-----------Original Message:
Date: Sat, 14 Apr 2007 07:51:58 -0700
From: "Chuck Guzis" <cclist at sydex.com>
Subject: Re: HP "calculators" (was Re: World's first computer on
ebay!)
Yes, it's true that BASIC doesn't differentiate lexically between the
assignment operator and equality test, but that seems to be unrelated
to the behavior of the two statements I gave. In fact, I don't know
what the operation performed by GWBASIC is in "X=Y=7".
Cheers,
Chuck
-------------Reply #2:
Maybe what's missing in my blathering is that when BASIC does a
conditional test it's actually looking for a non-zero number in whatever
its argument is, which could be a logical expression (which returns 0 or -1)
or simply a variable.
So instead of IF A <> 0 THEN... you can just say IF A THEN...,
something you'll see throughout my sloppy code ;-).
m
I've recently been looking at setting up an old HP Vectra P-90 to run Linux
(or some form of *nix) and started wondering what ports exist for "classic"
systems.
Looking around I see that "NetBSD" in particular supports quite a few
architectures including VAX, SGI, NeXT etc etc, but I've struck out
regarding Linux ports.
So, can anyone point me at a website listing Linux ports to architectures
other than PC's. And what modern *nix ports, if any, do other list members
use on their classic iron?
TTFN - Pete.
------------Original Message:
Date: Sat, 14 Apr 2007 07:51:58 -0700
From: "Chuck Guzis" <cclist at sydex.com>
Subject: Re: HP "calculators" (was Re: World's first computer on
ebay!)
I said "X=Y=7" in GWBASIC changes both X and Y the same way
regardless of their initial values. LET X=Y=7 changes them a
different way, regardless of their initial values. If there's a
conditional operator in either of those statements, I can't find it.
Yes, it's true that BASIC doesn't differentiate lexically between the
assignment operator and equality test, but that seems to be unrelated
to the behavior of the two statements I gave. In fact, I don't know
what the operation performed by GWBASIC is in "X=Y=7".
Cheers,
Chuck
----------Reply:
We can't be talking about the same thing here; I program a fair
bit in BASIC and use this technique quite often; (Y=7) is equivalent
to 0 or -1 (0000H or FFFFH, depending on Y, when it is in a place
where a numeric variable is expected. So if I read your examples
correctly, when Y is 7 then (Y=7) is -1; in essence TRUE has a
value of -1 and FALSE has a value of 0.
The conditional operator is implied, which is why it's useful when
an explicit IF/THEN is awkward; If Y=7 then (Y=7) is -1 else (Y=7) is 0.
Y never changes.
No?
Just try: Y=7:PRINT Y=7: Y=6: PRINT Y=7
or: Y=7:PRINT Y>6:Y=6:PRINT Y>6
or: A$="X":PRINT A$="X":A$="Y":PRINT A$="X"
or, to make it even more obscure:
INPUT "Guess a letter":A$:PRINT mid$("WRONGRIGHT",-(A$="Z")*5+1,5)
Admittedly, it's counterintuitive; X=Y=7 sure looks like a multiple assignment.
m
-------------Original Message:
Date: Sat, 14 Apr 2007 10:37:11 -0400 (EDT)
From: der Mouse <mouse at Rodents.Montreal.QC.CA>
Subject: Re: HP "calculators" (was Re: World's first computer on
ebay!)
>> The odd thing is that GWBASIC will pass a statement of the form
>> X=Y=7, but Y will be 0 and X will be -1 at the conclusion of
>> execution, regardless of the value of y or x at the beginning.
>> OTOH, LET X=Y=3 will set both X and Y to 0. GWBASIC is full of
>> stuff like this.
> Eh???? Did I miss something?
Yes, I think so.
> Those are not multiple assignments but very useful logical tests.
Then why do they change Y?
------------Reply:
They don't. That's why I think I must have missed something.
m
>
>Subject: Re: Dreaming of a lean installation method
> From: davis <davis at saw.net>
> Date: Sat, 14 Apr 2007 02:36:18 -0700
> To: General at saw.net, "Discussion at saw.net":On-Topic Posts Only
> <cctech at classiccmp.org>
>
>Ensor wrote:
>
>> Hi,
>>
>> > I bought the MINIX version for the Atari ST and installed it, but
>> >I used it as a toy, I don't remember recompiling the kernal. It had
>> >full kernal source, and all the tools to rebuild....
>>
>> It sure came with the sources, but I'm 99.999999999999% certain you
>> could not rebuild the kernal because no compiler was included in the
>> distribution.
>>
>> >....I think you're mistaken saying it was useless.
>>
>> It certainly fulfilled it's aim of being a platform to learn about
>> modern OS design and implementation, but without applications or a way
>> to develop them, it's usefulness is ultimately limited (by what is
>> included in the distribution).
>>
>> Don't forget, very few people had 'net access back then, those of us
>> who were on-line were mostly on FidoNet and there was nothing online
>> for Minix that I could find (and I tried pretty hard).
Minix is not up to V3.x and easily found on the net even old V1.5. It's
sources are available. It does not come with a compiler after all it's
an OS. It was expected that MINIX would be cross compiled on a platform
that supports a suitable compiler. there is nothing to say it's not
possibel to create a native MINIX compiler and then run/compile natively
assuming the platform is big enough to support it.
I might point out that CP/M was initially cross compiled on a PDP10
for the 8080 and in the later years rewritten in C and a VAX was used.
It's not unusual to have a OS that does not come supplied with
compilation tools.
Allison
> Subject: Dreaming of a lean installation method [was Re: *nix on
> "classic" systems]
I'll expand on my original suggestion: fork something from the NetBSD 1.x
line, call it ClassicBSD or OldSchoolBSD or whatever, pick a fairly fixed
set of hardware to support and commit to keeping things small and stable.
This includes the painful process of throwing things out (in the kernel and
in userland) that aren't essential to keeping old gear running or flat too
big to keep around. Something vaguely like the 4.3BSD-Quasijarus
philosophy, but with less....weirdness.
To be survivable, you'll need to make sure what you roll out is the most
conservative set of things that can actually be maintained. You need to
have the guts to put the foot down and say things like "Your favorite
program is too big, too slow and way too much work to keep current on the
kind of machines we support, so we won't even try to support it" and "even
though one of the machines we want to support theoretically could have a USB
card in it, almost all don't and we can save X amount of memory and Y amount
of support time by stripping the whole subsystem out, so that's what's going
to happen" or "SMP? No. Just no.".
ClassicBSD would need to be about 'have to have' functionality, not 'nice to
have' functionality. Just as long as we all agree NetHack stays.
Note that "fork" != "stagnate". You'd need to build a community, but for a
handful of committed folks folding security and other patches back in is
straightforward if very time consuming, especially if you've already
stripped out big, impractical stuff. Lots of projects do it with more or
less success.
BTW...if you aggressively avoid eye candy, even X11 can be relatively useful
on old hardware. Strip out all the bloaty modules and drivers (you probably
don't need OpenGL and font smoothing is a luxury you can't afford) turn off
all the fancy options (no, you can't have opaque moves...not yours) and keep
to a bare minimum WM.
There are even lighter choices...fork a "classic" X release (like 11.3...as
if your hardware supports any of the functionality added since then) or
someone could raise from the dead/port MGR or some such.
Ken
As I go through & clean up my unibus type stuff I have come across a few
things I know for sure I don't want or I have way more of than I thought....
RX01 drive (needs a lot of cleanup), two RX01 unibus controllers
a large stack of M7800's
DR11 dec and clones
A lot of dec corp cabs (white/beige) in various sizes
Items I'll need if anyone has available:
RX02 unibus controller
Cable that connects RL02 controller to backpanel bulkhead (and bulkhead
connector), or controller to drive cable
Cable and rackmount panel for DZ11
Top cable clamp for BAll-K
Misc. 861 power controller 3-pin molex cables
Console cable from M7856 to DB25
RL02 cable from controller to back of rack
I have a 3rd party 4 serial port unibus board - I'd love the cables that go
>from the card to the backpanel bulkhead & connector
Hello, Retrocomputing fans,
My first full-time job in the software field was at a company called
Systems Research, Inc, in Okemos, Michigan, in the U.S.A. Our company
made a front-end, based upon the HP-21MX machine, which could be used on
a Burroughs Medium System, and take the load of Burroughs' burdensome
TD-830 terminal protocol off of the main processor. They also produced
terminals which emulated the TD-830. These terminals were called the
SRI Century Terminal. The software engineers managed to fit the entire
code for the protocol into 7K of ROM memory. That left 1K on the ROM
open. Some of us were asked to write a TTY protocol for the same
machine, so it could be a dual-purpose terminal -- and to fit it into
the remaining 1K. To make a long, reasonably interesting story short,
we did. In the process, we wrote some of the UGLIEST code I've ever
seen, including a jump into string storage, just because, if one pushed
a register or two on the stack, it would get done what was needed, and
save almost ten bytes. <Shudder>
Anyway, I digress. I have a fondness for this example of jungle
coding, and would like to obtain one of these terminals. Realistically
speaking, the terminal was not a remarkable item; it was made on a
standard OEM frame. It does have sentimental value to me. I've located
some of my old cohorts, and none of them knows where to find one of
those terminals. Burroughs ended up buying SRI, causing me, and many
others, to leave, and they gave the Century their own name.
Unfortunately, I don't know what that name might be. And, whatever THAT
name might be, it could have been changed to another when the Unisys
name took over. If anyone has any information about this machine, I
would greatly appreciate it. Thanks!
Peace,
Warren E. Wolfe
wizard at voyager.net
Rumor has it that Ensor may have mentioned these words:
>LOL, the 6809 assembler I'm using at the moment is run from the command
>line (though I'm firing it up through a batch file and editing the source
>with "notepad" - who needs an IDE?).
Needs? Well... dunno about "needs" but Roger Taylor over at www.coco3.com
has written a beautimus IDE called "Rainbow" that you can interface your
favorite assembler into (It comes with CCASM, a very nice 6809 assembler)
and is interfaced right with M.E.S.S. so with several classic machine
emulations, click a button, all the assembling is done, if there's no
errors, it automagically creates/formats/writes the .dsk files & starts
M.E.S.S. with the disk already mounted & ready to go.
It makes writing & debugging code for the CoCo a dream, and he's expanding
it to include other machines & assemblers so you can write code for the
Commodore 64, Atari 800, etc. type machines.
The software is new, and technically you could use it to write code for an
offtopic machine, but it really is aimed at the ol' 8-bitters, so hopefully
this isn't too offtopic. IMHO, it's a lot closer than where the whole Linux
thread...[1]
I'm not affiliated with him other than being a very satisfied customer.
Laterz,
Roger "Merch" Merchberger
[1] Damn Small Linux (http://www.damnsmalllinux.org/). Whole distro's 50
Meg, boots from CD/USB/HD, runs in RAM, with X. Pretty slick. ;-)
--
Roger "Merch" Merchberger | "Profile, don't speculate."
SysAdmin, Iceberg Computers | Daniel J. Bernstein
zmerch at 30below.com |