adventofcomp.bsky.social
A podcast about the history of computing
https://adventofcomputing.com/
162 posts
121 followers
100 following
Regular Contributor
Active Commenter
comment in response to
post
That's a really good point. It almost feels like a leaky exception handling to me.
comment in response to
post
Function definition is also super nonintuitive.
comment in response to
post
I'm using the spitbol 64 compiler github.com/spitbol/x64
And yes, the success vs fail signaling is fascinating to wrap your head around!
comment in response to
post
Did Linux format get canceled!?
comment in response to
post
I mean, I really liked my 286 but idk the full story yet.
comment in response to
post
That series was getting up to the 8086 specifically. I haven't gone back to grab more Intel chips. I'll probably get to the 286 eventually.
comment in response to
post
Oooh. No.
comment in response to
post
That's so cool! Congrats!
comment in response to
post
One thing I'm realizing is that LGP-30 tapes don't have headers per say, but they _do_ have instructions to the p104 program loader. So they don't work without p104 being loaded. At least... some of them.
comment in response to
post
The displays weren't buffered, from what I can tell. You had instructions to plot pixels directly, and the manuals give time measurements for phosphor persistence. So it was all direct.
comment in response to
post
A lot of char encoding in that period was 5 or 6 bit. I'm not sure how much that actually contributed to word sizes, however. In some machines it definitely seems like text was an afterthought. I'd have to look at LINCs keyboard instructions again.
comment in response to
post
Combine that with the cast that all code for the LGP-30 is self modifying and, well, 🤯
comment in response to
post
I'm trying to get to the point of running the example program at the back of the programmer's manual. But... that needs a subroutine loaded which means you have to have the loader loaded which means............. a lot of work. I'm starting to see how hard programming was in the 50s.
comment in response to
post
That's beautiful
Hahaha
comment in response to
post
Haha. Who even wrote DOS at this point? Seems like it was a real tapestry of other people's code.
comment in response to
post
I'm in a crossover episode?
comment in response to
post
Hell yeah. I appreciate you.
comment in response to
post
...but there is nothing at 0600 to call! It's trying to run some standard subroutine that I haven't found yet. So it's back to the archives until I find some "print integer number formatted" subroutine tape.
comment in response to
post
Woo! Very excited for next year!
comment in response to
post
*now, not not. Lol
comment in response to
post
I am planning to expand things once everything fully works. Going to have resources about programming and operating the machine, and all that kinda stuff.
comment in response to
post
And as for using TS: I prefer it to JS. It's just an improvement. Since it's compiled I can target for web easily, or for nodejs. The deep cut is that with TS Native announced this _should_ be a le to be compiled for native execution... One day
comment in response to
post
The other fun part is I'm implementing this as a finite state machine. Each operation updates a state. That means I can should be able to do funky things like save states. Maybe that could be used for real-time animations.
comment in response to
post
I tend to be a very hands on programmer, so I do a lot of rewrites. Now that I know roughly how I want things laid out and that the code can actually work the refactor should close things out nicely.
comment in response to
post
I'm up to the point where I have started trying to bootstrap programs. Very excited!
comment in response to
post
The big annoying part is dealing with the Flexo. Char encoding is... interesting. The chars off the flexo can flow as either 4 or 6 bit, and in 4 bit mode there is a char -> CPU instruction equivalency.