[SystemSafety] Does "reliable" mean "safe" and or "secure" or neither?

Dave Banham dave.banham at gmail.com
Sun Apr 24 19:20:32 CEST 2016


I strikes me that a lot of this discussion has revolved around various 
definitions that are either generic and abstract, or have a specific context 
without (generally) making it clear what that context is.

I do agree that we need good (if not sound) definitions for reliability, 
dependability, integrity, because there is considerable propensity to use them 
in everyday English with overlapped meaning. Moreover, I propose that there is 
subtly different application in these terms when considering physical things and 
non-physical things. This is further compounded by the behaviour of some 
physical things to be defined by non-physical things, as is the case in point of 
a computer based system whose behaviour is defined by the set of instructions in 
a so called program. A further complication also arises when emergent behaviours 
(intended/desired, undesired/unwanted but known, and undesired and 
unanticipated) also exist in the thing that we want to describe with such 
predicates. The main difference between physical and non-physical things is that 
physical things decay; they age and they wear, whereas non-physical things do 
not, but even here we have to be clear over the difference between information 
and the data that represents that information using a physical means.

What is reliable software? Answer: software that performs as required. (This of 
course assumes that by "software" we mean a program that is being executed by a 
machine.) Unreliable software does things that were not required of it, often 
with the connotation of undesired behaviour. The likelihood of software doing 
something that is not desired is a function of how well we understand the 
process of software execution (or indeed system use) that produces the undesired 
effect. When its is well understood then we can say we have a systematic defect 
of some sort, and when it is not at all understood then it will manifest as a 
statistical process. Our ability to shift from the latter to the former is all 
about the amount of effort we are prepared to invest in gaining that understanding.

Dave Banham

On 24/04/2016 17:12, paul_e.bennett at topmail.co.uk wrote:
> On 24/04/2016 at 4:55 PM, "Chris Hills" <safetyyork at phaedsys.com> wrote:
>> Hi All
>>
>> As I started this (and was only expecting a couple of replies
>> referring me
>> to the same definition)  I thought I should chip in.
> We do need to have some general agreement about certain terms. The
> three main words that we should be absolutely clear on are:-
>
>    Reliability
>    Dependability
>    Integrity
>
> A couple of other terms that should also be defined are:-
>
>    Safe
>    Secure
>
> however, I would expect the last two to be defined as part of the
> requirements definition of terms.
>
> I use the OED definitions for the most part (always one that seems
> to fit the bill).
>
> Regards
>
> Paul E. Bennett IEng MIET
> Systems Engineer
>



More information about the systemsafety mailing list