This reminds of a quixotic quest I undertook once - implementing enough of a SLIP and TCP/IP stack to send mail with SMTP and retrieve it with POP3, all in PC BASIC (the kind that still required line numbers).
Amongst other problems, it wouldn't work properly (dropped characters) at anything faster than 9,600 baud, and on an original PC needed to go a bit slower than that; computing things like checksums was exceedingly painful; there wasn't enough RAM (it was limited to 64K for program plus data, which in practical terms meant a PC with at least 96K of RAM); it would have to drop the connection if it shelled out to an external editor. But it did work.
Might as well go the full insanity.. if you can get it work on the ACE, then surely you can make it work on a ZX81 with a rampak?
I remember as a kid in the 80s writing a program in BASIC to get my Spectrum 128 to connect over serial to my Sirius 1 PC so my buddy and I could have a chat application across my bedroom. Wild times.
"Keep that kernel version history in mind for when we get to oddiments of the C compiler. As for networking, though, with the exception of UUCP over serial, none of these early versions of Venix on either the PDP-11 or 8086 supported any kind of network connectivity out of the box."
You would not have known it was a F-11 CPU unless you opened it up.
The LSI-11/23 also could have a programmers utility, which was why Microsoft used one to develop Its DEC version of Xenix. So the LSI-11 birthed TWO separate wanna be UNIXes. Venix ran on the 8088, Xenix ran later on the 8088. Venix ran on the IBM AT, Xenix Ran on the IBM AT. What was DEC doing while all this was going in in the UNIX wars?
Although that this claims pre-C89, it was at least 5 years earlier, and running the portable-C Compiler, apparently Xenix version was a bit better, but not by much.
Which now leads me to speculate that after Microshaft divorced themselves from Xenix, they may have left their legacy email server ( a VAX ) for the use of SCO to keep developing Xenix on.
Being a bit lazy to cross-check dates, maybe their OpenVMS efforts?
I learned C before having access to Xenix, this was definitly before any portable C compiler, I imagine, because the book used RatC, a K&R C dialect.
I read somewhere that originally Microsoft did not develop MS-DOS directly on PC hardware, rather they would cross-compile/assemble, maybe they were using systems like these?
This reminds of a quixotic quest I undertook once - implementing enough of a SLIP and TCP/IP stack to send mail with SMTP and retrieve it with POP3, all in PC BASIC (the kind that still required line numbers).
Amongst other problems, it wouldn't work properly (dropped characters) at anything faster than 9,600 baud, and on an original PC needed to go a bit slower than that; computing things like checksums was exceedingly painful; there wasn't enough RAM (it was limited to 64K for program plus data, which in practical terms meant a PC with at least 96K of RAM); it would have to drop the connection if it shelled out to an external editor. But it did work.
Shit I'd love: minimal slip/ppp over serial enough to run a gopher client with a Jupiter ACE and some RAM expansion (16k or 32k).
If a Speccy can be connected to the internet with a Gopher client...
A small IRC would he half useful too.
Might as well go the full insanity.. if you can get it work on the ACE, then surely you can make it work on a ZX81 with a rampak?
I remember as a kid in the 80s writing a program in BASIC to get my Spectrum 128 to connect over serial to my Sirius 1 PC so my buddy and I could have a chat application across my bedroom. Wild times.
Forth was and it's much faster than Basic sadly.
Real BASIC, the one invented at Dartmouth, was compiled to native before execution and relatively fast for the hardware.
It was fitting BASIC into 8 bit home computers limited hardware that gave fame to its slowness, given the interpreter approach.
By the 16 bit days having access to compilers was already not an issue. It was Turbo BASIC that started my journey as happy Borland customer.
There was a Twitter client for ZX Spectrum, using Spectranet
Now seems to be the trend to bring TCP to retro machines from first principles.
8088 PC's already have M Brutman's "mTCP" driver. And DogCow (D. Finnigan) has been working on and off on the "Sabina" networking suite for the original 128K Macintosh on the MacGUI site: https://macgui.com/sabina/ https://macgui.com/news/article.php?t=550
Though it's a been slow going, apparently because of a day job. (Also note that the MacGUI site now requires an account to read progress posts.)
Interesting piece of history.
Some more market and technical information about the DEC PRO/VENIX system here:
https://oldvcr.blogspot.com/2025/03/more-pro-for-dec-profess...
And this wonderful quip here:
"Keep that kernel version history in mind for when we get to oddiments of the C compiler. As for networking, though, with the exception of UUCP over serial, none of these early versions of Venix on either the PDP-11 or 8086 supported any kind of network connectivity out of the box."
https://oldvcr.blogspot.com/2025/04/lets-give-provenix-barel...
Thanks for sharing, great content there.
Yes, Very ancient machine.
You would not have known it was a F-11 CPU unless you opened it up. The LSI-11/23 also could have a programmers utility, which was why Microsoft used one to develop Its DEC version of Xenix. So the LSI-11 birthed TWO separate wanna be UNIXes. Venix ran on the 8088, Xenix ran later on the 8088. Venix ran on the IBM AT, Xenix Ran on the IBM AT. What was DEC doing while all this was going in in the UNIX wars?
Although that this claims pre-C89, it was at least 5 years earlier, and running the portable-C Compiler, apparently Xenix version was a bit better, but not by much.
Which now leads me to speculate that after Microshaft divorced themselves from Xenix, they may have left their legacy email server ( a VAX ) for the use of SCO to keep developing Xenix on.
Being a bit lazy to cross-check dates, maybe their OpenVMS efforts?
I learned C before having access to Xenix, this was definitly before any portable C compiler, I imagine, because the book used RatC, a K&R C dialect.
I read somewhere that originally Microsoft did not develop MS-DOS directly on PC hardware, rather they would cross-compile/assemble, maybe they were using systems like these?