Re: [on safety cases]



Re: [on safety cases]

From: Yiannis Papadopoulos <yi_papadopoulos_at_xxxxxx>
Date: Thu, 19 Nov 2009 12:28:42 +0000 (GMT)
Message-ID: <247896.64227.qm@xxxxxx>
Martyn, 

Standards are inevitably imperfect, we agree on this.

However, I am not sure that we should not strive for good standards; how is safety going to be regulated, don't we need some commonly accepted ground? 

Perhaps, the way to go is to be open and recognise the limitations of standards (e.g. what you can prove and what you cannot prove regarding safety) within the standards. The same applies to safety cases.

Standardisation is a useful process. A standard is a public document which aspires to address the problem in its entirety and which is open to scrutiny. Anyone who has better ideas can oppose the standard and propose an alternative. There is a lot of bedate about 61508 but, to my knowledge, no one has proposed a much better alternative. 

What will happen if we don't have standards? 

Is everyone supposed to do whatever they like in terms of assurance and some other people to judge them by applying their own ad hoc criteria?

best regards

Yiannis


>----- Original Message ----
From: Martyn Thomas <martyn@xxxxxx>
To: safety-critical@xxxxxx
Sent: Thu, 19 November, 2009 11:49:17
Subject: Re: [sc] Re: [on safety cases]



Nancy Leveson wrote:
>
> Nothing can guarantee safety to a specified level of probability or risk,
> particularly for losses involving design errors (including software errors
> and requirements errors) and human errors. Creating a standard does
> not make it possible,  it simply creates activities that distract people
> from
> concentrating on the things they *can* do that will make systems safer.
>
> Nancy
>  

Well said. I agree completely.

Martyn


      
Received on Thu 19 Nov 2009 - 12:28:47 GMT