The software development process.

But better is to...

1. Write code
2. Read code
3. Test code
4. Release

but most people skip 2, sometimes 3, and even 4, in the sense that they don't formally release it, they just sort of set it free in the wild.

But the fastest way to get it working is to slow down and do it right.

John

In message , dated Sun, 10 Sep 2006, John Larkin writes

I agree, but why are programming languages (except vintage BASIC) written so as to be nearly (or quite) unreadable by humans?

Such dangerously revolutionary ideas should be banned from Usenet.

--
OOO - Own Opinions Only. Try www.jmwa.demon.co.uk and www.isce.org.uk
There are benefits from being irrational - just ask the square root of 2.
John Woodgate, J M Woodgate and Associates, Rayleigh, Essex UK

Amen! Quote for the week!

My other advice is not to disparage error-detecting tools (of any kind) or error-preventing disciplines as "training wheels." Too many programmers want to ride without training wheels or without a helmet. But surely the one thing every programmer knows is his own fallibility.

And the code the release shows the brain damage from falling on their head too many times. ;-)

--
Service to my country? Been there, Done that, and I\'ve got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

You refer of course to write-only languages like C++. It's because the practitioners enjoy the cryptic-ness of it all, and are usually inept at the use of real (human) languages. Posters to this group have noted that, if they take over someone else's C code, they first run it through a comment stripper.

I like to program in PowerBasic for PC-type stuff (engineering apps and test software) and assembly for embedded stuff, both heavily (and, I hope, literately) commented.

John

.SBTTL . MACRO : GEAR

; SIMULATE A GEAR PATTERN WITH POSSIBLE SHORT TEETH ; ; MACRO PARAMS ARE... ; ; 0 CHANNEL NUMBER ; 1 START POINTER ; 2 NUMBER OF PULSES 'N' 1..512 ; 3 DC BASELINE ; 4 WAVESHAPE (NOT IMPLEMENTED YET) ; 5 TOOTH WIDTH ; 6 DEFAULT TOOTH HEIGHT ; ; 7 SHORT PULSE INDEX 1..N } REPEAT, UP TO 16 ; 8 SHORT PULSE HEIGHT } EXCEPTION TEETH

; A0 SWEEPS PARAMETERS ; A1 AIMS AT PTRn REGISTER ; A2 AIMS AT WAVn ; D0 PULSE LEVEL ; D1 IS BLOCK START POINTER ; D2 PULSE POSITION, NORMALIZED ; D4 TOOTH WIDTH, AS DBF ; D5 TOOTH INDEX, 0..N-1 ; D6 TOOTH COUNT N

GEAR: MOVE.W MP0.W, D0 ; FETCH CHANNEL NUMBER JSR MAPP.W ; AND MAP SOME STUFF TST.W D7 ; DID THAT WORK? BNE BADMAC ; NOPE.

LEA.L WAVX(A1), A2 ; AIM A2 AT WAVn REGISTER LEA.L PTRX(A1), A1 ; AND A1 AT PTRn

MOVE.W # 0300h, D7 ; ANTICIPATE ERROR CODE 3 ; == ILLEGAL TOOTH COUNT ==

MOVE.W MP2.W, D6 ; EVAL NUMBER OF TEETH: BEQ BADMAC ; 0 IS SILLY, FOLKS. CMPI.W # 512, D6 ; CHECK LEGALITY BHI BADMAC ; BAIL IF TOO MANY

; FIRST, FILL THE WAVE BLOCK WITH THE BASELINE VALUE

MOVE.W MP1.W, D1 ; GET START LOCATION MOVE.W D1, (A1) ; POKE FPGA 'START' POINTER MOVE.W BSIZ.W, D7 ; AND NAB BLOCKSIZE SUBQ.W # 1, D7 ; -1 FOR SILLY 'DBF' LOGIC ; WHICH *DOES* WORK OK FOR 65536! MOVE.W MP3.W, D0 ; FETCH THE GEAR ROOT LEVEL  GROOT: MOVE.W D0, (A2) ; POKE BASE VALUE INTO RAM DBF D7, GROOT ; FOR THE WHOLE BLOCK

; NOW LOAD THE BASIC GEARTOOTH PATTERN

CLR.L D5 ; START WITH TOOTH # 0 MOVE.W MP6.W, D0 ; AT THE DEFAULT TOOTH HEIGHT MOVE.W MP5.W, D4 ; GET NORMALIZED TOOTH WIDTH MOVE.W DNORM.W, D7 ; AND DENORM FACTOR BEQ.S BEER ; 0 MEANS DON'T SHIFT! LSR.W D7, D4 ; ADJUST WIDTH TO ACTUAL BLOCK SIZE BEER: TST.W D4 ; DID WE CRUSH WIDTH TO ZERO? BEQ.S NORTH ; DBF COUNT OF ZERO MAKES WIDTH 1! SUBQ.W # 1, D4 ; OR ELSE USE REQUESTED WIDTH, DBF FORM.

NORTH: JSR TOOTH.W ; CREATE ONE TOOTH ADDQ.W # 1, D5 ; COUNT IT DONE CMP.W D6, D5 ; DONE ALL, 0 TO N-1? BLO.S NORTH ; HANG IN THERE, KID.

; NOW DO THE EXCEPTIONS: FOOLISHLY, WE IDENTIFIED THE EXCEPTION ; TEETH AS 1..N, SO WE'LL HAVE TO ATONE FOR THAT.

MOVEA.W # MP7, A0 ; AIM AT FIRST POSSIBLE EXCEPTION ENTRY MOVE.W # 16-1, D2 ; ALLOW UP TO 16 EXCEPTION TEETH

TOPIX: MOVE.W (A0)+, D5 ; NAB THE TOOTH INDEX BEQ.S GROUT ; AN INDEX OF 0 ENDS THE LIST CMP.W D6, D5 ; IF INDEX IS SILLY, BHI.S GROUT ; WE'LL QUIT AS WELL.

SUBQ.W # 1, D5 ; OK, ADJUST THE INDEX. MOVE.W (A0)+, D0 ; FETCH THE 'SHORT' PULSE HEIGHT JSR TOOTH.W ; AND OVERWRITE STANDARD GEAR TOOTH DBF D2, TOPIX ; DO ANOTHER TOOTH, MAYBE.

GROUT: JMP OKAY.W ; DONE THE LAST OF THE MEAN MOLDY MACROS!

Don't I wish. The only way to deliver good code is to assume that the author is a sloppy idiot whose work must be kept as simple as possible and checked at least twice.

John

You seem to have missed the most important step

1. PLAN the code.

"Those who fail to plan, plan to fail".

"Every failed project goes wrong in the first two weeks. After that, it is a losing battle".

Usually because no humans are involved.

Bill Gates' first 8080 Basic was somewhat bug ridden - but worked well enough. I once saw a reproduction of the first page - it was almost all comments. This is the sign of someone who knows what he is doing. I wonder what ever happened to him?

There is nothing wrong with most languages. There is plenty wrong with the people and the process used.

In message , dated Sun, 10 Sep 2006, John Larkin writes

Yes.

How is that different from an ordinary stripper?

I can ONLY do BASIC, and now you wrote that I'm not ashamed to admit it. (;-)

It wouldn't win a pullet surprise (not an original JMW; a US schoolgirl, I believe), but it is comprehensible.

--
OOO - Own Opinions Only. Try www.jmwa.demon.co.uk and www.isce.org.uk
There are benefits from being irrational - just ask the square root of 2.
John Woodgate, J M Woodgate and Associates, Rayleigh, Essex UK

I program bottom-up, so the planning happens continuously during the coding, not really before it begins. This only works if one is ruthless about re-writing anything that's not perfect, and if step (2) above results in substantial cleanup.

John

In message , dated Sun, 10 Sep 2006, Homer J Simpson writes

If every program has to start with:

main{

which does nothing, why bother with it?

--
OOO - Own Opinions Only. Try www.jmwa.demon.co.uk and www.isce.org.uk
There are benefits from being irrational - just ask the square root of 2.
John Woodgate, J M Woodgate and Associates, Rayleigh, Essex UK

Vintage BASIC is readable? Single letter variables and legions of gotos. Nearly all programs I've seen written in vintage BASIC has been excreable. Quite unreadable enough to be contenders in an obbsfucated code contest.

Microsoft's and other vendors more recent structured variations are much better having borrowed from other structured languages.

Robert

I have my own little mental collection of massive program failures but I wish there was a website to post the all on like

- sort of ?

perhaps if the FBI had read up on failures first?

I note that it's almost always large, complex database programs that fail.

Basic is cheaper to code and cheaper to maintain. You need a good reason to use another language.

Gee, I've picked up 20+ year old basic programs that I wrote and I've had no trouble reading or modifying them.

--
Service to my country? Been there, Done that, and I\'ve got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

PowerBasic is great. It's a serious compiler that lets you do almost anything that the OS allows.

John

It's the "subroutine" that the OS "calls" to run the program. It sort of means START HERE. The Windows (32 bit) versions of PowerBasic all start with PBmain() as a declared subroutine, so that poor little Windows doesn't get confused as to what we want it to do.

John

So we know where the program is supposed to start?

Remember non-trivial programs are made from many different source files. Also the parameters to main() can be used to access the parameters given by the user on the command line.

--

John Devereux

There are several "modern" BASICs around, and they can be quite readable. But to be fair, you should compare GWBasic to 8088 assembly, as they were contemporaries.

John

' ' BRICK.BAS : ROM-IMAGE FILE BUILDER ' ' by John Larkin ' Highland Technology, Inc ' 320 Judah Street ' San Francisco '

' ' November 9, 2000 ' ' Copyright (c) 1999 Highland Technology, Inc. '

' THIS PROGRAM WILL MAKE A ROM IMAGE FILE FROM MOTOROLA 'S28' ' FILES AND XILINX FPGA 'RBT' FILES. ' ' THIS VERSION CAN BUILD A ROM OF UP TO 4 MBITS (524288 BYTES). ' WE USE COMMAND-LINE ARGUMENTS TO SPECIFY BUILD PARAMETERS, ' AND CREATE LOG FILE BRICK.LOG TO DOCUMENT THE BUILD. ' ' COMMAND LINE SYNTAX IS DESCRIBED IN THE 'HELP' PROMPT BELOW.

' BRICK IS CODED FOR PowerBasic/CCv2.

' NOTE : NOV 9, 2000 VERSION FIXED CHECKSUM CALCULATION ERROR

#COMPILE EXE ' AS OPPOSED TO... WHAT?

#OPTION VERSION4 ' COMPILE FOR Hairball95

DEFLNG A-Z ' DEFAULT 32-BIT VARIABLES

FUNCTION PBMAIN() AS LONG

DIM ROM(524287) AS BYTE ' IMAGE WORKSPACE, 4194304 BITS

DIM H(100) ' HEX VALUES FROM INPUT STRING

DIM SP AS STRING PTR * 256 ' USED TO GOBBLE ROM() FOR OUTPUT

K256 = 256 ' A SCALER K64 = 65536 ' DITTO

COLOR 15, 0 ' TURN UP THE LIGHTS!

'

' INPUT THE FILENAMES FROM THE COMMAND LINE...

Y$ = COMMAND$ ' TRY COMMAND LINE FIRST

IF Y$ = "" THEN ' NONESUCH? ALLOW MANUAL ENTRY

PRINT PRINT " BRICK.EXE: Rom image builder Copyright (c) 1999 Highland Technology Inc" PRINT PRINT "Enter a command line as..." PRINT PRINT "OUTFIL.ROM/256 = MACFIL.S28 FPGA1.RBT/aaaa FPGA2.RBT/bbbb /C:cccc" PRINT PRINT " where..." PRINT PRINT " OUTFIL.ROM ROM-image output file. Specify ROM size" PRINT " as /256, /512, /1M, /2M, or /4M" PRINT PRINT " MACFILE.S28 68332 uP code in Motorola S28 format. Filename" PRINT " extension must be .S28 or .HEX. PRINT PRINT " FPGAn.RBT FPGA config file(s) in Xilinx RBT format;" PRINT " ASCII header loads at aaaa, config data begins" PRINT " at aaaa+100h. File extension must be .RBT. PRINT PRINT " /C:cccc Specifies address of optional 16-bit checksum" PRINT PRINT " /P:dddd@aaaa Pokes 16-bit value dddd at address aaaa" PRINT

LINE INPUT "BRICK > " Y$ PRINT

END IF '

Y$ = UCASE$(Y$) ' FORCE UPPERCASE IF Y$ = "" THEN GOTO XNULL ' NO COMMAND? JUST EXIT

' OPEN AND BEGIN THE LOG FILE

OPEN "BRICK.LOG" FOR OUTPUT AS # 7

PRINT #7, " BRICK.EXE ROM BUILD LOG FILE "; DATE$; " "; TIME$ PRINT #7, PRINT #7,