Send IT workers to jail

[The Lone Gunman]

Sorry, you are just plain wrong here.

There are requirements that this type of software is developed "properly" by qualified people.

Just because some fly-by-night companies have managed to get away with not doing it properly, doesn't mean that they wouldn't be up before the beak if it had killed someone.

tim

Dont you just love it when some tosser pipes up and tells you that you are wrong. Tim (nice but dim) again too.

I have been in the avionics industry for years. I have met dozens of people who have no qualifications at all in software. Senior engineers with degrees in all sorts of subjects, but no software quals. Lots of engineers with loads of quals and experience in electronics who learned to do C when soemone suggested using software to do hardware tasks. Self taught. I have sorted thousands of lines of code from these bods. Massive monolithic code structures. Discrete numbers in the code. References to board addresses in the code. No concept of software engineering.
BAe Systems, EADS and others. Not fly by nights.

[Francko]

Dont you just love it when some tosser pipes up and tells you that you are wrong. Tim (nice but dim) again too.

I have been in the avionics industry for years. I have met dozens of people who have no qualifications at all in software. Senior engineers with degrees in all sorts of subjects, but no software quals. Lots of engineers with loads of quals and experience in electronics who learned to do C when soemone suggested using software to do hardware tasks. Self taught. I have sorted thousands of lines of code from these bods. Massive monolithic code structures. Discrete numbers in the code. References to board addresses in the code. No concept of software engineering.
BAe Systems, EADS and others. Not fly by nights.

LG, do you really believe that a degree can make that much difference? And I am speaking as someone with a formal degree in IT. Yet I realised that most of the things I have studied I have to re-learn them again in the proper way. I have met people with no degree being excellent at their job and people with a degree of no value at all (surely the terms can be inverted, it's not all black and white).

What we might miss is, perhaps, a formal exam to assert whether the person is qualified to do the job (despite having or not having a degree). However, up to today no organisation has ever managed to achieve that. Perhaps because of the multiple proprietary technologies but it's hard to say who is a qualified IT consultant. And no, I think that discriminating only on the basis of the degree is totally unfair and will screen out many excellent people out of IT (and don't forget that academic studies are normally well behind the real world in this field). There are certifications, no one cares, and even the British Computer Society tried to make a standard qualification but yet no success. I think it will never be possible to really achieve some level of formal qualification standard in IT. So I wouldn't spit so much on self-learning skills as I think they are definitely the most important to have.*

*Despite all of this I believe too that are too many "cowboys", just wanted to share my doubts about formal qualifications...

[NickFitz]

The problem is that software is such a new and strange thing that it's very hard to fit it into the same space as the engineering of physical stuff. I personally (although I have held the title of "software engineer") don't believe that software development equates in any way with the scope traditionally attributed to engineers.

If I am responsible for engineering a bridge, there are certain fundamental and well described principles that I can reasonably be expected to be knowledgeable in, and there are (for example) clearly understood data relating to the load bearing capabilities of various materials under varying circumstances that can be seen to be matters of which I ought to be aware.

If I am responsible for engineering a jet fighter aircraft, I have to deal with pretty much the same information as the engineer building the bridge - different materials, different circumstances, but pretty much the same stuff.

Software? We make it up. It is such stuff as thoughts are made of. This is its wonder, and its - well, not failing, but the reason why we are foolhardy to put much faith in it.

Granted, one ought to be aware of the order of (for example) sorting algorithms, but it's entirely possible that a quicksort will be ideal every time in ten million - yet then the one time in one hundred thousand billion turns up when a bubble sort would have been not only better, but necessary in order to get things sorted (pun intended) in the necessary time with the current dataset.

We should stop expecting software development to be as predictable as a physical engineering problem, and accept all software solutions as having all the flaws of a human solution: it will probably be great most of the time, but when it loses it, it loses it big time.

Granted, it's possible to create mathematically-provably-correct software... but that only works if it runs on always-correct hardware. Given that a cosmic ray in the wrong place at the wrong time can flip a bit in your RAM, it's probably best not to rely on it.

After all, even "real" engineers don't always get it just right

[HairyArsedBloke]

Whenever I hear someone claim to be a ‘software engineer’ I think of wet behind the ears fresh graduates or self-righteous Indians (which is often the same thing) with delusions of adequacy.

[BrilloPad]

God help anyone who puts faith in software built on Windows.

[The Lone Gunman]

You are all missing the point.
I am not as concerned with the qualifications as a I am with the quality of the engineers. Self taught is not always bad.
The fact is that there is no standard level of knowledge or achievement required to develop safety critical software.
Gas fitter have to be corgi registered.
Electrical instalations have to be done or inspected by qualified people.
Accountants, Surveyors etc have to be chartered.... Bar exams, doctors blah blah.
Some of the above quals evolve and practitioners have to keep up.

There are certain basic minimums I would expect to see in safety critical code.
I would expect practitioners to be aware of those basic minimums.

The BCS chartered status may be a good standard to start with, but while there is no requirement to hold chartered status why would anyone bother?
How about if IR35 didnt apply to chartered engineers, would that make it worth having?