UVA Computer Science Museum 186
Cryptographrix writes "Just came across this site, thought slashdot users should check it out, definately worth a read, has everything from the original Osborne portable computer to such memorables as the Altair...supposedly from the UVA staff's personal collection. Even has old (1950's and another board that looks like ESS3, maybe) telephone switching equipment."
50 years from now... (Score:3, Interesting)
Just imagine high school science-class field trips laughing at the very system you're using now...
---
Destiny-land [blogspot.com].
The happiest blog on earth.
Re:Punch cards (Score:3, Interesting)
"A master control program to automatically manipulate machine programs, allocate memory, assign equipment, and route all information.
Found that quite humurous - I wonder if that is where the tron script writers got the idea? Reading the brochure was odd - I am a youngin' and know very little about very old computers (relatively...), and was quite curious about the description [virginia.edu] of the chip: "processors operate on 49 bit words (48 bits plus parity bit)"... where these chips then 49 bit? From the sound of the brochure it makes it seem like the entire system was 49 bit (memory, storage, etc). Or was it like a 4 bit processor that just used 49 bit commands?
Anybody know?
Another UVA, another museum (Score:2, Interesting)
Re:TRS-80 (Score:3, Interesting)
It's been sitting in my garage since the early '90s, when I switched first to a Sun 3/80 and then to Linux on a 386DX/25.
I've also got a TRS-80 Model I system with monitor, expansion unit and floppy drive sitting in the garage, but I don't think I'll part with that one yet...
What about all these machines... (Score:4, Interesting)
But what about the earlier machines that broke new ground:
The CompuColor. This was a great machine. It only had an 8080 processor but was one of the very first "off the shelf" machines to come with amazing (from memory) 128x128 8-color graphics. It also had the disk-drive built into the color screen with a whole 84Kbytes of formatted storage.
The Commodore Pet. Just as every movie ever made to day has an apple of some flavor in it, the Commodore Pet used to be the favorite choice of movie makers when they needed to show a microcomputer somewhere. It's very distinctive looks made it instantly recognizable -- but its lackuster performance and monochrome character-based graphics was a disappointment
The TRS80 model 1. This was the main competition to the Apple II in the late 1970's. I actually preferred it to the Apple as it had a much more powerful BASIC interpreter (double-precision math!) and could be easily converted to display proper lower-case characters. It also had a decidedly flakey expansion unit that could hold up to 32 or 48K of RAM and from which up to four floppy drives could be daisy chained. Add some double-sided, double-density 80-track drives plus a copy of NewDos80 and you could get up to 1.6MB per drive for a whopping total of 6.4MB of online storage!!! Woah, be still my beating heart.
The Intertec SuperBrain. This was a really odd box that looked just like a mainfraime terminal with keyboard, screen and drives all integrated into one whopping great case. It had two 4MHZ Z80 processors -- but only one was ever processing at a time because the second was dedicated solely to the task of polled disk I/O. Looking at the schematics and firmware it appears very much as if the designers used this method because they were too stupid to write good software for a single CPU. Its real claim to fame was that it was one of the first microcomputers with any real networking capability. If you bought one of their enormous 8MB server boxes (with a 8" hard drive) you could then connect up to 255 SuperBrain computers to it using a star topography network that ran over an inflexible and awkward 40-way ribbon cable.
There were numerous other very popular machines out there such as the Ohio Superboard -- a real hacker's delight. For your money you got a built-up circuit board with a full QWERTY keyboard right their on the PCB. You had to add your own power supply, case, monitor, etc -- but they were dirt cheap.
I used to love going to computer shows back in the late 1970's and early 1980s because there was always something *radically* different to see.
These days everything's just a slightly different flavor of IBM PC
Of course I'm a *real* hacker from way-back who built my first computer from scratch back in 1977 and then had to write and hand-assemble my own macro assembler before I could write a BASIC interpreter.
The processor was a Signetics 2650 CPU running at a whopping 1MHZ.
I started with just 1KB of of static ram and when I spent a small fortune to 4Kbytes I thought I was in heaven.
Believe it or not, I actually made some money from programming way back then. I'd hire out my computer to various shops where it would display a scrolling message I'd programmed (in my own BASIC) on a computer screen in the store Window.
In those days, the whole idea of a small computer and computer-generated scrolling text on a screen was so unusual that people would stop and look for many minutes. Great advertising for the stores which hired my little box and paid me to program in their message.
Geez I feel old
Re:Punch cards (Score:3, Interesting)
The machine was optimised for Algol and used a stack based architecture meaning that your arithmetic ops etc were done on the top elements of the stack rather than numbers in registers. There was hardware support for creating an Algol stack frame. I'm not 100% sure but I think there was a set of registers to keep track of the scope levels (C has only two levels of scope: local and global. Algol like Pascal can define procedures that contain other procedures recursively which complicates the scoping somewhat).
The programmer only actually saw 48 bits of the 49 bit word. For real numbers, each word was divided into a mantissa and an exponent + 2 bits for the sign of each of those. An integer was merely a real number with a zero exponent. I'm a bit hazy, but I think it used ones complement i.e. (assuming the mantissa sign bit is bit 47, -1 is represented as 800000000001 in hex, not FFFFFFFFFFFF, so you could negate a number merely by flipping bit 47.
If your program crashed, a crash dump would be produced on the line printer. Usually it would take about half a box of fanfold paper which you'd then have to wade through in conjunction with the program listing matching stack frames and variables to the correct names. I remember how we rejoiced after one MCP upgrade when the lines of the crash dumps suddenly started coming out with variable names printed next to them.
The CompuColor pretty much sucked. (Score:2, Interesting)
I'm not even going to get started on the NorthStar Horizon [hartetec.com] (64K of RAM!, dual floppies!, case made of WOOD!), or I'll start showing my age.
Whoops, too late.