Comment by hotsauceror | Hacker News Reader

Comment by hotsauceror | Hacker News Reader

My dad subscribed to one of the hobbyist computer magazines back in the 70s and 80s, I think it was Creative Computing. One summer, they printed out the BASIC source code of Colossal Cave, I don't remember which platform though. Probably a PET or a TI-99.

Our whole family took turns typing in the source code on our IBM PC XT, it took us a couple of weeks in our spare time, and with no one knowing how to type besides mom and dad. Dad showed me how to make the necessary tweaks to get it to run on our flavor of BASIC, and then we played it all summer.

I definitely felt something click into place, after that experience. I ended up doing the same thing years later on my Commodore 64, manually typing in the assembly code that they printed out in Compute!, so that I could play various homegrown games. You could subscribe to the magazine and get all the monthly featured programs on floppy disk, but that was for Rich People, not peasants like us.

You talk about mind-numbing... A9 00 C0 A0 D4 01 03 05 0A! for pages and pages and pages.

I got my start writing compilers from when BYTE published the complete source code to a Pascal compiler & interpreter.

My parents bought me an early Commodore 64 for Christmas in 1982 and I spent untold days typing in source code from Compute!'s Gazette which was the Commodore-dedicated companion to Compute!.

Their MLX system for entering assembly language programs with a checksum was an amazing way to enter programs from a magazine. I remember enlisting my sister to help me type in the programs and I told her, "Type the line of numbers here. If you type one wrong, it will tell you." She quickly retorted, "If it know what's write, why doesn't the computer type it in!?"

You talk about mind-numbing... A9 00 C0 A0 D4 01 03 05 0A

Apparently I did that too much because I can convert that to 6502 Assembly language in my head.

Yes! Compute!'s Gazette, that was it! Every month users would submit their homegrown works, like "I made a game that can display 16 sprites at once instead of 8 because of this clever IRQ trick" and it would contain the full assembly code, along with their checksum like you said. Compute!'s Gazette.

Actually got some submissions to COMPUTE!'s Gazette accepted in high school (though I also kept all my rejection letters). By then, though, they were technically COMPUTE Gazette Edition after the General Media buyout (yes, Bob "Penthouse" Guccione's outfit).

A9 = LDA, right? :)

All that remains in me' noggin...

I am not sure about that D4 6502 op code

    echo 'A9 00 C0 A0 D4 01 03 05 0A' | mondump -r | l

    0300- A9 00     LDA #$00
    0302- C0 A0     CPY #$A0
    0304- D4        ???
    0305- 01 03     ORA ($03,X)
    0307- 05 0A     ORA $0A

using KEGS emulator with 65816 op codes

    call-151

    300:A9 00 C0 A0 D4 01 03 05 0A

    300L

    1=m   1=x   0=d   1=LCbank (0/1)
 
    00/0300: A9 00        LDA #00
    00/0302: C0 A0        CPY #A0
    00/0304: D4 01        PEI 01
    00/0306: 03 05        ORA 05,S
    00/0308: 0A           ASL

>>> You talk about mind-numbing... A9 00 C0 A0 D4 01 03 05 0A! for pages and pages and pages.

Man, seriously, can't even think what would happen if you made a typo and how you would debug it

There was some amount of "standardization" of checksum methods to help with debugging typos. For example, Nibble magazine used a couple of programs-- the one I remember using was Key Perfect-- that would scan your program in memory after you typed it in, and report a series of short hex checksums, one for each reasonably small chunk of program (bytes for machine language, or lines of BASIC code). The magazine article would include the correct Key Perfect checksums for the program; compare against the generated report, and a difference meant you "only" had to scan a dozen lines of source instead of hundreds.

Worse, imagine going through all that, with no typos, and it still not working.

Friend and I paired up doing that once. Boy, I tell ya, you don’t have to tell us 4 times the code won’t work.

Oh, and “3rd times a charm” is a lie.

>Boy, I tell ya, you don’t have to tell us 4 times the code won’t work.

this might be the first time i've seen this, but love it

A lot of machine code listings had per-line checksums which could be checked with a small auxiliary program. PC Magazine had assembly source for its free utilities, alongside a BASIC program that would produce the utility from machine code encoded as hexadecimal numbers in DATA statements; the BASIC program also included checksums and would check the machine code as it ran.

As others have mentioned, the publishers had developed and distributed a checksum tool, written in BASIC, that you could type in from scratch as well. Although I think I copied mine from a friend who did subscribe to the monthly diskette version.

So you couldn't enter a line wrong, it would get flagged right away.

As far as debugging the whole thing, I think they just published a dump of the known-good copy they received with the user's submission. I think you had to provide two separate, working copies of your program along with the source code?

i can remember what would happen. of course, it depends on where the bug was and what it was for. i remember the bits of hex being proceeded by the DATA keyword. in one instance, the DATA was for the audio, and it played a horrendous noise. as the multiple typos that were present were fixed, the audio went from horrendous to glitchy. some of the DATA was for the graphics, and they'd present as glitches as well.

a friend of mine and i "developed" a coding technique (we were 10 years old) where i would type in the initial load (i seemed to be a better/faster hunt&peck typist). after that, we'd switch. i'd read the DATA hex, and he'd follow along on the screen and make corrections. it wasn't until many years later before i realized that the coding technique wasn't an original something that we "developed".

Reply on news.ycombinator.com