Test Equipment Anonymous (TEA) group therapy thread

[GregDunn]

Over my career from college through systems engineering through software development I think I've probably learned over 30 programming languages.  BASIC is fun to use, but you can't really extend the language without moving to a compiled version - which negates most of the instant feedback you get from an interpreter.  If I was after readability + speed, I always ended up using C (not C++, C# or the other horrid encumbered variations).  As part of my last job, I was essentially forced to program in Java; over the 18 years I used it, I developed a loathing for it that is hard to communicate in a few sentences. 

When I discovered Python, it was almost like that cartoon - a language which was powerful and yet reasonably efficient, and you weren't forced to use unnecessary layers of objects for simple things.  I found excuses to use it to replace Java everywhere I could.  I even prefer it to shell scripting if it's installed on a machine I use.

Yes, the "Hello World" program has always been a sharp test of how much overhead is required by a programming language; if you can perform it in a handful of machine language steps it shouldn't require a megabyte of Java libraries to do the same.

[bd139]

Python seems to be a good language for getting things done quickly and I've been intending to learn it for a long time now but I never get around to it.

But I think it's kind of ironic that python fans brag about being able to do "Hello World!" in one line.  Seems to me we could all do that back in the 1980's with any computer that booted right into BASIC, or any computer that had a BASIC interpreter.   

This is my favorite way to do a Hello World program:

Python is one of those things. It's like a good box of tools. But all the tools are from Star Trek TNG.

Big fan of debug.  MSDOS and me go a loooong way back. One of my prized posessions back from the distant past. IBM PC UG meeting in the late 1980s I snagged this just before heading off to university...

[Mr. Scram]

What kind of crummy foreword is that!?

[bd139]

Hahaha 

[GregDunn]

I once found this quote in a book on Digital Electronics:

Nothing Dignifies labor so much as the saving of it.

A rule I try to live by. 

We have gotten so far away from that in today's software.  I feel that a large part of the race toward faster machines with more memory was a direct result of programmers' unwillingness to exert a little more effort when coding.  That also leads to multi-megabyte programs which fail in subtle ways and require multi-megabyte monitoring and debugging suites to isolate (a large part of my job in the final years before I retired).  Any labor saved by bodging together huge libraries instead of writing more efficient programs is obliterated by the outages and recovery efforts resulting from that mass of ugly code.  Now that we've hit the wall and Moore's Law is no longer a valid guide, people are going to have to learn how to program smarter instead of just faster again.  You can't simply multi-thread an existing program and expect it to run faster.

And now, my Digi-Key order has arrived and I need to hunker down in the lab for a while.

[bd139]

Incidentally you get paid better to handle this crap now though. The binary payload on the last .net thing I worked on was about 350mb (!)

Doug stuff properly means knowing what you’re doing which is in short supply.

[Specmaster]

Whats wrong with Sinclair basic?

10 PRINT "Hello World"
20 GOTO 10

That bloody simple. 

[bd139]

Mainly the shitty keyboard you have to use to get it in

[Specmaster]

Not on my Spectrum, it has a normal keyboard

[bd139]

Waste of a good keyboard 

[GregDunn]

Incidentally you get paid better to handle this crap now though. The binary payload on the last .net thing I worked on was about 350mb (!)

Doug stuff properly means knowing what you’re doing which is in short supply.

I must admit we were paid well to do it; unfortunately, AT&T loves to cut corners up front and then pay out the arse to keep dodgy software working later.  Most of the sw "engineers" we worked with had no clue what debugging was.  They just slapped the code together and did no regression testing or trace analysis before dropping it into production - and remember, this code all directly or indirectly interacts with the phone system.  A careless DBM in a hurry to start their holiday deleted the wrong partition during a backup and took down our 600TB (that's not a typo) database for most of a week.  They had never tested the backup/restore function so it was a huge surprise that nothing had been backed up properly in two months.  That was a very entertaining Thanksgiving weekend, let me tell you...

[bd139]

Unfortunately that hasn’t changed. I was shovelling similar smelling turds today

[Cerebus]

Today's XKCD:

[Cerebus]

...  Most of the sw "engineers" we worked with had no clue what debugging was.  ...

I'm of the opinion that there is no such thing as "software engineering" and am deeply suspicious of those who claim to be "engineers" in the realm of software. Some small part of the business of producing software is amenable to actual engineering, but most of it is not, and the activity of making it is a "craft" not engineering. Good software is craftsmanlike and the producers of the same are craftsmen and master craftsmen, titles anyone ought to be proud to be capable of wearing.

[bd139]

I don’t think it’s engineering or craft related. It’s neither. When even the best quality tools disintegrate in your hands, you’re no better than a monkey painting with its own excrement.

[Carl_Smith]

Yes, the "Hello World" program has always been a sharp test of how much overhead is required by a programming language; if you can perform it in a handful of machine language steps it shouldn't require a megabyte of Java libraries to do the same.

And that is why I like the "debug" version of the Hello World program.  It produces a .com file that is 28 bytes long.   

I remember way back when I first learned a little about Visual Basic, probably version 4.  I was appalled when I made a program that produced a dialog box with nothing but the text "Hello World!" and an OK button and the resulting .exe file wouldn't fit on a 1.44MB floppy disk. 

[Carl_Smith]

Today's XKCD:

Every XKCD comic has "mouse over" text with another bit to the joke, which doesn't show when you link to the image here.   This one's mouse over text says:

I'm working on assembling a combination declinometer, sclerometer, viscometer, aleurometer, stalagmometer, and hypsometer. I'm making good progress according to my ometerometer, a device which shows the rate at which I'm acquiring measurement devices.

[Specmaster]

...  Most of the sw "engineers" we worked with had no clue what debugging was.  ...

I'm of the opinion that there is no such thing as "software engineering" and am deeply suspicious of those who claim to be "engineers" in the realm of software. Some small part of the business of producing software is amenable to actual engineering, but most of it is not, and the activity of making it is a "craft" not engineering. Good software is craftsmanlike and the producers of the same are craftsmen and master craftsmen, titles anyone ought to be proud to be capable of wearing.

I agree with you, its not engineering, its a form of technical writing. You don't call an author of a book, no matter how important or big seller it is etc. an "engineer" of any sort, they are writers or authors so why should a person who writes software code be an engineer? 

[GregDunn]

...  Most of the sw "engineers" we worked with had no clue what debugging was.  ...

I'm of the opinion that there is no such thing as "software engineering" and am deeply suspicious of those who claim to be "engineers" in the realm of software. Some small part of the business of producing software is amenable to actual engineering, but most of it is not, and the activity of making it is a "craft" not engineering. Good software is craftsmanlike and the producers of the same are craftsmen and master craftsmen, titles anyone ought to be proud to be capable of wearing.

I agree with you, its not engineering, its a form of technical writing. You don't call an author of a book, no matter how important or big seller it is etc. an "engineer" of any sort, they are writers or authors so why should a person who writes software code be an engineer? 

I had one software job which did require engineering - we were writing the control code for a programmable airborne radar system.  It had to be able to talk flawlessly to the hardware at every level and be able to withstand the most egregious abuse from the end users.  The amount of regression testing and timing analysis which went into that code was extreme.  When you're approaching a target at 600 knots and tracking it in 2D, the last thing you want is for the UI to freeze or the tracking algorithm to lose its mind; you don't get a do-over.

The language we used, incidentally, was not Java, Ada, or C - it was PL/M, which unquestionably produced the most efficient code I've ever seen.  We actually inspected the compiler output on a couple of occasions in hope of improving a few things and finally gave up because none of us could write better machine code by hand.  That software was a wonder.

I ended up using PL/M to write a real-time AGO keyboard scanner with force and velocity detection, using a 12 MHz 8051 and very little added hardware.  It's still working well 30 years later...

[bd139]

That's the sort of job I'd have liked to have done if I'm honest.

Really I think modern programming is a result of the ability to quickly screw things up and then quickly unscrew them in a never ending cycle. People tread more carefully when they had an EPROM erase / burn cycle every time they did something.  Engineering has a long lead time between design and outcome thus the risk benefit is balanced differently.

Edit: most fun I ever had in this space was hand assembling Z80 code with a pencil and paper. Took forever but it worked every time as I was scared of getting it wrong.

[GreyWoolfe]

I love it when I am actually good for something instead of good for nothing!!

Hey what's the status with your meter?

It is supposed to be landing at ordinaryman1971's domicile today for calibration.  I included a return sticker to send it back.  I can't wait to have it back.  With the buttons replaced and feet added, it already has its spot ready on the bench.  That will be the first meter the DMM Check plus will attach itself to when it arrives next month.

[mnementh]

...  Most of the sw "engineers" we worked with had no clue what debugging was.  ...

I'm of the opinion that there is no such thing as "software engineering" and am deeply suspicious of those who claim to be "engineers" in the realm of software. Some small part of the business of producing software is amenable to actual engineering, but most of it is not, and the activity of making it is a "craft" not engineering. Good software is craftsmanlike and the producers of the same are craftsmen and master craftsmen, titles anyone ought to be proud to be capable of wearing.

The dingle who coined the phrase was probably a fucking MBA... their entire degree consists of learning the same hundred or so ways to make numbers tell lies that have been taught for generations, and the current buzzwords in use for those lies. Of course they wouldn't get why the two are an oxymoron. 

An engineer deals in absolutes and concrete (sometimes literally) concepts; what you put in your brain, what comes out of it and what goes down on paper are all representations of tangibles in the real world, and that translate directly to THINGS.

Programming is by definition almost entirely abstract, the polar opposite; even when you start out with something concrete, you disintegrate it into a collection of interdependent abstact objects that don't really exist anywhere until somebody makes the thing those abstracts represent.

Yes, there's a LOT of hard math involved... and that usually represents some real physics in the real world... but it's STILL ABSTRACT objects used to describe abstract concepts to represent reality. It is more a high-precision art than any kind of science.

*Currently overrun with two small children*

[GregDunn]

That's the sort of job I'd have liked to have done if I'm honest.

Really I think modern programming is a result of the ability to quickly screw things up and then quickly unscrew them in a never ending cycle. People tread more carefully when they had an EPROM erase / burn cycle every time they did something.  Engineering has a long lead time between design and outcome thus the risk benefit is balanced differently.

Edit: most fun I ever had in this space was hand assembling Z80 code with a pencil and paper. Took forever but it worked every time as I was scared of getting it wrong.

It was a lot more rewarding than what I ended up doing at AT&T; and yes, we had to mass-burn EPROMs every time there was a software change.  I believe there was just over a megabyte of code in that three-processor system; all 8086 mil parts running at a whopping 5 MHz.

Before we had a native assembler, a friend and I used to hand-assemble all our 6502 code; it was easy to test small routines that way.  Even today, he or I will msg one another with a string of hex bytes to make the other laugh, because we remember the story behind much of it.

[Mr. Scram]

I agree with you, its not engineering, its a form of technical writing. You don't call an author of a book, no matter how important or big seller it is etc. an "engineer" of any sort, they are writers or authors so why should a person who writes software code be an engineer? 

If writing code is technical writing then drawing up a schematic or blueprints with the correct numbers in the right places and testing it is just technical drawing. I'm not sure what would constitute an EE, because you just end up with people who draw and technicians.

[mnementh]

That's the sort of job I'd have liked to have done if I'm honest.

Really I think modern programming is a result of the ability to quickly screw things up and then quickly unscrew them in a never ending cycle. People tread more carefully when they had an EPROM erase / burn cycle every time they did something.  Engineering has a long lead time between design and outcome thus the risk benefit is balanced differently.

Edit: most fun I ever had in this space was hand assembling Z80 code with a pencil and paper. Took forever but it worked every time as I was scared of getting it wrong.

It was a lot more rewarding than what I ended up doing at AT&T; and yes, we had to mass-burn EPROMs every time there was a software change.  I believe there was just over a megabyte of code in that three-processor system; all 8086 mil parts running at a whopping 5 MHz.

Before we had a native assembler, a friend and I used to hand-assemble all our 6502 code; it was easy to test small routines that way.  Even today, he or I will msg one another with a string of hex bytes to make the other laugh, because we remember the story behind much of it.

I think you plucked a cosmic string with that one... I mean, I traded ASCII humor with friends back in the day, but jokes in hex?

That's like finding your girlfriend's phone number in your own DNA...



mnem