10 September, 2009

restored minicomputer, 80's-style


Reinhard Heuberger made a YouTube-video showing off the booting of his restored Digital Equipment Corporation PDP-11/23 SE minicomputer (last used in 1987).

I think it's pretty funny how futuristic looking the casing and controls of the machine are. But to be fair, it's a die-hard fan's restoration project, so it doesn't surprise me he built it to look like some machine taken from first-generation Star Trek. I think it's awesome! ;P However, I do not know how an original PDP-11 looked like, so for all I know, it's restored using mainly original parts.

Pay attention to how much the UPC is overwealming the sound of the 5Mbyte harddrive while booting :P classic. It almost sounds like an industrial-sized airconditioner for supermarket freezers.

Remember, this is a computer from the early days of computing. No fancy graphics, no sound feedback, no removable storage, no simplified standard controls. It's all completely hardcore low-level. I cannot imagine how much heat dissipiates and how much power is required to operate the beast. Like illustrated in the video, it takes both an A/C supply and a D/C supply just to wake it up.

These early computers are the reason why I prefer a CLI instead of a GUI. Gives me a nostalgic operator feel, since I never had the opportunity to become one ;D

The video showcases the programming language BASIC (RT-11* revision for the PDP-11) running on the PDP-11/23 minicomputer through a command prompt (serial connection?) on a Windows XP machine. It also depicts how ASSEMBLER can be used with RT-11* BASIC to operate a computer-controlled CNC-machine and request i/o from it.

* RT-11 - single-user operating system developed by DEC for their PDP-line of minicomputers.

As a little sidenote, I think Reinhard had to physically modify his machine to incorporate a serial interface to connect a terminal some of the PDP minicomputers had to be extended with serial interface cards to connect serial consoles. PDP minicomputers did not come with either serial consoles/screens for i/o as standard, instead they used punched paper tape, control knobs and flashing diode lights for manual debugging and operation. The most important tools a hacker of those days had to have, were the things we stopped using by the end of the 90s, namely: a pen and paper. To hold a general idea of what the PDP was doing, you had to organize the punched paper tapes you were feeding the reader, note the flashing diodes and keep track of the various manual monitor-functions to debug and operate the system.
MY point is really this: computers were created and used in the 60s/70s by huge corporations, universities and science labs, when all computers were nothing more than headless, insanely expensive calculator-boxes. It's the later development and pioneering which followed the footsteps that has made computers what they are today.

Additional ramblings:

The shell UI and control-functions looks a lot like BSD to me. I remember my first BSD-experience back in the 90s. I had received a CD-R containing FreeBSD from a fellow geek-buddy, who wanted to introduce me to the world of *nix. But until that point, I had no experience what-so-ever with POSIX-systems (were BSDs classified as POSIX back in the day? if somebody knows, please drop me a line, I'm really curious about this kind of stuff).

Anyway, my point being; FreeBSD at that time was as UNIX as you could get it without actually using UNIX. This also meant there were no auto-completion functions or "automagic" functions as we have in Linux today, no, everything had to be manually defined and executed. Suffice to say, the learning curve was STEEEEEEP...

I COULD have downloaded the FreeBSD manual, and do what I normally do when I'm presented with a new system. Read, read, search online, and read some more. But this was in the time before I became obsessed with low-level computer technology. I was like most other computer-users, not bothering about the internals of neither machine, nor software. I basically "just wanted things to work", not much unlike today's paradigm of "out-of-the-box" implementation. But, having the wealth of computer experience I have today, has shifted my viewpoint entirely.

Now, I usually start off an interest by tinkering, destroying, repairing and reading manuals, forums and wikis on anything technology-related.

2 kommentarer :

Stephane said...

The PDP-11 is not a mainframe. It's a minicomputer.

The PDP-11 displayed (11/23) is microprocessor based. The older models, like the 11/05 or 11/40 etc. use TTL chips assembled on large boards connected through UNIBUS. The 11/23 uses tiny cards connected through Qbus.

Most CPU boards in Qbus format contain a debugger, called ODT. Most of them do not have a serial interface builtin, they must use a serial card (the console register is probably at 777560 or thereabout).

The older ones have just keys and lights as their front panel. But serial interfaces were always available (need to add 1 or 2 cards). But most config had at least a TTY (the console).

PIZslacker said...

thanks for the feedback :) I've brushed up on my computer machine-size definitions ;) I should have paid more attention to the internal specs of the machine while I wrote this post.

Post a Comment