student project

now

k,

up

=A0... no comments, insane logical

I had to debug some code where the comments were in Finnish. I found out after researching the language a bit that Finnish is really rather simple --- if you already know Estonian.

te:

ur

er!)

sh.

=A0 =A0 ... no comments, insane logical

ed

me

ts

able

want...

A logical design process starts with requirements, which can exist in a variety of forms, and then a formal specification, which, if everyone does not agree on it, at least serves as a record of what the disagreements are.

Of course, in the background some lab/code tinkerer is often lashing together a prototype based only on what he thinks the final product should be.

are

es

'd.

I heard a VP of Engineering say almost that - once we win the contract, then we can write the real spec and rebid everything.

programmable logic,

Every two weeks you had a board spin? Now *that's* productive!

How did you know that you were done with the design?

saved.

Right, no more meetings. ;-)

switches...

happened.

I assume the place folded.

Yeah, but that's a dangerous thing to tell the peons.

=A0 =A0(Richard Feynman)

The most beneficial element of the original C++ release was the // comment delimiter.

I have two, one at work and one at home (in the guest bathroom!) Epson LQ-2090s, beautiful gadgets. Before I ever run a piece of code, I print and read/tweak it several times. Before I release it, I read it again. I deliver embedded code about 5x or 10x faster than the average programmer, seldom more than 2 weeks per project, because I spend very little time debugging.

All my programs begin with a table of contents, and several pages of introduction/version control/change log.

check

If they finish soon, and leave good documented code, why not?

Two is the minimum. This also helps with checking code since

We hired a contract programmer to rewrite out in-house parts database system, the original of which I did about 15 years ago. He was a typical programmer, no social skills, typed real fast, made bugs faster. The Brat rode herd on him: told him exactly what to do every day, tested the results every day. He soon got trained to please her. It worked great. So all you need is one very good manager per programmer.

ftp://jjlarkin.lmi.net/MAX.jpg

John

Phil Hobbs expounded in news:V7udnXy-tbeEMnTRnZ2dnUVZ snipped-for-privacy@supernews.com:

I only like REXX when I have to do something on the mainframe. I do like the fact that I can usually test most of the REXX script on an open system platform.

But for pure open systems work, script wise, a combination of bash/ksh and awk is all I usually need.

Warren

Joel Koltner expounded in news:VsCFo.374028$ snipped-for-privacy@en-nntp-12.dc.easynews.com:

It?s much easier to get programmers to use a feature with a simple compiler switch, than it is to get them to code the safety into the code. The best part is that the compiler won?t miss a thing and will always get it right.

Warren

John Larkin expounded in news: snipped-for-privacy@4ax.com:

But nobody objects to C on that basis. But you hear it a lot when you mention Ada.

Warren

Nico Coesel expounded in news: snipped-for-privacy@news.kpnplanet.nl:

Availability of an Ada compiler was a big impediment to the language for a long time. GNAT?s been around for a while (early 90?s?), but with Adacore?s integration of GNAT into the gcc project has really been a big boost. Now Ada is available almost on any platform now, just by an install. It used to be quite the process to build GNAT from sources. But with its inclusion in gcc, it is very available now.

Here is what I used for AVR (atmega/arduino) projects:

If you want to purchase support, you can now get it from Adacore as well:

They now support the AVR platform with GPS 5.0.

Warren

You make it sound like she dresses in black leather & six inch stilettos, and carries a whip that she isn't afraid to use! ;-)

Interestingly there seems to be no ARM port available.

Why's that? If the guy is good -- commenting his code, using decent algorithms, not using spaghetti structures (all the stuff we've been discussing here) --, if he is run over by a bus it shouldn't be particularly difficult at all for the next skilled programmed to pick up the pieces and keep going. I realize there's a transition time there, but I almost think I'd rather go with that approach than having two guys on small projects that are easily doable by one guy.

Nico Coesel expounded in news: snipped-for-privacy@news.kpnplanet.nl:

I believe there is a separate project for that but I'm not familiar with it. I'm guessing it is related to a AVR-32 port somewhere.

Warren

Pelosi is Larkin's brat?

Two projects done by two guys? I think the point is, though, that "skilled programmers" are hard to come by. Half of the above problem can be solved by paying the "skilled programmer" what he's worth and stop treating him like dirt. IOW, pay him enough to drive his own car. ;-)

Agreed. With programming and often engineering, I think it's often the case that one really good guy can often easily be as productive as two mediocre guys... and the really good guy will be quite happy with a salary that's still noticeably less than twice what you're paying each mediocre guy. Hence, both the really good employee and the company win!

But as you say, finding those really good guys is not always easy...

No problem, but the car will have to be a Tata Nano.