Are programmers engineers?

There are some places where calling oneself "engineer" is legally controlled. I think a couple of US states restrict the usage, but it's not much enforced.

So "engineering" describes a vague range of activities.

I don't think I can become a PE in any state for a couple hundred dollars. There's plenty of bogus print-it-yourself certs I could get dealing with software, systems or networking for that price.

I've also not met any civil engineers that had stories about how all the stuff they put their stamped on collapsed or failed time after time.

What I'm not really clear on electrical engineering and what legal responsibilities are involved there though.

For example, there's plenty of garbage designed electronicy stuff out there somebody came up with, and even if it's horrible, it doesn't matter.

What about more critical stuff like airport runway lighting systems? EEs must be involved and I can't imagine these designs don't get approved by somebody with a stamp, like in civil engineering. What't the title or role of who signs off on things like that?

After Fukushima, I would not assume any engineer would solely sign such documents.

The problem with Fukushima was that the design assumption was that there might be individual diesel failures in the emergency cooling system, but when the tsunami took out all diesel generators ...

AIUI, it is enforced in Texas. I don't know of another state, though.

I think that's been pretty well established, here. ;-)

It doesn't cost that much. The tests have little to do with reality, is the point. At least when I looked into the exams, they didn't even have any electrical questions on them. What's the point?

You've got to be kidding. First, how many electrical engineers sit around bragging about all the stuff they've designed killed people. Second, have you ever done any research into engineering disasters? There certainly are enough to go around.

Hire a lawyer.

Does it involve safety?

Likely a civil engineer, though the law makes no distinction. A PE is a PE.

I guess that I am quite old fashioned since I think that safety critical systems should be implemented using simple laws of nature, such as gravity or the use of springs, not using some sequential or programmable logic. However, now it seems to be acceptable to use some programmable logic to do that and various systems such as SIL

are used to justify that usage. I haven't heard of any certifications required to program these systems.

Springs and wedges aren't going to help land a 777.

I think you'll find certifications in all of these fields, not so much on the people but corporate lawyers and insurance companies are going to have a much bigger real impact than any government intervention.

In California, you need a license to cut someone's hair, but not to design electronics.

Hardware engineers need to make a product work in a year long test cycle. S oftware programmers, can make mistakes every day and fix some and ship it. Sloppy programming is allowed because of the fast time to fix it. Engineers who are sloppy lose money for the company for years after each error. Veri fication and rigor are important to engineers. Programmers do not have thos e limitations, they do not have rigor, just slop.

Not so much. It's essentially an inability or unwillingness to properly measure and estimate the costs of defects.

But this money goes unmeasured. So it doesn't exist. Therefore, defects are "free".

Speak for yourself. My code works, to the limit of the team's ability to specify the problem.

--
Les Cargill

So, because bugs aren't specified, they don't exist? ;-)

Seriously, this is one of the biggest problems I see. Rarely is anything specified such that it's clear what's required. When I was in IBM, specifications were very good (there is a reason that programmer productivity was set at one to three LOCs a day). It's the only way to build complex chips or operating systems. Since, I've been in situations where there were no specifications at all. Sometimes even the features list was in some email trail. Somewhere. After we were done, management brought out the "well, that's not what we meant" crap.

Provably correct programming starts with an unambiguous and demonstrably specification of what the complete program is required to do.

This dates from 2001, but is an example of the approach.

--
Bill Sloman, Sydney

That's why a prototype can be vital. A prototype with all the options encoded into it and a quick way to switch them keeps the team bandwidth channel filled.

The way to build complex thing is to build them just like the last one only with this small difference.

I won't speak for having to code up driver after driver but a *basic* O/S is a student-scale project.

Just one driver - like for a complex graphics card - may never be actually finished.

Yep.

You have to learn how to ... interrogate people.

--
Les Cargill

Sure, in fact that's pretty much what my job is now. OTOH, few companies have the money or foresight to do real prototypes. At the PPoE, they wanted everything done with one pass but didn't want to specify anything. When that didn't work, they'd hurry out fixes on top of fixes, without testing the first fix. It's surprising that anything worked. It was a good thing the owner had a decent source of money.

Evolution is good, except when it isn't. If evolution was the only mode, we'd still be typing punch cards, if perhaps virtual.

Basic OSs don't make money, either.

Sure it will be. Sooner or later there will be a new graphics card. ;-)

The problem was that they didn't want to admit that they hadn't considered all the possibilities. When the issue arose, it was the designer's error. That's when we started writing specifications before we started but management wouldn't read them, much less sign off on them.

What, you aren't?

Cheers

Phil Hobbs

Punched cards were an immense improvement over paper tape. You could edit a program by shuffling a tray of cards. You could color-code the edges and see the program structure. You could copy subroutines and keep a library of good stuff around.

I interfaced a number of card readers to a PDP-11, and interfaced an IBM 029 card punch. And I patched the FOCAL language to read cards.

That was all cool, but really primitive. Piece-o-junk that it is, Windows with a terabyte of hard drive is better.

and they you are wrong. Software Engineering has been a professional engineering discipline for some time now. Course are accredited by the Washington Accord. Not so for computer scientists of course. Likewise electronic engineers and even mechatronics are acccredited.

e

r

o

es.

and you talk rubbish. Although maybe a large proportion (not all) of electr onic engineers don't bother to register themselves as they have no need, th at is not to say that they couldn't. Most Uni course are accredited by the Washington accord (in Western countries that is). It is perhaps exactly why engineers are often looked upon with scorn that they do not in greater num bers join a professional institution to protect their interests and promote the profession - CPD etc

Been there, done that. I had an undergraduate research assistantship in

1979-80 working for a millimetre-wave radio astronomer (Prof. Bill Shuter, may his tribe increase). He handed me a hardcopy Fortran-77 listing of a Monte Carlo radiative transfer code from some Danish observatory, and asked me to get it working on UBC's Amdahl 470-V8 machine and do some simulations. (You can do radio astronomy through clouds--probably nobody since Tycho Brahe has tried to do optical astronomy in Denmark, or Vancouver for that matter.)

I learned FORTRAN by porting and debugging somebody else's grad student code, which was a slice. Wound up working well, though, and produced some very interesting results about carbon monoxide emission line shapes in interstellar giant molecular clouds. (AFAIK it was the first time that anybody showed that you could get self-reversal of emission lines from a cloud not undergoing inhomogeneous collapse. Self-reversal is when absorption in colder outer layers knocks out most of the central peak but leaving the wings, so that what is nominally a single emission feature has a three-pointed shape.)

I couldn't type back then, so I handed the listing to the last surviving keypunch girl, who typed it onto a card deck. (I thought card decks were cool, and knew that they'd be going away soon.) I did all my editing and so on in a room full of 3270 terminals in the basement of Hennings (the main physics building).

Yeah, but the chrome bling on the card verifier was awesome.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net