On 15-Feb-02 Martyn Drake wrote:
Hi folks,
Just for a bit of fun (since it's Friday) - what was your very first exposure to Linux and what was your first distribution that you ever installed (and can you remember the kernel version?!).
Well, since you ask, the story (for good reasons) goes back to 1980 or so when I was using CP/M -- brilliant OS considering what it packed into the top 8K of 64K RAM (Oh, not forgetting the 256 bytes it also needed at the bottom, but never mind). And the ZILOG Z80 processor it ran on was not a bad beast either, for the time.
What was amazing (and still is, looking back) is the performance of the best CP/M software -- remember WordStar, dBaseII? VisiCalc? And I was compiling FORTRAN on it, and later C (Aztec C; excellent). Mind you, you had to know how to handle "overlays", but in the circumstances of the times that was forgiveable.
Later, it was a plunge from CP/M into Unix: 1984 or so, and a few of us statisticians had become computer missionaries in the Dept of Pure Maths at Cambridge, not only using them for our number-crunching but also advocating the use of microcomputers (as they were called then) in teaching (and being accused of "throwing money at the problem"; what's new?). But then the PMs noticed that the Applied folk were printing their own articles, lecture notes and exam papers just like on a real printing press, using a Unix system running on a ZILOG ZX8000 machine and typesetting software called "troff". So it was "Us too, please"; and at the end of the day yours truly made the inspired mistake of volunteering to get the Pure counterpart installed and running, and to manage it. Not knowing anything much about Unix.
Well, that was a Unix machine with a 68000 processor, and it had all of 1MB RAM and 10MB hard drive. "WOW!!" from your boy from the backwoods of CP/M. Also, on the back of the box was a bloody great serial board with 32 ports on it, from which cables fanned out to Wyse terminals across the Dept, to secreraries and Professors alike. (And remind me sometime to tell the story of "device full" due to trying to log-in the mains hum one night ... ).
Well, up to say 6-7 users could be on at once, and you wouldn't notice that anyone else was using the machine. Above that and, well, it did slow down ... Nevertheless, your boy from CP/M was vastly impressed; and in the process learned a lot about Unix and what capabilities should be expected from a really good OS; and also why, and how, Unix was one such. And that's why that mistake, of "volunteering", turned out to be inspired.
And so it goes ... after a few years, the early IBM PCs appeared; some people turned to these, but the Unix fans amongst us managed nicely without. What we couldn't (for some reason) achieve on our Unix machine we could still do on CP/M and/or BBC micros.
By 1989 I was in Manchester, and before long grappling with the PC and with DOS. It rapidly became clear that DOS was CP/M-like with a few Unix "features" tacked on: mainly directory trees and a really nasty kludge for pipes. Oh, yes, there was a DOS program called "more". Big deal. I'd already written my own in C for CP/M ...
The one major advance was in the area of graphics: PC hardware was much better at this than what we'd had on CP/M or the Wyse terminals. Apart from that, little could be done with DOS software that couldn't have been done before. But multi-tasking? WHAT???? [Aside: hats off to Quarterdeck and their "Desqview", which finally did manage to achieve fair multi-tasking on the PC with DOS.]
Then it struck me: The PC I was using (386SX 25MHz processor) itself had 4 MB Ram, and a massive 40MB HD -- more hardware power on every front than the Unix machine I'd been running (except for that serial board). And the best that could be done with it was to run DOS! Blimey. [And the serial board would have been irrelevant anyway.] Actually about the same capability as I was getting on CP/M with 64K RAM.
Well, you don't have to guess for hard the next thought that struck me. Where's Unix for the PC then? We're up to about 1990/1 now.
Well, there were one or two knocking around. SCO Xenix comes to mind. But you paid the earth just for the Unix "base", and then some if you wanted to run anything useful on it. Also on the scene around then: The Next Box, with NextStep OS (basically Unix with a graphical front-end that used Display PostScript to relate to the user). A beautiful thing, I was told. But costing a good fraction of a year's pay.
By 1992, I was hyped up into "mission status" on this issue. No way was I putting up with Unix-capable hardware that was only running DOS; hardly an advance on CP/M in 64K.
Early MS Windows had come out, and I had a look at that; those of you who remember it don't need telling what I thought of it (and those of you who don't remember, but do remember more recent series of Windows versions, can extrapolate backwards with confidence.
Had another look at NextStep. By now the Next Box was dead, but the OS was now available for the PC. Three snags. (1) The OS was 1,000 quid. (2) Appications were extra. (3) No way would my 386SX-25 box cope with running NextStep. Hardware of that strength would have set me back about 2,000 quid. 3,000 quid all told. Plus.
Also had a look at IBM's OS/2, which seemed to be a good approximation to the real thing. And, believe it or not, at "Windows NT", already lurking back-stage by 1992; and this, according to its description, would turn Windows-on-DOS into Windows on something that might have the multi-user, multi-tasking capabilities that I declared to be the birthright of every user of sufficiently powerful hardware. Wasn't sure about it having good tools, though.
But it was the wrong moment to be looking at either. IBM and Microsoft got into a tangle, which wrapped itself round OS/2's neck in the womb and strangled it. Windows NT retired into an extended development phase, and remained irrelevant for some years to come. And I began jumping up and down.
* * * * * * * * * *
At this moment, lateish 1992, in the Newsletter of the Manchester Computing Centre, appeared a small announcement:
"Free Unixes for the PC"
placed by Owen Leblanc (one of the most notable Linux pioneers in the UK, and contributor to the ext2 filesystem). It described FreeBSD, and Linux. FreeBSD was "more mature", but wanted the whole hard drive to itself. Linux, while "experimental", could share a drive with DOS. After some hesitation, this persuaded me to try Linux.
So I jumped over to MCC, and sought out Owen. Armed with about 25 floppies, he took me to a PC where we downloaded SLS Linux (kernel 0.99 if I remember), and I took this lot back home.
Repartitioned the 40MB drive into 10MB for DOS, 5MB swap, and 25MB for Linux and user workspace. [Don't ask what it was like doing this for the first time, in those days. But Owen had written excellent documentation of the details of what was going on, so I got there, making progress notes on paper on the way.]
Then inserted floppy after floppy, picking and choosing, being very very careful ... After a long time of this, finally the thing booted into Linux, and sitting at the foot of the console was
root #
and a flashing cursor. The dream come true. Unix on the PC. And, to check I was really home, I entered "ls /usr/bin", and the familiar faces rolled past. Finally, to culminate the ceremony,
shutdown -h now
and it wrapped itself up and went to bed. I was home again.
Well, after that no looking back. No X, of course; not with that capacity hardware. But SVGAlib was your friend for graphics, and the console terminal was really (as it still often is) all you really needed. But those 25MB for Linux and user workspace allowed a lot of useful real work to get done.
So that's how I came to Linux, what my first distribution was, and what the kernel version was.
After that, it was MCC-Interim 0.99, Slackware (which I stuck with for a good few years), Red Hat, briefly Linux Universe, a brush with Debian unfortunately at the bad moment when it tried to move from 0.99 to 1.0 and got it wrong, then SuSE. Currently running on different machines Debian, SuSE-5.1, SuSE-7.2, and Red Hat 7.1. And remembering that interesting time of transition from "a.out" to "ELF".
I still feel that angry missionary urge which gripped me around 1990, still trying to show the cannibals what can really be done with a proper OS and proper tools.
And yet, more and more marvelling at how it is that what used to be possible in 4MB RAM and 40MB HD, with room to spare, now wouldn't even install on that.
Well, you did ask ...
Two days ago, someone wrote to the groff mailing-list:
"Of course, choosing an embedded language must be done in terms of the target audience which is *not* jaded old hackers who got in touch with computers in the early eighties at the least, that is in times so Paleolithic to have known *roff as the tool of choice to typeset their documents, and find that writing programs in VM/CMS pseudo-assembler is the closest to nirvana one can get while incarnated."
The cap fits.
Best wishes, therefore, from your Jaded Old Hacker, Ted.
-------------------------------------------------------------------- E-Mail: (Ted Harding) Ted.Harding@nessie.mcc.ac.uk Fax-to-email: +44 (0)870 167 1972 Date: 15-Feb-02 Time: 21:57:47 ------------------------------ XFMail ------------------------------