[SystemSafety] Fwd: Re: CbyC and unit testing

[David Crocker]

I've  been involved in producing CbyC tools for many years. My view:

- If the complete tool chain (CbyC tools, compilers, linkers etc.) and hardware work as intended, the software will perform according to the specification at the unit level.
- The primary function of unit testing in a CbyC process is to check that the tools and the process used to combine them are working as intended. We expect no errors to be found by unit testing. If we do find errors, there is something wrong with the tool set or the process.
- If we use the same tool set and process for enough projects, we may gain enough evidence and confidence in them to omit unit testing in future projects using the same tool set and process.

On 3 July 2020 21:28:15 BST, Olwen Morgan <olwen at phaedsys.com> wrote:
>
>
>
>On 03/07/2020 16:04, Peter Bernard Ladkin wrote:
>
>
><snip>
>
>
>In keeping with what I'd said previously, I do not regard mathematical 
>truths as analytic. (The assertion that they are is, IMHO, part of 
>Kant's monumental clusterf*ck - there's bigoted Buddhism for you :-).
>Hence:
>
>1. I'd hoped you'd realise that I might take a non-standard (to put it 
>mildly) view of the interpretation of logical formulae and that my 
>concocting a deliberately careless, gung-ho traduction might help to 
>elicit what you actually meant. I'm not as stupid as I made myself out 
>to be. ... ;-)
>
><snip>
>
>>>     My position is: ¬[](Use-CbyC -> ¬[]Perform-UT). That is to say I
>
>>> do not think it necessarily
>>> the case that use of CbyC renders UT unnecessary.
>> Your statement is logically equivalent to <>(Use-CbyC & []Perform-UT)
>
>> which is that it is possible
>> that you use CbyC and you have to perform full unit testing.
>> Those two statements aren't inconsistent with each other.
>
>... and ...
>
>2. In consequence of holding a purely empirical view of logical 
>modalities, I view claims of these statements' equivalence based solely
>
>on logic as profoundly meaningless word-games.
>
>
>(For readers who are by now totally confused, I hold to Arnold's 
>position that mathematics is part of natural science and that 
>mathematical statements and scientific statements are not different in 
>character.)
>
>
>So, after now having elicited what seems to me to be a helpful 
>clarification of your position, I repeat a previous question: If you 
>hold that the use of CbyC means that you can possibly avoid 
>unit-testing, are there any circumstances in which you would concede 
>that you can use CbyC and yet in which it is not possible to avoid unit
>
>testing? Please explain your reasoning.
>
>Olwen
>
>PS: Readers please be aware that PBL has already, privately, suggested 
>that I am a nutcase ... which I take as a great compliment ... :-))
>
>
>
>
>_______________________________________________
>The System Safety Mailing List
>systemsafety at TechFak.Uni-Bielefeld.DE
>Manage your subscription:
>https://lists.techfak.uni-bielefeld.de/mailman/listinfo/systemsafety

-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.techfak.uni-bielefeld.de/pipermail/systemsafety/attachments/20200703/24f0cbdd/attachment.html>