[SystemSafety] Making Standards available .....

[Peter Bernard Ladkin]

On 2016-05-14 09:16 , Michael J. Pont wrote:
> .... if we have the time
> available to teach functional safety – we should be introducing practical techniques for developing
> safe systems (and discussing various case studies).  

We have found it takes students 150 hours to learn causal analysis of incidents, and 150 hours to
learn hazard identification and a bit of hazard analysis. I distributed our curriculum already.

Then there's risk. Jens Braband teaches that, and has a good text on practical engineering
approaches (in German).

So we can go for about 450 hours just for the Guide 51 basics.

> It seems to me that one of the most influential “standards” that has emerged in recent years is
> MISRA C.  The standard is not free (but neither is it expensive).   It has (in my view) made a
> positive contribution to the goal of making the world a safer place.

That's about software quality. Add another 300 hours for software quality. Software quality is
certainly an important part of programmable-digital-electronic-system dependability, but it's only a
part.

> MISRA C is (of course) a coding standard.  What would also be useful would be a similar, pragmatic
> document that discussed design guidelines for software in safety-related systems.  We also need a
> document that describes how to record safety requirements and system requirements.
>  
> This (in my view) is the kind of material that we should be teaching our students. 

Well, yes, for the software quality part.

> Would anyone have any interest in getting involved?

Many people on this list have been involved for years trying to get people together on joint
projects to do stuff like this, and there is some success to report. Besides our efforts with
Finsbury on software quality, which led to actions within the formal standards framework, at
national and IEC level, on assuring objective properties of software, and on assessing the quality
of existing software through use data, there is the Data Safety Initiative run by SCSC, which came
out with a fine document which they distributed at SSS2016 and which I have used in recent
commentaries on draft standards and new work item proposals in the IEC.

So, yes, there is lots of interest on this list in getting involved in such efforts. Such past
efforts have been fruitful, so there is every chance that future ones will be also.

"Getting round a table" is nice, but I incline towards the use of teleconferencing facilities such
as Cisco's WebEx. It is far more resource-efficient.

Daniel Grivicic suggested:
> Marvin Rausand's book from the course 'Reliability of Safety-Critical Systems' is not inexpensive however I think it is easy 
> to follow and provides useful references for further information on each topic covered.

Yes, but it's only about HW reliability assessment.

Then you need good material on specification, including nonambiguity, non-contradiction, correctness
and completeness. And that material has to be for systems, not just SW, but the books on
specification are generally just about software.

If you really want to cover all the bases minimally, you probably need about a dozen books of this
sort. And for some topics (software reliability and reliability assessment for safety-critical
systems) there is no one good source.

I think designing an ideal curriculum for safety of E/E/PE-based systems would be a useful project.
Here's my start on a classification. Here, I mean "dependability" in the IFIP WG 10.4 sense rather
than in the IEC sense.

* Guide 51 basics
* Specification techniques
* SW dependability and quality assurance
* HW dependability and quality assurance
* Maintenance, logging of operational data, system change management
* Incident analysis and countermeasure design

PBL

Prof. Peter Bernard Ladkin, Faculty of Technology, University of Bielefeld, 33594 Bielefeld, Germany
Je suis Charlie
Tel+msg +49 (0)521 880 7319  www.rvs.uni-bielefeld.de





-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 455 bytes
Desc: OpenPGP digital signature
URL: <https://lists.techfak.uni-bielefeld.de/mailman/private/systemsafety/attachments/20160514/3513e4f2/attachment.pgp>