student project

Programming is no more and no less of art, then, say, a plumbing. Someone could find a fascination in the properly layed pipes, and I am sure there is a lot of sacred knowledge about it. The result of broken pipe could be catastrophic, too. But nobody talks about plumbing as a culture or as an art.

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

Plumbers make more money than doctors, and their liability insurance is cheap :-) ...Jim Thompson

The figure of merit is time spent coding divided by time spent debugging. I've known C programmers that ran in the 0.005 ballpark. Type code for 5 minutes, spend a week debugging. And, of course, not find all the bugs.

John

Huh yourself. My wife works with austistic kids. She can spot the trends in their fathers, who are often some sort of IT worker.

Not here, it isn't.

The flaw in that approach is that it's much faster to do it right.

John

SEMI-USELESS BRIGHTNESS FACTOR

same program later.

Well, PowerBasic has a nice set of PARSE functions.

Not sure what you mean by that.

John

On a sunny day (Thu, 18 Nov 2010 20:44:44 +0000 (UTC)) it happened Warren wrote in :

OK, but I do not understand all that stuff 'why better', I usually develop code in modules, and those modules are pretty much tested in real life and bug free. In this case I am re-using maybe 75% of old code, changing some parameters, removing some stuff that is not needed in this case, and then a new setup with some new code to be tested. All in PIC asm. For example all the LCD routines, timing, EEPROM related stuff, communication, has been tested in real life and I just re-use that code. Ada generated code probably would not fit in the 4 kB memory...

Yes you can write mathematical stuff to test for errors in software. But then that has its limitations too, and how is the correctness of that verified. I think for programming you need a logical and clear mind. Not everybody can do it, think that way. Maybe for those very high level languages, where you say something like 'computer do this' or computer do that' are the way to go. And even then they will mess things up: Defuse bomb manual:

Some people ARE dangerous

Here are so many things that can go wrong at so many levels, even the simplest PIC: earlier this year I tried SPI mode on a PICF14K22... it simply does not work (after hardware revision 4 IIRC still not). Glitches... you have to trust something at some point.

This is the whole way our physics works. We try some reaction 1000x, and then ASSUME it will always be that way. Then comes QED, and things at the quantum level are highly uncertain. They are more certain for macroscopic things, but not *absolute* certain. Same with software, the more complex it becomes, the more unexpected things will happen. Same with the ever more complicated processors, not all parameters are tested. What this says is that basically there is a limit on how much testing for software makes sense. Probabilities. We are, in fact, a neural net, and we can give the right answer most of the time, but not all of the time, more with more training. The world is probably quantum coupled, the whole universe, and us as part of that. No free will, things come from the past experiences, the present inputs, the structure of our brain (neural net), and the coupling with the rest of the universe, all that springing up from the subconscious and making us think 'Oh yes, I want this, I do this, I understand this'. Just a wheel in a machine we are, math, that you use to test your software, is just a small subset of neurons playing a simulation of reality that is in fact extremely limited. People think math is the answer to anything and everything. In my personal view math starts where you no longer understand what happens. It is a painful long road to solutions. As an example, consider somebody throwing you a ball, and you catching it. Now the neural net had no problem whatsoever to do that in real time. Take seals, those can catch a ball with their nose, and keep it there. Now let us take our 2 digital cameras, analyse the pictures at 30 fps in real time with that latest Nvidia based Chinese world speed wining supercomputer, calculate how and where to move your hands and fingers to catch the ball..... the equations, the processing power, the heat. We are just playing in the sand really. Sure the little gadgets we design can be very useful, I use those, but nature often was already there and did a better job. Maybe we can augment what nature does a bit. And as long as we have fun doing it that is just great. So much for todays philosophy, it was too late at night to fire up the scope again to see if my software changes worked... LOL

On a sunny day (Thu, 18 Nov 2010 13:58:12 -0800) it happened John Larkin wrote in :

Someone claims to be a great programer who writes code for 5 minutes in BASIC and then compains for 6 month in newsgroups about C programmers and C.

On a sunny day (Thu, 18 Nov 2010 14:03:15 -0800) it happened John Larkin wrote in :

You are mis-using statistics. As most people are -, or claim to be, or are thought to be 'some sort of IT worker', that statement is equvalent to: Most people that died today ate bread, so bread is very dangerous.

Autism is a politically correct word for anything from bad temper to imbecility (typical for IT workers).

That's why Bill Gates offers Windows 7 for $100, whereas John Larkin can only offer a piece of cable for that price.

The flaw in that approach is that you need many right people to do it right. Right people cost right money and aren't interchangeable. There are only few critical things that has to be done right and there are only so many right people.

VLV

On a sunny day (Thu, 18 Nov 2010 14:08:48 -0800) it happened John Larkin wrote in :

SEMI-USELESS BRIGHTNESS FACTOR

same program later.

Say you have started the program, and it is now doing X, with 20 different parameters you entered, 40 files open, all modified in some way, incuding screen formatting, selected forms, connections, and a power failure happens. Power comes back, PC reboots, program starts again, continue where you was? No files damaged? No data lost? No strange output to the network?

C is an outdated dialect of C++, beloved by some geezers.

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

My son, Aaron, is just the opposite, he's a commenting extremist. Gets all over my case if I don't comment virtually every line :-) I think he's writing predominately in C# right now. ...Jim Thompson

The UK formulation of that idea is "we never have time to do it right, so we always have to find time to do it again".

I had my nose rubbed in it with my advanced lectron beam tester project. We'd gone for a complicated electronics running fast (with GigaBit Logic's GaAs logic for the fastest bits).

Management wanted it yesterday. When I concocted the project plan I put in a couple of months for debugging, but noted that we'd recently had a couple of small projects that hadn't needed any debugging at all

- using computer-aided design and simulations run directly from the production working drawings. Both had been done by unusually competent engineers, and both had benefited from careful design reviews.

Nett result was - predictably - that the debugging time was cut, half the engineers were sub-contractors, and the designs were rushed to printed circuit layout without benefit of design review (to save time). Debugging was something of a nightmare. The high speed aspect of the design didn't give us any problems that we couldn't fix quickly, but the complicated sequencing turned out to have been poorly thought out and we reprogrammed a lot of 22V10s a lot of times before we got them doing what we needed them to do.

The fastest (multulayer) board turned out to have been built with its ground planes in the wrong places in the stack-up. My layout notes specified exactly what layer had to go where (and why) but the printed circuit department "knew" that the inner layers could be in any order.

Lots of high speed tracks had to metamorphose into lengths of minature coax before we could get the first version of the board to work. That didn't take long to do, once the penny had dropped.

It took about six weeks of me walking past the engineer testing the board and repeatedly pointing out to him that his stack of layer drawing was in the wrong order before he got the idea to take a drill and find out exactly how the board was actually built up.

-- Bill Sloman, Nijmegen

programming Language' by K&R.

consider BASIC again.

thinking about how good BASIC is.

As for the the time, it's an old story:

- When you are young, you like to do things and you have plenty of time but no money.

- Then you grow older and get some money and still like to do things but you have no time for it.

- .

As for the code, it's the kind of puzzle I like.

petrus bitbyter

"John Larkin" schreef in bericht news: snipped-for-privacy@4ax.com...

C is a programmers language so only programmers can value it :) (You can value the programmer by the way he is using it.)

petrus bitbyter

Can't say I've ever had a pencil that did any vague approximation of black.

The nearest hex equivalent might be around #A0A0A0. For sharp points, maybe as dark as #808080. That's against a background of around #E0E0E0 for correctly lit paper (give or take optical brighteners, which tend to tilt it towards #E8E0F0 or so).

I prefer pens over pencils. A good dark pen gets in the #606060 range easily, though I've seen too many that are as weak as pencil despite very hard pressure.

Your images are poorly lit, cutting everything pretty much in half. Pencil under those conditions is more like #606060 against an #808080 background. Terrible. At least learn photography and get out a flash or halogen light when you're going to snap something. And lots of post processing. At the very least: contrast, brightness/gamma, crop, resize. Optionally: color balance, filters (usually sharpen), geometry (stretch, skew, perspective, warp, etc.). Maybe some redraw and touchups with paintbrush, clone tool, that sort of thing.

There isn't any use releasing your pet project under GPL if no one can see what the hell it is.

All my (also hand-drawn!) computerized schematics are precisely #FFFFFF and #000000 for maximum contrast. They print quite nicely at 100 DPI, blockiness notwithstanding. My 300 DPI schematics are almost vector quality by eye.

Tim

"Jan Panteltje" schreef in bericht news:ic3ti4$vgi$ snipped-for-privacy@news.albasani.net...

FAIK C++ is never spoken.

petrus bitbyter