[SystemSafety] The Importance of Standards
Les Chambers
les at chambers.com.au
Mon Jan 9 12:56:13 CET 2023
Well said Peter and Andrew
I feel your pain. But courage, your suffering is insignificant compared to
that of Catholic Priest Jan Hus (1369 - 1415) as the Pope's men burned him at
the stake for his sins.
Hus' special crime was to push back against pathological behaviours of the
Roman Catholic Church of his day. These behaviours repeat in various contexts
throughout history. With the IEC we have deja vu all over again.
My best pass at a behavioural model is as follows:
An elite, having come into possession of a body of knowledge (BOK), sought
after by the masses, declares its members holy and proceeds to limit the
common folk's access to that BOK. Typical motivations are ego, power and
money.
The actions that lead Hus to the stake included:
1. Calling for a higher level of morality among the priesthood. Financial
abuses, sexual immorality, and drunkenness were common among the priests of
Europe.
2. Calling for preaching and Bible reading in the common language.
3. Opposing the sale of indulgences.
(These were documents of personal forgiveness from the Pope which were sold
for sometimes exorbitant prices. You could go to heaven for a pot of gold ...
apparently.)
The essence of these actions projected onto the IEC's role in today's safety
critical systems development are as follows:
1. Morality among the priesthood. Religious belief is an apt metaphor for
alignment with the principles of safety critical systems development. Your
belief needs to be strong to engage with confidence in exchanges such as ...
CEO: "Well son, what do I get if I spend an additional million dollars on this
smoke extraction system?"
Systems Engineer: "Sir, trust in IEC 61508, if you spend an additional million
dollars nothing bad will befall you."
(followed by industry sector specific disaster case studies - my favourite for
chemical processing is Bhopal preceded by this image:
famouspictures.org/bhopal-gas-disaster-girl/)
Systems Engineer: "Can you imagine being held responsible for this sir?"
A good Christian needs easy access to the Christian Bible. A good Muslim needs
ready visibility of a Quran. While working in a Muslim country I was impressed
with the way Muslims consider their spirituality at prayers five times a day.
Would that systems engineers were that devout.
It is well known that communities with free access to information flourish.
Those whose enlightenment is blocked by the vested interests of a few "holy
men" remain in the dark ages. The keepers of the IEC business model should
reflect on their actions and cease their immoral blocking, rent seeking
behaviour.
Choosing to do nothing is to invite the modern equivalent of a Reformation.
See Martin Luther and the "Ninety-Five Theses", 1517. The BS of indulgences
ran too deep for Luther. He pushed back in the tradition of Jan Hus 102 years
after his brutal execution.
2. Preaching and reading in the common language. Standards such as IEC 61508
need to be explained line by line. In the 1990s I did exactly this,
explaining ISO 9001's application to software development to a few hundred
Australians. I was blown away by how such a simple standard could generate so
much confusion and so much conversation. Personally, it was fun. I attracted
much positive feedback. The most common comment was. "Oh! Thank you Les. Now I
understand WTF I'm supposed to do."
Without dialog with knowledgeable humans (or smart AIs), expecting these
standards to be understood and implemented ... ever ... is analogous to
expecting your congregation to do good and be good by lecturing them in Latin.
A contributing factor to the success of my ISO 9001 courses was that I could
afford to include an original ISO 9001 in my course notes. This is not
possible with IEC 61508. Copyright restrictions must represent a minefield for
people who attempt to provide training in this standard.
I note that, in scope and complexity, ISO 9001 is to IEC 61508 as Conrad's
Heart of Darkness is to Tolstoy's War and Peace. You need a multidisciplinary
team to "preach" it effectively and over a period of weeks. I suspect that
this task will ultimately fall to an AI. We are already embarked on that path.
Go to chat.openai.com and ask "How do I comply with IEC 61508" or command the
AI to "Generate a IEC 61508 compliant safety plan outline". The openai chat
bot is a general tool, it will fit you up with a delicious linguine recipe -
with white clam sauce - if you're that way inclined.
Imagine the possibilities if it was trained in the minutiae of 61508
compliance!
3. Selling indulgences. The modern business "goes to heaven" when it gets its
invoices paid. Yes indeed, the standards compliance imperative becomes a
potent moral force when attached to conditions of contract. If you don't
comply you don't get paid. The contractor's first step therefore is to
purchase the indulgence of the IEC but unlike the indulgences of old, the
whole team requires elevation to heaven not just one rich person. You
therefore need multiple copies of the standard, which at its current list
price creates a problem - which is usually solved by a descent into copyright
criminality at the Xerox machine. Wither morality now?
But what of your average bear, an open source developer for example, without
the luxury of a wealthy company to provide a copy of the standard or an
employee of a wealthy company without contracts that require standards
compliance?
I'll tell you what ...
The grave's insatiate maw,
The void, the vast abyss,
The nothingness, the shade,
The silence and the sleep. - "The Dark" by Emily Brontë
... of the rank ignorance that kills people.
Each year we add more punters to the ranks of this dark rank. Consider only
two companies in one industry sector. Ford and GM alone have committed to a
collective > 40 billion EV development budget for 2020 - 25. Given the
productivity of the average developer and the need for millions of lines of
safety critical code in an EV we have a massive demand for software engineers
with safety credentials. With the glacial takeup of 61508/26262 style
standards, actively obstructed by IEC "holy men", these jobs will be taken by
inexperienced people with no concept of a hazard or a hazard reduction
technique and no craving for the fellowship of the meaty red team code review.
Ergo the IEC's pricing policy is a hazard with potential for devastating down
stream effects. They should be leading but instead they are acting as dead
weight.
My suggestions for breaking down barriers to adoption are:
1.Packaging. Package standards with an AI.
2. Accessability. Make them available on line at a peppercorn rental eg as a
Netflix subscription.
3. Synthesis. Provide open access to standard updates. My personal experience
in begging visibility of candidate 61508 updates is akin to a known pedophile
requesting access to a childcare centre. I am, apparently, insufficiently
holy. A career, commencing 1975, in developing, living with and training the
faithful in safety critical systems development is insufficient. I've searched
my soul. Could it be that I also need to give up the women?
No matter, the naysayers of open access should study the evolution of Linux -
an infinitely more complex piece of intellectual property (ask
chat.openai.com).
Finally I say to you IEC "holy men".
Your business model is not sustainable. The demand of the faithful is too
great, the supply of wisdom is too pricey, the disappointment in your low
level of morality is growing.
There will be blood.
There will come a Martin Luther.
There will be a Reformation.
The communities that depend on these standards will split just as the
Protestants parted from the Catholic Church.
Repent now, open your standards, or risk "the vast abyss" of irrelevance!
As for you Andrew Banks, keep the faith, its unlikely they'll burn you at the
stake for speaking the truth. The holy have evolved past that (one would
hope); evidence Martin Luther dying quietly in bed aged 62. Be aware though
that attempting to sell an overpriced closed standard to an open source
community WILL get you singed. The failure of the IEC to spot the oxymoron
betrays an eye-watering lack of insight.
Cheers
Les
> Hi Peter
>
> A good post. For my sins, I've recently picked up the role of chairing an
ISO study group into the use of ISO/IEC standards in the open source community
- and in particular the barriers to adoption.
>
> When I suggested that cost was a serious barrier, this didn’t go down well
- for example, even the flagship software life cycle processes standard,
ISO/IEC/IEEE 12207, is CHF 208 for a single-user licenced copy. Heck, people
get hot under the collar when asked to pay £5.00 for a copy of MISRA C.
>
> The sale of standards (nationally and internationally) keeps a lot of people
in employment - but not those of us doing the work (we rely on supportive
employers).
>
> Andrew
>
> -----Original Message-----
> From: systemsafety <systemsafety-bounces at lists.techfak.uni-bielefeld.de> On
Behalf Of Peter Bernard Ladkin
> Sent: 09 December 2022 11:46
> To: The System Safety List <systemsafety at techfak.uni-bielefeld.de>
> Subject: [SystemSafety] The Importance of Standards
>
> Back in 2016, I wrote a short article comparing the approach of EUROCAE ED-
153 to software safety and that of IEC 61508. It was presented at the 11th IET
International Conference on System Safety and Cyber Security in 2016.
>
> Papers weren't required to accompany talks. But I wrote the paper and it was
duly "published" in that USB sticks containing written accompaniments to talks
were distributed with the conference materials. (Martyn Thomas pointed to the
irony of a Cyber Security conference distributing USB sticks without any kind
of assurance.)
>
> The proceedings are available. The IET offers them for sale for £79.
> https://digital-library.theiet.org/content/conferences?pageSize=100&page=1
>
> Most academic publishers offer individual papers for sale. I do not agree
with the kinds of prices they charge, but £79 is way above those prices.
>
> I did approach the IET about this. They responded that individual papers are
available at no charge to all IET members and affiliates (that is, people
entitled to log in to IET on-line services). I guess that solves it for 160+K
people.
>
> But no one has to go that route. It's up on ResearchGate, a
preprint+published-paper collector.
> ResearchGate informed me today that 6,000 people have read it.
>
> That paper was written when I was a German public servant with my salary
paid by German taxpayers, as also Managing Director of the tech-transfer
company Causalis on whose behalf I donated it pro bono publicum. I do not
agree with restricting its distribution. I want it to be open access and,
thanks to ResearchGate and my still-maintained Uni WWW site (and its mirror,
paid for by Causalis) it is.
>
> The main point I wish to make is this. 6,000 people have wanted to know the
similarities and differences between ED-153 and 61508. They have their
reasons, and I surmise it is not because they find my paper more entertaining
to read than Stephen King.
>
> Engineering standards are a public good. Except they are not public.
>
> We are currently running through the German comments on IEC 61508 Ed3 CD. We
have had three full days of discussion, just on which ones to forward to the
IEC and which ones not to, and a further three full days are planned. That is
300 person-days, just for that one task (and then there is the huge effort put
in by particular people to collate and sort the comments and ease the
discussion. I am currently in awe at the skill of one colleague who conducted
yesterday's 7-hour meeting flawlessly, without apparently cognitively phasing
in and out, probably beyond my capabilities.)
>
> When this decade-long task is finally finished, the result will be proudly
........ sold by the IEC to anyone with CHF 1400. (And, may I say, actively
copyright-protected.)
>
> John Knight, RIP, Martyn Thomas and I have repeatedly expressed our
discontent with this and other aspects of engineering standardisation
https://scsc.uk/scsc-126
>
> Standards are important. We need to move to a model in which they work as a
public good.
>
> PBL
>
> Prof. i.R. Dr. Peter Bernard Ladkin, Bielefeld, Germany
> Tel+msg +49 (0)521 880 7319 www.rvs-bi.de
>
> _______________________________________________
> The System Safety Mailing List
> systemsafety at TechFak.Uni-Bielefeld.DE
> Manage your subscription: https://lists.techfak.uni-
bielefeld.de/mailman/listinfo/systemsafety
--
Les Chambers
les at chambers.com.au
+61 (0)412 648 992
More information about the systemsafety
mailing list