Stephen BissAll of the Software is Legally Relevant
Updated: Sep 7, 2022
Purpose:
To establish that the international literature indicates that all of the approved instrument's software is legally relevant to reliability of the measuring system.
To distinguish "audit trail" (e.g. COBRA data) from instrument software (e.g. source code)
To clarify the relationship beteen software transparency and type approval
Note also the identification of software version in the type approval certificate. We don't require that in an Approved Instruments Order by the Minister. Quaere: Why not, when software version may change a threshold value such as the section 320.31(1)(a) requirement for a 10 mg/100mls maximum on a blank test?
Excerpt from cross-examination of a CFS scientist:
Q. Now I just want to go back to R-1-26 again, you, please – that’s the thick O-I-M-L bundle.
A. Okay.
... A. Tab 7, okay. Q. Tab 7, page...
A. Page 18.
Q. ...18, top of the page under 6.4, under A. Software, yes. Q. And if you can just read that paragraph, A. “The whole software of the breath alcohol analyser should be considered as a legal – as – should be considered as legally relevant. In the event of a software separation as described in 5.2.1.2 of O-I-M-L D-31:2008(7) the whole software is considered as legally relevant.” Q. So, it appears that, from the perspective of the international standard, of O-I-M-L R-1-2-6 that their perspective is that the whole of the software of the instrument should be considered legally relevant, but I understand that’s not the position taken by the Alcohol Test Committee or the Centre of Forensic Sciences recommendation.
A. That's correct. That’s an O-I-M-L as it says on the first page...
Q. Yes. A. ...of their chapter seven, it says recommendation.” I believe there’s also a
document from an organisation in the United States saying that the software is irrelevant as well
Q. And what organisation is that? A. I would have to get it from my vast number of documents there in order to tell you which exact organisation it was. Q. Maybe we should just check it during the coffee break? A. Sure.
...
U P O N R E S U M I N G : THE COURT: Okay, sir. MR. BISS: Thanks Your Honour. Over the break, Mr. P was able to find the paper that he referred to with respect to software. THE COURT: Okay. MR. BISS: And I should clarify what I’m – for this - purposes of this discussion right now, we’re talking about software, source code for the instrument, which is not the same thing as data, the audit trail for the instrument. A. Correct. The software operates the instrument and the source code is the written computer language that says ‘do this function or do this function’ whereas the information that the instrument collects, is the downloaded data, and you use software called Cobra to download the data from the instrument about all the parameters of what occurred at a particular time, whether it’s stand alone calibration check or a breath test sequence or a diagnostic failure. Q. All right, but just with respect to software. This came up when I was pointing out something in the O-I-M-L R-1-2-6. A. Yes.
Q. And you referred to.... A. Six point four. Q. Yes. A. On page 18. Q. In the 2012 version, Tab Number... A. Seven. Q. ...seven, yes? A. Yes. Q. Right. And you mentioned that there was an article that you knew of and specifically something from the National Safety Council’s Committee on alcohol and other drugs. A. Yes, it’s actually a position paper. Q. It’s not a scientific opinion? A. No, it’s a position paper. Q. Right, and the date.... A. Similar to the one of – with the A-T-C with respect to the proper working order and the materials required for determining that. Q. Yes. This position paper is dated – it says it’s adopted February 16th, 2009. A. Correct. Q. So, that would be before the O-I-M-L standard R-1-2-6 of 2012. A. The recommended standard, yes.
Q. And on – I just notice on – on page two of three, in this document about – and under “comment” about one third of the way down, or a quarter of the way down the page, it says, “Likewise, the international recommendation on breath alcohol analysers O-I-M-L R-1-2-6 evidential breath analysers, promulgated by the international organisation of legal metrology, an international treaty organisation of which the
United States of America is a member state, does not refer to or include the source code of analysers in its recommendation.” A. Yes. Q. Right. And so, obviously, they must have been referring to O-I-M-L R-1-2-6 from – the previous one, which I gather was 1988, I think it was, which is the Tab Number 6 in the book. A. 1998. Q. Oh, 1998, I’m sorry. That’s Tab 6, whereas 2012, the O-I-M-L took a completely different position from what we saw when you were answering my questions this morning. A. Yes. MR. BISS: All right. Your Honour, in fairness to the Crown this should be made an exhibit. THE COURT: Okay, this is National Safety Council’s committee on alcohol and other drugs position paper and Exhibit 39. EXHIBIT NUMBER 39: National Safety Council’s committee on alcohol and other drugs position paper – produced and marked.
See also:
OIML R126 1988 at oiml.org
OIML R126 2012 at oiml.org
OIML D31 2008 at oiml.org
Update: Note the 2021 revision to OIML R 126 states in Part 1:
7.1.10 Software The following requirements (7.1.10.1 to 7.1.10.7) established in OIML D 31 [5] shall be fulfilled. The severity of testing shall be selected independently for each requirement. The whole software of the EBA should be considered as legally relevant. However, if the software of the EBA is separated into parts, each part shall separately conform to these requirements. 7.1.10.1 Softwareidentification The software of the EBA shall be unambiguously identified with its version number and by the result of a hash function or by a checksum. The identification shall be inextricably linked to the software itself and shall be calculated, then presented or printed, on command or displayed during operation or at start-up. The software identification and all its parts shall be stated in the type approval certificate/certificate of conformity.