The code had already undergone two tests. 1. Electrical Rig tests. 2. Physical Rig Tests. Which showed the code meets the customer requirements.
The unit test step was the third step . Unit tests are written by independent developers. But not by looking at the procedure parameters and return values from code. But from the specification document where the algorithms were detailed for each procedure and each variable. That is to say detailed documents on code. Before code. The unit tests are to make sure the developer followed the documents. I think the problem here on this email line is that detailed documents of code before code is written is a challenging concept to grasp. This is also is known as best practice. Underlying my point that programmers want to dive into code. Then later they get unstuck. I dislike the concept of Javadoc for same reason that documents are generated from the code which flies in the face of best software development practice as I understand it. On Thu, 5 Dec 2019, 22:00 Martin Gainty, <mgai...@hotmail.com> wrote: > DOB correct > > if the dev isnt willing to verify their code works with at least one unit > test then whats the point of development? > > dave did you see the case of the pacemaker's recording of heart electrical > activity solved a homicide in Australia? > > https://www.smh.com.au/national/pacemakers-electrical-memory-that-foiled-a-murderers-alibi-in-court-20030120-gdg4uz.html > > best regards > m. > > ________________________________ > From: Dave Newton <davelnew...@gmail.com> > Sent: Tuesday, December 3, 2019 11:31 AM > To: Struts Users Mailing List <user@struts.apache.org> > Subject: Re: OFF TOPIC: Introduction > > On Tue, Dec 3, 2019 at 9:43 AM Zahid Rahman <zahidr1...@gmail.com> wrote: > > > If the developer tests his own code then there is a danger that he will > > pass his misunderstanding of specification. > > > > Correct. > > > > Anyway in the safety critical environments the one who wrote the code > does > > not do the testing. > > > > That's simply not the case across the board: there are multiple tiers of > testing, and the onus is on the developer to do the initial testing. > > For example, when I was writing pacemakers--you better believe we tested > the *crap* out of our code. Were we the *only* ones testing? Of course not; > there were *many* levels of testing--but it *started* with us. > > > > Tests are also done on hardware electrical rigs , simulating feed to > > sensors. > > > > Yep. And when I was writing fuel pump controllers not only did *I* have a > set of tests, so did the hardware guys, so did the combination of HW/SW > guys, so did the verification group, so did whoever gave us our > certification, etc. > > It's not entirely clear to me what your point is, or if you're just arguing > for the sake of arguing. > > Testing is, by necessity, the responsibility of everyone involved in the > project, which includes the developer. > > d. >
