Definition Test Software - Definition - ISO 9001:2000 Clause 7.6

[gpainter]

Your definition of Test Software? Thanks in advance.

 

[Graeme]

Re: 7.6 Test Software

gpainter said:

Your definition of Test Software? Thanks in advance.

Please clarify: are you referring to software as mentioned in 7.6 of ISO 9001:2000? (My copy of ISO/IEC 17025:1999 stops at 5.10.9.)

 

[gpainter]

ISO 9001-2000

 

[Atul Khandekar]

Can someone provide a definition or any examples please?

Would CMM Software qualify as Test Software?
How about an SPC software?

 

[Mike S.]

"Test software" (using 9001-1994 terminology) and "computer software" 9001-2000 section 7.6 terminology) to me means this:

I need to test a certain parameter(s) of a product.

I make a measurement(s) of some type and input the measurement(s) into a computer program to do some complex and/or repetitive math calculations for me and spit out the answer.

Also, some of my testing is automated. I hook a computer up to a piece of test equipment, load a sample, run the program, and let the computer control the test instrument, query the instrument for data, calculate (as required) and print-out a test report.

In both cases I need to verify that the computer program does the job correctly before I approve it for use.

I also take some steps (as appropriate) to make sure no one screws with my test software so that it starts doing the wrong things. A password, restricted access, and/or verification of proper functioning before each test are possible appropriate methods to do this.

I hope this helps! If not, write back.

 

[Atul Khandekar]

I make a measurement(s) of some type and input the measurement(s) into a computer program to do some complex and/or repetitive math calculations for me and spit out the answer.

Thank you Mike. I understand from this that software that's not actually involved in 'measurement', but does complex calculations - spcifically SPC or MSA software - is also Test Software and needs to be validated periodically. Is this validation done by the users or developers of the software - or both?

 

[Ryan Wilde]

Here is a file that may help. It is geared toward 17025, but it is fairly universal.

Ryan

 

[Atul Khandekar]

Ryan, Thank you very much!

 

[gpainter]

Thanks Graeme and Ryan.

 

[Mike S.]

Thank you Mike. I understand from this that software that's not actually involved in 'measurement', but does complex calculations - spcifically SPC or MSA software - is also Test Software and needs to be validated periodically. Is this validation done by the users or developers of the software - or both?

Ryan's post/attachment is certainly authoratative, especially with regards to 17025, and should cover any critical needs/situations, but it may be a bit of overkill for some with only ISO 9K in mind, especially for custom stuff. ISO 9k requires that you confirm "...the ability of computer software to satisfy the intended application..." prior to initial use and reconfirmed as necessary.

When I write a simple BASIC or Excel etc. program to do some calculations for me, I simply run it to exercise all the options and compare the program's result(s) to my hand-calculated result(s) and if everything matches I consider it confirmed suitably and I'm done -- kinda. As for re-confirming, my long-standing habit is to measure a "standard" (even if it is just an internal "working standard") at least once a day on every test setup that will be used that day prior to measuring any unknowns. This simple step gives me great confidence that not only is the software okay, but so is the test instrument(s), cables, connectors, and my setup. You may have had something with a 6-month cal. cycle "calibrated" formally just yesterday, but who is to say something in the instrument doesn't go haywire today, knocking the thing out of cal? Or that I might set it up wrong? This little habit helps me to make Mr. Murphy's job just a little more difficult. JMO.