[eichin]

I hadn't realized Domain/OS emulation was viable these days. It's one of the few systems that has actually "lost" features - the terminal-window-like thing (called pads, I think?) when in line mode had a dividing line at the bottom where your unconsumed typeahead was visible and you could continue to edit it until it got read - not just one line, the entire unconsumed input. (Not that it's a particularly desirable feature - it's just one that I'm pretty sure you can't implement with ptys...)

[bilegeek]

Unfortunately, pre-Domain/OS AEGIS is basically lost. One person popped up with talk of imaging their 9.6 floppies, but I haven't seen anything since then.

[1]https://www.facebook.com/groups/retrocomputers/posts/7062462...

[FarmerPotato]

I just received from a retired engineer, a binder of 8” floppies that says Jan 1984, AEGIS 6.0 / Mentor 3.0, Full Backup, WBAK. The owner got them from a dumpster 40 years ago, but suspects someone just reused the binder to store blank floppies. Anyhow I’m working on it.

I’ve also found source for an AEGIS menu system (mouse, hotkeys) written in Forth.

[em-bee]

it's probably not old enough, but in the mid 90s i acquired a working apollo domain workstation that was functioning as a doorstop at a university library. it came with a full set of documentation, but no floppies, i think. i don't know which version, and i don't know if it is still working now. it's gathering dust at my mothers home in europe.

[neilv]

I wonder whether this could still pop up at estate sales, or when a retiree is cleaning out their garage.

Not all gear got junked. When I was a teen intern, I got some obsolete Apollos (and 2 logic analyzers and a terminal) from my employer, and other people were also bringing home gear the company "sold" them.

Somewhere, there might well be an industry or university sysadmin or programmer who brought home a box of old QIC tapes, and one of them says "AEGIS" on the label, and it's in a garage/attic.

Also, rumor has it that at one point Boeing physically archived at least one Apollo network, because they apparently take documentation integrity extremely seriously. If that's true, they might have an engineering librarian or someone who could take an interest in making sure any versions of Aegis/Domain they need (and have preserved media for) can run on emulators or something?

[andreww591]

Yeah, I'd definitely like to see older versions of AEGIS as well

[andreww591]

Yeah, MAME has had working Apollo emulation since around 2010. Domain/OS is definitely pretty odd. You could almost mistake SR10 for a normal functional Unix if you use the SysV or BSD universes rather than the AEGIS one, but while it is clearly Unix-like, it's also quite Multics-like as well and is pretty distinct from the typical functional Unix family.

[jerf]

Not only can you implement that with PTYs, it's how they operate by default. That's why you can telnet to an HTTP server and make a mistake and use backspace to fix it. The terminal will only send lines over. You have to use a command to put it into "raw" mode so the application gets every keystroke immediately. You have to ask for your PTY to not work that way.

[eichin]

That's not the same thing - with a pty in line mode, you get one line that you can edit, and when you hit enter, it gets sent. With Domain/OS, you still get to edit your last line until you hit enter... but if the process is otherwise blocked and doesn't consume it immediately, you can up-arrow and continue editing the earlier bits too. (Pretty sure it didn't work with telnet in particular, the common cases I remember were issuing a foo;bar;baz commandline and "working on" the input to baz while foo and bar were still running.)

[glhaynes]

What an amazingly goofy (but also kinda maybe makes sense?) feature!

[compsciphd]

why could you not implement it as ptys.

Currently the terminal doesn't really process input itself, it just gives the program running the "raw" fd.

If instead the terminal gave the processes a pipe (for instance) and consumed all the pty input itself (and its end of the pipe being a buffer of that content), why wouldn't it be the same?

[eichin]

pipes buffer, and you don't know when the child is actually reading? (Maybe you could cook something up with ptrace, since it would be a blocking read in this case...)