donderdag 18 september 2008

Never again and public service announcement

I hereby solemly plegde never to change an article or presentation again because a couple of colleagues say so.
What happened? In the beginning of this year I made a synopsis about HumanBasedAnalysis (then called HumanBasedTesting as it was intended as antagonist of Risk or Requirementbased Testing), it was reviewed, but some colleagues said I should narrow the subject down. I did so, by applying it to the Teststrategypart and calling it thus HumanBasedAnalysis, wrong, wrong, wrong.
I got to present my piece at the Najaarsevent of Testnet last tuesday, it became a fiasco. Technically it wasn't bad, but with respect to the content, basically: it Sucked!
Due to 'jibbers' I mixed up a technique; namely Functionpointanalysis and ProductRiskAnalysis, this wouldn't have happend if I had stuck to my original plan, the HumanBasedAnalysis is namely applicable within all phases of the test-trajectory, allthough mainly at the left-upper part of the V-model- and not only at Strategy definition.
So after a couple of sheets I was 'saved' by one of the audience who said he couldn't quite put his finger on the spot. I tried to explain, but again, the mixup of techniques was my 'millstone'.
Luckily I got a bit of explanation done, but also I discussion had arisen in the audience.

I thought the discussion was good though; it confirmed my original idea of Human Based Analysis (because part of the discussion broadened to the fact that this should also be used as soon as the phase where requirements are set; good; because requirements should be validated and verified (in my opinion with the end-user as reference). The discussion was also proof of that this kind of thinking is perceived as important but to put it in a context like I did, caused a great deal of controversy. Note: when looking at the V-model then the Human Based Analysis is done between Setting Requirements and Designing the FO.

OK, for keeping the peace: I'm solely responsible for the content on HumanBasedAnalysis. My company is not responsible and does not explicitly support me in this way of thinking, my idea is - seemingly - a bit difficult to understand and with controverse, so I still have a lot of explaining to do.
Let it be stated that when - in the future- it catches on; it was solemly my idea to introduce it
I stand by my subject and am proud of it.

For the people who attended: I'm sorry to have desillusionised you. Perhaps the following definition will clearify the intention somewhat more (I'm better on script than on speach)

Human Based Analysis
Human Based Analysis is a compilation of techniques where the end-users or experts - other than appointed stakeholders - are explicitly asked to point out the risks within the proces to be automated or the software to be developed or what requirements are the most important.

The Human Based Analysis is primarily done between Requirement Specification and Functional Design, but can be done on other part within the V-model allthough when at 'User Acceptance Test' level, you're (probably) too late.

The Human Based Analysis has in a lot of cases the consequence that two things are to be adjusted: 1. The requirements and 2. The testdesign


Hopefully to be continued....

Geen opmerkingen: