It is very hard to have a serious discussion on any topic if we can’t agree on what words mean. In particular, definitions are important and of course the relevant terminology.
In an upcoming blog I plan to cover how standards are written, but in many cases they are written and read by non-native English speakers. This makes the correct use of definitions and terminology even more important.
An interesting example is the use of the word “duplication” in Annex E of IEC 61508-2:2010. According to various online dictionaries, duplication means “making an exact copy of something”. However, speaking of one of the authors of the Annex, they did not intend to limit it to exact copies but rather any form of on-chip hardware fault tolerance. You then wonder did some of the committee members agree to it only because of the identical redundancy within scope or did they all understand the same thing when it was agreed upon.
This raises an important question as to whether to claim compliance with a standard. You must comply with what is written in the standard or what was intended by the authors. However, I am getting off topic. Another example is the use of PFH (IEC 61508) and PFHd (IEC 62061). In both cases it means the average frequency of dangerous failure per hour. In that case PFHd is the clearer term to me as the “d” emphasizes that it is the dangerous failure rate which is involved.
Most single part standards have a clause or sub-clause containing definitions and are always worth reading. Multi part standards such as IEC 61508 can have a section containing nothing but definitions such as IEC 61508-4:2010. Standards derived from a basic safety standard such as IEC 61508 should justify any changes in the definitions from the basic safety standard but this is not always done, and leads to confusion to those who fail to spot the changes. Standards such as those in the IEC 61800 series highlight any definitions used in the standard by putting them in italics, which I believe is helpful. However, this technique does not seem to be widely used.
Outside of the definitions with the standards themselves (which you have to buy to access) the standardization bodies make some very useful technical dictionaries available.
- IEC – Electropedia http://www.electropedia.org/
- ISO – Online browsing platform http://www.iso.org/obp
- ISO/IEC/IEEE 24765:2017 International Standard - Systems and software engineering – Vocabulary
- ISO concept database – see https://www.iso.org/obp/ui/
The ISO concept database is particularly interesting as it searches across the various ISO standards for whatever term you enter – try for instance “embedded software”.
Electropedia is useful because not alone does it give you the definition of the item but also gives you a translation in various languages including French, German, Spanish, Arabic and Korean.
Also of note are the ISO/IEC directives part 2 where guidance is given on how to write standards. Sub-clause B.2 lists reference works for language.
Figure 1 - Reference works from ISO/IEC directives part 2 2011
Sub-clause D.1.4 gives addition guidance on which terms need further explanation.
Figure 2 - Requirement to define terms from ISO/IEC directives part 2
And later gives the guidance that “The form of a definition shall be such that it can replace the term in context”.
And now a confession, a habit related to terminology which I am finding hard to break is the use of the term “FIT rate”. FIT stands for failure in time and for integrated circuits, it is normally expressed in units of failure per billion hours of operation. However, it has been pointed out to me that since FIT already implies failure per billion hours of operation. The use of the term “FIT rate” is wrong and you should only say “FIT”. For example “The FIT for this IC is 5” as opposed to “The FIT rate for this IC is 5”. Personally, I don’t find the use of the somewhat redundant “rate” offensive as I believe it clarifies what is meant by FIT, but some do and the functional safety profession may attract more pedants than general engineering.
Lastly, I can never understand why one standard would use fault injection while another would use fault insertion to mean the same thing. It just adds confusion and I can never remember which one is more appropriate for a given domain. Not only is it confusing but it is also open to abuse or misuse when terms such as safe failure and dangerous failure change between standards such as IEC 61508 and ISO 26262.
I have been looking for an excuse to use this week’s video – the commentary is from the 2018 Irish vs France six nations match closing stages. The commentary is in French but you don’t need a dictionary to understand the reaction or for what one of the French players say as the final kick goes over – see https://www.youtube.com/watch?v=rRFw54vUEPY