Is Quality Assurance solely reliant on software?

[Asher]

Originally posted by Japher
I never thought PH would actually be writing the software. I assumed he would be helping developing the parameters and representing the end user as to the application of the software.

Well, that's good for you, but PH:
1) Mentioned he was sought after for his "IT skills" in making a program -- if he was to set parameters for the program, he should have said so. He was intentionally misleading people into thinking it was something more complex than it was. Setting requirements and parameters for a program is hardly a great career opportunity, IMO...
2) Made mention to "computer programming" in reply to me questioning his competence in software development
3) Made no effort whatsoever until just now to clarify he's not developing a damn thing, he's writing a list of requirements and banging away on someone else's code
4) I said, and I quote: "I can't imagine someone like PH in a development position, because he knows the sum of jack **** about it" which clearly indicates it was my assumption, way at the start, that he'd be in a development position. He did not correct this, because he wanted people to think he was more diverse than he lets on.

It is a reasonable assumption to make, given the above, that he'd be writing code.

But, apparently Software Writers are Omniscient and don't need any practical knowledge of the process for which the software is being writen.

I've said repeatedly the end users are to be involved to set requirements and do acceptance testing.

I'm beginning to see a trend among Pharmaceutical employees and reading comprehension skills...

Just curious, when you hear "software developer", do you think of someone like PH writing requirement lists and doing testing, or do you think of a programmer and/or designer?

[duke o' york]

Surely it means both. Both of these groups should be involved in software development, although only one would actually call themselves "software developers". Without involving the user, then the software will be utterly useless. You need to know their requirements before you even begin to design software.

[Jon Miller]

Originally posted by Asher
The debate in the other thread is should anyone be allowed to develop critical software.

Should aeronautical engineers program the flight system for space shuttles and the landing algorithms for Airbuses?

Should mechanical engineers program the on-board computers in cars?

Should biochem scientists program QA software for drugs?

The answer, overwhelmingly in responsible organizations, is no.

But who but physicists know the physics to program the software used in experiments, simulations, and models?

JM

[Asher]

"software developers" are people who develop software.

Using it otherwise is a misuse of the term, and an error on the part of the person using it.

"Requirements Engineer" is the term applied to people who take requirements and turn them into a formal document and charts for the "Software Engineers" to then make the architecture for, where the "Software Developers" then implement.

Requirements Engineers are usually specialist Software Engineers...

[Asher]

Originally posted by Jon Miller
But who but physicists know the physics to program the software used in experiments, simulations, and models?

JM

Do physicists make commercial programs?

It's one thing to use programming as a tool for your personal use/project to aid in, for example, research. That can and usually is done by whoever is doing it.

I'm talking about developing real software, not just big calculations for research.

[Jon Miller]

Originally posted by Asher
Buy books! They'll help. Not programming books -- software engineering and architecture books. I could give recommendations.

I could probably use some of these. Some of the people who are long time software are really good (I think). But my few small scripts, while I tried, aren't as well written as they could be (I think). If I am going to be on the software side of experiments, I need to get much better.

Jon Miller

[Jon Miller]

Originally posted by Asher

Do physicists make commercial programs?

It's one thing to use programming as a tool for your personal use/project to aid in, for example, research. That can and usually is done by whoever is doing it.

I'm talking about developing real software, not just big calculations for research.

They write programs that a lot of other physicists use. For example ROOT http://root.cern.ch/

JM
(edit another example in nuclear/particle physics http://wwwasd.web.cern.ch/wwwasd/geant4/geant4.html)

[Asher]

Originally posted by Jon Miller
I could probably use some of these. Some of the people who are long time software are really good (I think). But my few small scripts, while I tried, aren't as well written as they could be (I think). If I am going to be on the software side of experiments, I need to get much better.

Jon Miller

The books I'm most familiar with aren't programming books, they're software engineering books. Dealing with a customer's requirements and developing that into a well-designed piece of software.

[Asher]

Originally posted by Jon Miller
They write programs that a lot of other physicists use. For example ROOT http://root.cern.ch/

JM

The authors appear to be computational physicists?

According to the about page, http://root.cern.ch/root/Authors.html they also have computer scientists working on it.

[Jon Miller]

Originally posted by Asher

The books I'm most familiar with aren't programming books, they're software engineering books. Dealing with a customer's requirements and developing that into a well-designed piece of software.

So I know the language fine (well, sorta fine ). It is making the software well-designed that I would say I am lacking in.

But maybe we are saying similiar words but meaning different things.

Jon Miller

[Jon Miller]

Originally posted by Asher

The authors appear to be computational physicists?

According to the about page, http://root.cern.ch/root/Authors.html they also have computer scientists working on it.

Well, it is half high energy physicists...

Maybe Geant is a better example?

JM

[Asher]

Originally posted by Jon Miller
So I know the language fine (well, sorta fine ). It is making the software well-designed that I would say I am lacking in.

But maybe we are saying similiar words but meaning different things.

Jon Miller

Well, it depends what kind of software you're writing. Most of the stuff in software engineering books is object oriented.

But for the purposes of physicists, you're probably better off with procedural programming (think C, Fortran, etc. instead of C++, Java, etc.), just because it's more natural when thinking of mathematical programs.

A general suggestion is to keep functions short -- if it's over one page long, it's too long.

[Jon Miller]

Yeah, I know procedural.

But ROOT is object oreinted, and that is what I am suppose to be doing (but not as well as I should be, I compare what I write to what others write and mine is much worst (and their programs are much more complicated)).

Jon Miller
(been using C++ for 4 months, have used Fortran 90 for 6ish years)

[Asher]

If you want you can send me a copy of some of your code, then I can give you some pointers.

Writing good OO code is hard, almost all of it around is pretty bad. You should never see a "switch" statement or long "if-else" chain in good OO code, for example.

[Jon Miller]

Thanks

I actually haven't written anything for a couple months now, wasn't really working for a while and then recently (last 8 or so days) I have been just trying to understand someone elses code (which I have done, and he writes a lot better OO then I).

Jon Miller