineer) have to be brought into the design and code level hazard 
analyses, so that they can be fed into the wider system safety 
assessment.  But the point that this responded to is not talking about 
failure rates of algorithms in this sense, but failure rates of the 
implementation in producing the 'correct' value according to the 
algorithm specification (and yes, the software engineer is responsible 
for the error bounds on the implementation method chose).

   [snip]

>     Last but not least, there is the ugly ghost of Godel.  Systems
> below a certain level of complexity can be 100% right or 100% wrong.
> Above that line, software systems look a lot more like hardware.

I wonder if this is where computer science and (my sort of) software 
engineering part company.  Or do you while away the hours when 
travelling by Boeing 777 wondering whether the flight control computers 
are about to be presented with an undecidable proposition?

> 
> --
> 
> 					Robert I. Eachus
> 
> with Standard_Disclaimer;
> use  Standard_Disclaimer;
> function Message (Text: in Clever_Ideas) return Better_Ideas is...
> 
> 
-- 
------------------------------------------------------------------------
| JP Thornley    EMail jpt@diphi.demon.co.uk                           |
------------------------------------------------------------------------