Simulated PDP-11 Blinkenlight front panel for SimH

[Jörg Hoppe]

-
A few weeks ago I presented "Project BlinkenBone":
www.retrocmp.com/projects/blinkenbone
A real PDP-11 front panel was connected to a simulated PDP-11.
Well, nobody but me had any benefit from this project ...

Meanwhile there is also a photorealistic simulation of the front panel
itself.
It can be connected to SimH as an alternative to the physical panel.
So now everybody can run his/her very own Blinkenlight panel.

I made a simple installation package: Just download, unzip, start, enjoy!

It's been tested under WinXP/Win7 and Ubuntu. See download link and
instructions here:
www.retrocmp.com/projects/blinkenbone/176-blinkenbone-download-and-run-simulated-panels-for-free

In case of problems please give feedback.

Joerg

[Renaissance]

Jï¿œrg Hoppe wrote:

> A few weeks ago I presented "Project BlinkenBone":
> www.retrocmp.com/projects/blinkenbone A real PDP-11 front panel was
> connected to a simulated PDP-11. Well, nobody but me had any benefit
> from this project ...

> Meanwhile there is also a photorealistic simulation of the front
> panel itself. It can be connected to SimH as an alternative to the
> physical panel. So now everybody can run his/her very own
> Blinkenlight panel.

> I made a simple installation package: Just download, unzip, start,
> enjoy!

THX! :-) At the company where I work, I planned an internal course of
computer architecture and operating systems, and I chose the pdp-11 for
a brief practical course of assembly language (since it's didactically
perfect for the purpose). I'll use the frontpanel simulator to make
things more interesting. :-)

bye G.L.
--
Da i.d.c.tutela:
P.S. Quando ci sarￜ lo switch-off, avrￜ problemi anche col
monitor del PC? Ho visto che ￜ collegato in modalitￜ analogica.

[jmfbahciv]

Renaissance wrote:

> Jörg Hoppe wrote:
>
>> A few weeks ago I presented "Project BlinkenBone":
>> www.retrocmp.com/projects/blinkenbone A real PDP-11 front panel was
>> connected to a simulated PDP-11. Well, nobody but me had any benefit
>> from this project ...
>
>> Meanwhile there is also a photorealistic simulation of the front
>> panel itself. It can be connected to SimH as an alternative to the
>> physical panel. So now everybody can run his/her very own
>> Blinkenlight panel.
>
>> I made a simple installation package: Just download, unzip, start,
>> enjoy!
>
> THX! :-) At the company where I work, I planned an internal course of
> computer architecture and operating systems, and I chose the pdp-11 for
> a brief practical course of assembly language (since it's didactically
> perfect for the purpose). I'll use the frontpanel simulator to make
> things more interesting. :-)
>
> bye G.L.

Kewl. Let us know how it worked out. If the lights are realistic,
I'd like to know how easy/hard it was for the kids of this
pointy/clicky generation to use. oh, yea..and how much they
liked it.

/BAH

[Michael Black]

On Mon, 11 Jun 2012, jmfbahciv wrote:


> Kewl. Let us know how it worked out. If the lights are realistic,
> I'd like to know how easy/hard it was for the kids of this
> pointy/clicky generation to use. oh, yea..and how much they
> liked it.
>

You move the mosue over the switch, and click the mouse button. How hard
can that be? Of course, if they don't know binary (it wasn't octal,
right?) and they don't know the op-codes, or know how to program the
tbing, they can't get very far.

Michael

[Renaissance]

jmfbahciv wrote:

>>THX! :-) At the company where I work, I planned an internal course of
>>computer architecture and operating systems, and I chose the pdp-11 for
>>a brief practical course of assembly language (since it's didactically
>>perfect for the purpose). I'll use the frontpanel simulator to make
>>things more interesting. :-)

> Kewl. Let us know how it worked out. If the lights are realistic,
> I'd like to know how easy/hard it was for the kids of this
> pointy/clicky generation to use. oh, yea..and how much they
> liked it.

Frankly: I began operate computers (a pdp-11/23+) in 1983 when I was
sixteen, when the switch panel era was gone by a few years. I learned
the basics of fortran, pascal and macro-11 under rsx-11/m+, how to use
function libraries and how to write assembly function/procedures that
can be called by programs written in fortran and pascal. Also I used a
didactical microcomputer, the nbs-z80, that provided a "thousand-holes
board" for interfacing circuits/chips/devices, and a led alphanumerical
display plus a keyboard to enter z80 assembly programs in hexadecimal
form.
Therefore, I hope will not be particularly difficult to me to learn
how to use the front switch panel of an old pdp-11, and let my
"students" understand how a computer works in "bare metal" programming
and under an operating system. My aim is, at least, a little program
that write something on the terminal console (just to show how a
device is controlled in a system with memory-mapped registries).

bye G.L.
--
Da i.d.c.tutela:

P.S. Quando ci sarà lo switch-off, avrò problemi anche col
monitor del PC? Ho visto che è collegato in modalità analogica.

[jmfbahciv]

It was octal the last time I looked at the lights. I didn't read binary;
I would think in octal.

/BAH

[Joe Pfeiffer]

The few -11 front panels I saw were binary, but with the lights in
groups of three (there weren't spacing differences, just markings in the
silkscreen) so reading them off in octal was automatic.

[Charles Richmond]

"jmfbahciv" <See....@aol.com> wrote in message
news:PM0004C25...@aca21d2d.ipt.aol.com...

BAH, I seem to recall that the switch panel had alternating colors every 3
switches (or alternating color switch paddles), so one could see the binary
as entering in octal. In fact, I have heard that there was a "three finger
technique" whereby the operator could enter octal digits using the
individual bit paddles.

--

numerist at aquaporin4 dot com

[Walter Banks]

Nice job. Brings back old memories of sore fingers and a boot loader
so familiar that it could be keyed in without thinking. Two minutes playing
with your very nice simulation and I know why I seriously don't want
to go back..

Thanks for doing this

Walter Banks

[Michael Black]

I come from the generation after the string of panel switches and lights.

So I wasn't really thinking, but so long as there is a string of lights
and switches, one for each bit, it has to be binary. But, as you point
out, octal is just an easier way of viewing those bits.

My aside was because when there was the move from a string of lights and
switches to 7segment readouts and a keypad, you really were entering
things in octal rather than doing an instant translation before you
entereed on the string of switches. At that point, it wasn't just a
mnemoic system for remembering binary, it was the only way to enter
things, which meant some systems used octal and others used hex, and never
the two shall meet.

Michael

[gareth]

"Walter Banks" <wal...@bytecraft.com> wrote in message
news:4FD8B400...@bytecraft.com...

>
> Nice job. Brings back old memories of sore fingers and a boot loader
> so familiar that it could be keyed in without thinking. Two minutes
> playing
> with your very nice simulation and I know why I seriously don't want
> to go back..

But for those of us brought up in the lamps-and-switches world, being
able to have such intimate contact with the computer is useful on those
few occasions when debugging takes you down to that level.

[Rod Speed]

gareth <no....@thank.you.invalid> wrote
> Walter Banks <wal...@bytecraft.com> wrote

>> Nice job. Brings back old memories of sore fingers and a boot loader
>> so familiar that it could be keyed in without thinking. Two minutes
>> playing with your very nice simulation and I know why I seriously
>> don't want to go back..

> But for those of us brought up in the lamps-and-switches world, being
> able to have such intimate contact with the computer is useful on those
> few occasions when debugging takes you down to that level.

Nope, I never did debug at that level on machines that still had them.

And I had the boot loader engraved in my brain like walter did too.

[Jörg Hoppe]

Hi,

I sometimes do presentations with these panels and have a tutorial prepared.

It consists of entering a little assembler program, verifying it and
let it run.
The program simple adds the first <n> natural numbers and dispalys the
sum. sum = n*(n+1)/2
For n = 100 (octal 144), the sum is 5050 (octal 11672).

(You can tell your audience the story behind:
The 8-year old Carl Friedrich Gauss was forced in school to add
the numbers 1+2+3+..+100, because a lazy teacher wanted to keep the kids
occupied for an hour.
Gauss quickly reorderd the numbers to
(1+100) + (2+99) + (3+98) ... (50+51) = 101 * 50.
and had the result after a few seconds.
)


See the detailed instruction in the attachement.
(I hope the mailer doesn't strips them off)

regards
Joerg




Am 11.06.2012 10:37, schrieb Renaissance:

[Walter Banks]

Each computer had its own recognizable light patterns that you could see
the pulse of the machine. The first computer I ever had intimate contact
with was an IBM 1620 whose data was stored as a variable length BCD
number. Each digit contained 6 bits 4 for BCD number and some special
symbols a check digit and flag bit. If the LSdigit had a flag bit set the
number was negative. The length of the number was determined by
setting the flag bit on the MSdigit.

The lights on the IBM 1620 developed a very familiar pattern when it
executed a math instruction that couldn't find a end to the number it was
using in a computation. Time for a stop button and reload or fix and
reload.
It was possible for a result to to overwrite parts of the number or the
look up tables for add and multiply and lights would dance in a familiar
pattern as the instruction ran chasing its tail. The IBM had a big front
panel
and all those lights were incandescent. The lamps alone threw off quite a
bit of heat.


I had a research lab full of PDP-11's of various types, running it
displayed
the PC and R0. While running R0 was rarely very useful but the PC display
would give a indication of where a program was executing and often would
appear to freeze when the processor was in a short infinite loop. Halting
and
single stepping would then identify where the offending problem was.
The PDP-11 front panel was set up for hands on software debugging.

The PDP-11 UNIBUS was a backpanel bus that passed though all of the
peripheral slots. It was easy to add user specific hardware, it used TTL
levels
and most of the key computer signaling was there. The bus itself was
available
through the card slots or wire wrap pins on the back of the card slots. One

of our applications was not running even close to our execution time
expectations ( a protocol simulator taking a day or more per run) One
morning at least 12 hours after the expected completion halting and
single stepping did not give any real indication where it was spending its
time. The PDP-11 could be restarted by hitting RUN.

I soldered two groups of 4 resistors together as two simple 4 bit DAC's
and pulled a cover off the running computer and wire wrapped the resistor
DAC's to the MS 8 address lines. X-Y displaying the two outputs on an
oscilloscope produced 256 dots whose intensity directly was related to
time spent in that address group. A primitive but effective code profiling
display. For months it just dangled there getting a lot of use until it was

cleaned it up and brought out through two BNC connectors.

Walter Banks

[gareth]

"Rod Speed" <rod.sp...@gmail.com> wrote in message
news:a3tmag...@mid.individual.net...

Altogether now ...
16701...26...12702...352...

[gareth]

"Walter Banks" <wal...@bytecraft.com> wrote in message

news:4FD9D176...@bytecraft.com...

> gareth wrote:
>> "Walter Banks" <wal...@bytecraft.com> wrote in message
>> news:4FD8B400...@bytecraft.com...
>> >
>> > Nice job. Brings back old memories of sore fingers and a boot loader
>> > so familiar that it could be keyed in without thinking. Two minutes
>> > playing
>> > with your very nice simulation and I know why I seriously don't want
>> > to go back..
>>
>> But for those of us brought up in the lamps-and-switches world, being
>> able to have such intimate contact with the computer is useful on those
>> few occasions when debugging takes you down to that level.
>
>

> I had a research lab full of PDP-11's of various types, running it
> displayed
> the PC and R0.

ISTR that R0 was displayed during the RESET instruction or after a HALT.

Otherwise, you could program the lights yourself.

The first 11 I worked on, an 11/20 in 1971 had miniature incandescent wire
ended
bulbs and if one needed replacing, such was the tenacity of the edge
connectors, that
the force required to re-insert the card would cause other bulbs to fail
through mechanical shock.

[jmfbahciv]

Few occasions?!!!!! not in my shop. ;-)

/BAH

[jmfbahciv]

The TOPS-10 Monitor Distribution tape has a saveset of ANF-10
sources. this was all PDP-11 based. RSX-20F was also a
PDP-11 which was the front end to all PDP-10 KLs. All the diagnotic
sources for the -20F should be around somewhere. They all
should have code which intereacts with the terminal.

/BAH

[jmfbahciv]

But I never "saw" the binary. It was always octal. My eyes would bugger out
if I had to view 36 bits as only binary. Even 16 bits would make my eyes
cross if I had to spend even 2 minutes looking at it. Octal could be
watched for hours and remembered over long periods of time.

did anyone here really memorize a program in binary? Octal was the way to
do that.

/BAH

[jmfbahciv]

But that's what was so great about octal. I used the octal as shorthand
for binary.

I never understood the enamour for hex.

/BAH

[gareth]

"jmfbahciv" <See....@aol.com> wrote in message

news:PM0004C26...@aca2359b.ipt.aol.com...

In 1978 was working on a SCADA system written in assembler on an 11/05,
with no memory protection available. When the client reported strange
behaviour
followed by a crash, it was necessary to drop everything and rush to the
site
to examine the nature of the crash using the switches.

[Walter Banks]

gareth wrote:

> "Walter Banks" <wal...@bytecraft.com> wrote in message
> news:4FD9D176...@bytecraft.com...
> > gareth wrote:
> >> "Walter Banks" <wal...@bytecraft.com> wrote in message
> >> news:4FD8B400...@bytecraft.com...
> >> >
> >> > Nice job. Brings back old memories of sore fingers and a boot loader
> >> > so familiar that it could be keyed in without thinking. Two minutes
> >> > playing
> >> > with your very nice simulation and I know why I seriously don't want
> >> > to go back..
> >>
> >> But for those of us brought up in the lamps-and-switches world, being
> >> able to have such intimate contact with the computer is useful on those
> >> few occasions when debugging takes you down to that level.
> >
> >
> > I had a research lab full of PDP-11's of various types, running it
> > displayed
> > the PC and R0.
>
> ISTR that R0 was displayed during the RESET instruction or after a HALT.
>
> Otherwise, you could program the lights yourself.

Your right.

> The first 11 I worked on, an 11/20 in 1971 had miniature incandescent wire
> ended
> bulbs and if one needed replacing, such was the tenacity of the edge
> connectors, that
> the force required to re-insert the card would cause other bulbs to fail
> through mechanical shock.

I remember the bulb replacements all too well. I had two PDP-11's
around that time followed by 2 more a few months later. They were
the same computer except for the silk-screen that said the later two
were 11/20. Eventually we had 11 PDP-11's of various sorts from
LSI-11's to 11/45's

w..

[Ahem A Rivet's Shot]

On 14 Jun 2012 12:45:44 GMT

jmfbahciv <See....@aol.com> wrote:

> But that's what was so great about octal. I used the octal as shorthand
> for binary.
>
> I never understood the enamour for hex.

Hex does the same thing as octal, but for groups of four bits. When
your word is a multiple of four bits long and not a multiple of 3 bits then
hex fits better. That being said the 8 bit 8080 instruction set makes sense
in octal but not in hex.

--
Steve O'Hara-Smith | Directable Mirror Arrays
C:>WIN | A better way to focus the sun
The computer obeys and wins. | licences available see
You lose and Bill collects. | http://www.sohara.org/

[gareth]

"Ahem A Rivet's Shot" <ste...@eircom.net> wrote in message
news:20120614150740....@eircom.net...

> On 14 Jun 2012 12:45:44 GMT
> jmfbahciv <See....@aol.com> wrote:
>> But that's what was so great about octal. I used the octal as shorthand
>> for binary.
>> I never understood the enamour for hex.
> Hex does the same thing as octal, but for groups of four bits. When
> your word is a multiple of four bits long and not a multiple of 3 bits
> then
> hex fits better. That being said the 8 bit 8080 instruction set makes
> sense
> in octal but not in hex.

And if 'twere not for the S bit (Update condition codes) the ARM 32-bit
instruction set would hexise neatly.

[Joe Pfeiffer]

Dunno. Having that close contact early in my career was very valuable
for building appropriate intuitions in me, but today when I'm debugging
'way down at the hardware level a little LCD screen is much more
helpful...

[Joe Pfeiffer]

Whenever I'd be working at that level I'd be thinking in octal, but
writing 5 as 101 (or reading 011 as 3) was so automatic that toggling in
the program, or reading the blinkenlights, in binary wasn't an issue.

[Joe Pfeiffer]

The crucial thing is the width of the fields you're looking at. For a
machine like a -11 with lots of three bit fields, thinking in octal was
appropriate. For a machine like a VAX with lots of four bit fields, hex
was what worked better. For hex instruction notation to seemingly be
popular with x86 seems just perverse to me.

Hex has additional smaller benefits: representing a 16 bit quantity in
octal takes five digits and an extra bit, which isn't really a problem
but is a bit annoying to the OCD. You can write a value in fewer
characters with hex.

The drawback to hex is you have to memorize twice as many binary bit
patterns.

But of course, the fact that you can fill unused memory with 0xDEADBEEF
is enough reason to prefer hex all by itself!

[Dan Espen]

And of course, packed decimal would be un-intelligible in octal.

Packed +123:

x'123C'

--
Dan Espen

[gareth]

"Joe Pfeiffer" <pfei...@cs.nmsu.edu> wrote in message
news:1b4nqek...@pfeifferfamily.net...

...and when it is the LCD driver that is troublesome?

[David Dyer-Bennet]

I never even *saw* equipment of that generation. The IBM 1620 and 1401,
and the PDP-8s and -11s I worked with all had binary blinkenlight
panels, the KL-10s had none (except for the -11 front end, which had a
conventional 11 binary panel except in special colors). Nothing I
worked with after that had any kind of blinkenlights panel.
--
David Dyer-Bennet, dd...@dd-b.net; http://dd-b.net/
Snapshots: http://dd-b.net/dd-b/SnapshotAlbum/data/
Photos: http://dd-b.net/photography/gallery/
Dragaera: http://dragaera.info

[David Dyer-Bennet]

Walter Banks <wal...@bytecraft.com> writes:

> gareth wrote:
>
>> "Walter Banks" <wal...@bytecraft.com> wrote in message
>> news:4FD8B400...@bytecraft.com...
>> >
>> > Nice job. Brings back old memories of sore fingers and a boot loader
>> > so familiar that it could be keyed in without thinking. Two minutes
>> > playing
>> > with your very nice simulation and I know why I seriously don't want
>> > to go back..
>>
>> But for those of us brought up in the lamps-and-switches world, being
>> able to have such intimate contact with the computer is useful on those
>> few occasions when debugging takes you down to that level.
>
> Each computer had its own recognizable light patterns that you could see
> the pulse of the machine. The first computer I ever had intimate contact
> with was an IBM 1620 whose data was stored as a variable length BCD
> number. Each digit contained 6 bits 4 for BCD number and some special
> symbols a check digit and flag bit. If the LSdigit had a flag bit set the
> number was negative. The length of the number was determined by
> setting the flag bit on the MSdigit.
>
> The lights on the IBM 1620 developed a very familiar pattern when it
> executed a math instruction that couldn't find a end to the number it was
> using in a computation. Time for a stop button and reload or fix and
> reload.

Yep! I think the rhythm we saw was it wrapping around memory (we got
used to seeing this in the clear-core instructions we always started a
session with, right?).

> It was possible for a result to to overwrite parts of the number or
> the look up tables for add and multiply and lights would dance in a
> familiar pattern as the instruction ran chasing its tail. The IBM had
> a big front panel and all those lights were incandescent. The lamps
> alone threw off quite a bit of heat.

Yep. And for some reason lights burned out less often than in a the
later PDP-8 front panels I used.

> I had a research lab full of PDP-11's of various types, running it
> displayed the PC and R0. While running R0 was rarely very useful but
> the PC display would give a indication of where a program was
> executing and often would appear to freeze when the processor was in a
> short infinite loop. Halting and single stepping would then identify
> where the offending problem was. The PDP-11 front panel was set up
> for hands on software debugging.

I had mostly moved into higher-level languages and hence used symbolic
debuggers on the PDP-11, but sure, the front panel would have been
useful for OS-level debugging.

> The PDP-11 UNIBUS was a backpanel bus that passed though all of the
> peripheral slots. It was easy to add user specific hardware, it used
> TTL levels and most of the key computer signaling was there. The bus
> itself was available through the card slots or wire wrap pins on the
> back of the card slots. One of our applications was not running even
> close to our execution time expectations ( a protocol simulator taking
> a day or more per run) One morning at least 12 hours after the
> expected completion halting and single stepping did not give any real
> indication where it was spending its time. The PDP-11 could be
> restarted by hitting RUN.
>
> I soldered two groups of 4 resistors together as two simple 4 bit
> DAC's and pulled a cover off the running computer and wire wrapped the
> resistor DAC's to the MS 8 address lines. X-Y displaying the two
> outputs on an oscilloscope produced 256 dots whose intensity directly
> was related to time spent in that address group. A primitive but
> effective code profiling display. For months it just dangled there
> getting a lot of use until it was cleaned it up and brought out
> through two BNC connectors.

That's cool, never had to implement my own hardware debugging aids!
Much later, I worked with in-circuit emulators for Intel processors.

[Rod Speed]

jmfbahciv <See....@aol.com> wrote

> Charles Richmond wrote
>> jmfbahciv <See....@aol.com> wrote

>>> Michael Black wrote

>>>> jmfbahciv wrote

>>>>> Kewl. Let us know how it worked out. If the lights are realistic,
>>>>> I'd like to know how easy/hard it was for the kids of this pointy/
>>>>> clicky generation to use. oh, yea..and how much they liked it.

>>>> You move the mosue over the switch, and click the mouse button.
>>>> How hard can that be? Of course, if they don't know binary (it
>>>> wasn't octal, right?) and they don't know the op-codes, or know
>>>> how to program the tbing, they can't get very far.

>>> It was octal the last time I looked at the lights. I didn't read
>>> binary;
>>> I would think in octal.

>> BAH, I seem to recall that the switch panel had alternating colors
>> every 3 switches (or alternating color switch paddles), so one could
>> see the binary as entering in octal. In fact, I have heard that there
>> was a "three finger technique" whereby the operator could enter
>> octal digits using the individual bit paddles.

> But I never "saw" the binary. It was always octal. My eyes would bugger
> out if I had to view 36 bits as only binary. Even 16 bits would make my
> eyes cross if I had to spend even 2 minutes looking at it. Octal could be
> watched for hours and remembered over long periods of time.

> did anyone here really memorize a program in binary?
> Octal was the way to do that.

Yeah, the boot loader was certainly memorised in octal, not binary.

[Joe Pfeiffer]

That's the first thing I get working... the code for it is short enough
that swearing and inspection seems to be possible.

[Andrew Swallow]

On 14/06/2012 13:45, jmfbahciv wrote:
{snip}

> I never understood the enamour for hex.

BCD 8 and 9 take four bits so BCD numbers can be read straight off.

Andrew Swallow

[Walter Banks]

David Dyer-Bennet wrote:

> Walter Banks <wal...@bytecraft.com> writes:
>
> > The lights on the IBM 1620 developed a very familiar pattern when it
> > executed a math instruction that couldn't find a end to the number it was
> > using in a computation. Time for a stop button and reload or fix and
> > reload.
>
> Yep! I think the rhythm we saw was it wrapping around memory (we got
> used to seeing this in the clear-core instructions we always started a
> session with, right?).

It was a strange almost IBM1620 unique, every session was always
started by typing 260000800009 clearing memory and then loading
the add and multiply math tables. I had to dig out a IBM instruction
book to remember what the 26 opcode was actually called. I knew
it was some sort of move or copy. The IBM mnemonic was TF for
Transmit Field.

26 TF
00008 LS BCD digit of destination
00009 LS BCD digit of source

Copy source into destination and if the field flag was zero decrement
the source and destination addresses and do it again checking the
field flag each time. This cleared memory by copying a 0 from the
instruction each digit destroying any possibly of the field flag being
for the next digit transfer.

The IBM addressed memory a digit at a time. Instructions took
either 7 or 12 addresses. Memory fetch time was on the order of
20 micro seconds per digit pair. Yes a two operand fetch took
120 microseconds.

There were a number of interesting tricks that could be done with
the IBM 1620 just by changing the math tables. The math tables
were in unprotected data memory

As primitive as the IBM 1620 was data and executable code
was typed in from the console typewriter or loaded through
a card reader.

w..

[Rod Speed]

"Jörg Hoppe" <j_h...@t-online.de> wrote in message
news:jrcepn$8v0$1...@news.albasani.net...

> Hi,
>
> I sometimes do presentations with these panels and have a tutorial
> prepared.
>
> It consists of entering a little assembler program, verifying it and
> let it run.
> The program simple adds the first <n> natural numbers and dispalys the
> sum. sum = n*(n+1)/2
> For n = 100 (octal 144), the sum is 5050 (octal 11672).
>
> (You can tell your audience the story behind:
> The 8-year old Carl Friedrich Gauss was forced in school to add
> the numbers 1+2+3+..+100, because a lazy teacher wanted to keep the kids
> occupied for an hour.
> Gauss quickly reorderd the numbers to
> (1+100) + (2+99) + (3+98) ... (50+51) = 101 * 50.
> and had the result after a few seconds.
> )
>
>
> See the detailed instruction in the attachement.
> (I hope the mailer doesn't strips them off)

It isnt the mailer that strips them off, it's the news server.

And it got stripped by mine.

[Charles Richmond]

"Joe Pfeiffer" <pfei...@cs.nmsu.edu> wrote in message

news:1bvciuj...@pfeifferfamily.net...

>
>
<
> But of course, the fact that you can fill unused memory with 0xDEADBEEF
> is enough reason to prefer hex all by itself!

You can also fill memory with 0xDEADFACE.

--

numerist at aquaporin4 dot com

[Charles Richmond]

"Jörg Hoppe" <j_h...@t-online.de> wrote in message
news:jrcepn$8v0$1...@news.albasani.net...

> Hi,
>
> I sometimes do presentations with these panels and have a tutorial
> prepared.
>
> It consists of entering a little assembler program, verifying it and
> let it run.
> The program simple adds the first <n> natural numbers and dispalys the
> sum. sum = n*(n+1)/2
> For n = 100 (octal 144), the sum is 5050 (octal 11672).
>
> (You can tell your audience the story behind:
> The 8-year old Carl Friedrich Gauss was forced in school to add
> the numbers 1+2+3+..+100, because a lazy teacher wanted to keep the kids
> occupied for an hour.
> Gauss quickly reorderd the numbers to
> (1+100) + (2+99) + (3+98) ... (50+51) = 101 * 50.
> and had the result after a few seconds.
> )
>
>

The way I heard the story, Gausss noticed the following:

1 2 3 4 5 ... 98 99 100
100 99 98 97 96 3 2 1

That woud be writing down the numbers *twice* but putting the bottom row in
reverse order. He noted that the sum of each column was 101, which is 100 +
1. Since there were 100 columns, you had (100+1)*100. And of course, since
you had each number twice, you had to divide by 2.

I guess the only difference in our stories is... my version emphasizes the
two rows of numbers written one above the other.

That's the story I heard about how Gauss developed this formula in the first
grade.

[Ben Pfaff]

"Charles Richmond" <nume...@aquaporin4.com> writes:

> The way I heard the story, Gausss noticed the following:
>
> 1 2 3 4 5 ... 98 99 100
> 100 99 98 97 96 3 2 1
>
> That woud be writing down the numbers *twice* but putting the bottom
> row in reverse order. He noted that the sum of each column was 101,
> which is 100 + 1. Since there were 100 columns, you had (100+1)*100.
> And of course, since you had each number twice, you had to divide by
> 2.

I had to take a BASIC programming class in grade school. We were
assigned to write a program to print the sum of 1 through N, with
N an integer input by the user. I turned in the program:

10 INPUT N
20 PRINT N*(N+1)/2

and received a failing grade.

[Dan Espen]

Oddly, our 5th grade teacher asked the same question.

Yes, I got the answer.
Seemed simple enough.

--
Dan Espen

[Peter Flass]

On 6/14/2012 7:56 AM, Walter Banks wrote:

>
> gareth wrote:
>
>> "Walter Banks"<wal...@bytecraft.com> wrote in message
>> news:4FD8B400...@bytecraft.com...
>>>
>>> Nice job. Brings back old memories of sore fingers and a boot loader
>>> so familiar that it could be keyed in without thinking. Two minutes
>>> playing
>>> with your very nice simulation and I know why I seriously don't want
>>> to go back..
>>
>> But for those of us brought up in the lamps-and-switches world, being
>> able to have such intimate contact with the computer is useful on those
>> few occasions when debugging takes you down to that level.
>

> Each computer had its own recognizable light patterns that you could see
> the pulse of the machine. The first computer I ever had intimate contact
> with was an IBM 1620 whose data was stored as a variable length BCD
> number. Each digit contained 6 bits 4 for BCD number and some special
> symbols a check digit and flag bit. If the LSdigit had a flag bit set the
> number was negative. The length of the number was determined by
> setting the flag bit on the MSdigit.
>

> The lights on the IBM 1620 developed a very familiar pattern when it
> executed a math instruction that couldn't find a end to the number it was
> using in a computation. Time for a stop button and reload or fix and
> reload.

> It was possible for a result to to overwrite parts of the number or the
> look up tables for add and multiply and lights would dance in a familiar
> pattern as the instruction ran chasing its tail. The IBM had a big front
> panel
> and all those lights were incandescent. The lamps alone threw off quite a
> bit of heat.
>

For some reason debugging from the front panel seemed easier than
running a debugger today, although they each have their uses. Maybe it
was because the dials and switches were better designed for ham-fingered
hands than mice and keyboards,


--
Pete

[Peter Flass]

On 6/14/2012 8:45 AM, jmfbahciv wrote:
>
> But that's what was so great about octal. I used the octal as shorthand
> for binary.
>
> I never understood the enamour for hex.
>

Hex is basically the same shorthand for binary, better suited to 8-bit
bytes, while octal is great for sixbit. You worked mostly with PDP-10s
where octal was a good fit. DEC also used it for the -11, where it did
not suit as well.

The 16-bit value -16657 is hex 'BEEF'x, in octal it's '137357'B3. If
you wanted to know what the individual (8-bit) bytes contain it's easy
to see in hex, but somewhat harder to see in octal.

--
Pete

[Harry Vaderchi]

On Thu, 14 Jun 2012 21:14:19 +0100, Charles Richmond
<nume...@aquaporin4.com> wrote:

> "Joe Pfeiffer" <pfei...@cs.nmsu.edu> wrote in message
> news:1bvciuj...@pfeifferfamily.net...
>>
>>
> <
>> But of course, the fact that you can fill unused memory with 0xDEADBEEF
>> is enough reason to prefer hex all by itself!
>
> You can also fill memory with 0xDEADFACE.
>

1DEFAC8
2B00B1E5


Hey, we have this competetion in the UK, to make the daftest personalised
number plate.

> --
>
> numerist at aquaporin4 dot com
>

--
[dash dash space newline 4line sig]

Albi CNU

[gareth]

>>>
>>> THX! :-) At the company where I work, I planned an internal course of
>>> computer architecture and operating systems, and I chose the pdp-11 for
>>> a brief practical course of assembly language (since it's didactically
>>> perfect for the purpose). I'll use the frontpanel simulator to make
>>> things more interesting. :-)

(Not sure of the original poster for the above)

When I was at the University of Essex in 1972, reading Computer and
Communication
Engineering as a specialisation of electronics, they used the PDP8 as the
example for both
hardware design and system software design.

Despite feeling avant garde at the time, it's all dinosaur technology today,
with no
call for an ability to design the electronics to drive a 3D-organised core
memory.

Half-select noise, anybody?

[jmfbahciv]

gareth wrote:
> "Walter Banks" <wal...@bytecraft.com> wrote in message

> news:4FD9D176...@bytecraft.com...

>> gareth wrote:
>>> "Walter Banks" <wal...@bytecraft.com> wrote in message
>>> news:4FD8B400...@bytecraft.com...
>>> >
>>> > Nice job. Brings back old memories of sore fingers and a boot loader
>>> > so familiar that it could be keyed in without thinking. Two minutes
>>> > playing
>>> > with your very nice simulation and I know why I seriously don't want
>>> > to go back..
>>>
>>> But for those of us brought up in the lamps-and-switches world, being
>>> able to have such intimate contact with the computer is useful on those
>>> few occasions when debugging takes you down to that level.
>>
>>

>> I had a research lab full of PDP-11's of various types, running it
>> displayed
>> the PC and R0.
>

> ISTR that R0 was displayed during the RESET instruction or after a HALT.
>
> Otherwise, you could program the lights yourself.
>

> The first 11 I worked on, an 11/20 in 1971 had miniature incandescent wire
> ended
> bulbs and if one needed replacing, such was the tenacity of the edge
> connectors, that
> the force required to re-insert the card would cause other bulbs to fail
> through mechanical shock.
>
>

Did you win when you played Whack-a-mole?

/BAH

[jmfbahciv]

Peter Flass wrote:
> On 6/14/2012 8:45 AM, jmfbahciv wrote:
>>
>> But that's what was so great about octal. I used the octal as shorthand
>> for binary.
>>
>> I never understood the enamour for hex.
>>
>
> Hex is basically the same shorthand for binary, better suited to 8-bit
> bytes, while octal is great for sixbit.

We generally didn't deal with character data when debugging. Most
of our work was delivery of I/O or calculations.

>You worked mostly with PDP-10s
> where octal was a good fit. DEC also used it for the -11, where it did
> not suit as well.
>
> The 16-bit value -16657 is hex 'BEEF'x, in octal it's '137357'B3. If
> you wanted to know what the individual (8-bit) bytes contain it's easy
> to see in hex, but somewhat harder to see in octal.
>

I'm not writing well; it's easier to remember and "see" binary using
octal. hex just seems to be too complicated by I've been breathing
DEC gear most of my computing work life.

/BAH

[jmfbahciv]

/DBD didn't do TOPS-10 lanugage work. There were blinken lights
in the middle and other end of the machine room. KL10s and the 11/70.

[jmfbahciv]

ROTFLMAO. Both of you are funny. What's missing is the opportunity
to kick the damned thing as a last resort, and getting it to work.

/BAH

[Charles Richmond]

There was a book written in the 1970's and the author said:

"Know the truth and the truth will get you a B. Know what the teacher
wants, and you get an A."

When Fred Smith was in college, he submitted the idea for Federal Express as
an assignment in a business class. The idea of course was to fly all the
packages to a central location (Memphis), sort them, and send them out to
arrive the next day. His professor gave him a C on the project. The
professor said it was *not* practical to do this.

In my experience, grade school is the *most* intolerant of alternate ideas.
If what you put down.... does *not* agree with what the teacher expects to
get, then you're *wrong*. There is *no* debate. Also, grade school
teachers have been know to subject a student with different ideas... to the
riducule of the entire class. I had a cow-orker back in the 80's... who
told the teacher in the 6th grade that he read that Christ was *not* born in
the year 0. Boy did he get riducule for that one!!!

In another newsgroup, I picked up the following piece of wisdom:

"If you're *not* catching FLAK, then you're *not* over the target." :-)

[Jean-Marc Bourguet]

"Charles Richmond" <nume...@aquaporin4.com> writes:

> "If you're *not* catching FLAK, then you're *not* over the target." :-)

The problem is that too many conclude that because they are catching
FLAK, they are over the target.

Yours,

--
Jean-Marc

[Charles Richmond]

"Peter Flass" <Peter...@Yahoo.com> wrote in message
news:jrdmnd$1i3$1...@dont-email.me...

To help remember the hex bit patterns, I found it helpful to remember:

A in hex is 10 decimal, and the binary is 1010. So 1011 would be B hex.

F in hex is of course all one's in binary 1111. One less is E hex... 1110.

C in hex is 12 decimal, and has a characteristic pattern of 1100. One more
is D hex... 1101. That was easy for me to remember somehow.

9 in hex is like a split in bowling... 1001. And 8 in hex is one less...
1000.


And to convert from hex to octal, (and vice versa) the easiest way is go
through binary.

0xDEAD => 1101 1110 1010 1101 => 1 101 111 010 101 101 => 0157255

0xBEEF => 1011 1110 1110 1111 => 1 011 111 011 101 111 => 0135357

0xFACE => 1111 1010 1100 1110 => 1 111 101 011 001 110 => 0175316

Just re-arrange the bits in groups of three instead of four.

[Joe Pfeiffer]

Yeah, it's definitely "kick it" OR "get it to work".

[Charles Richmond]

"jmfbahciv" <See....@aol.com> wrote in message
news:PM0004C28...@aca23683.ipt.aol.com...
> gareth wrote:
>>
>> [snip...] [snip...]
>> [snip...]

>>
>> The first 11 I worked on, an 11/20 in 1971 had miniature incandescent
>> wire
>> ended
>> bulbs and if one needed replacing, such was the tenacity of the edge
>> connectors, that
>> the force required to re-insert the card would cause other bulbs to fail
>> through mechanical shock.
>>
>>
> Did you win when you played Whack-a-mole?
>

It's too bad that the blinkenlight machines are *not* still prominent. The
LED was a Godsend for blinkenlights!!! I remember in the late 70's at
univeristy, there were Harris minicomputers running the remote job
submission systems (printer and card reader) in the Engineering and Business
buildings. The remotes interfaced to an IBM 370/155 in the administration
building basement. The blinkenlights on those *were* LED's. LED's last
longer, survive greater mechanical shock, switch the light faster, and are
cheaper than the old incandescent bulbs used in the PDP-8 front panels.

Just when electronics was getting good at blinkenlights, the trend in
computers was to get rid of them altogether. :-)

[Ben Pfaff]

"Charles Richmond" <nume...@aquaporin4.com> writes:

>> I had to take a BASIC programming class in grade school. We were
>> assigned to write a program to print the sum of 1 through N, with
>> N an integer input by the user. I turned in the program:
>>
>> 10 INPUT N
>> 20 PRINT N*(N+1)/2
>>
>> and received a failing grade.
>
> There was a book written in the 1970's and the author said:
>
> "Know the truth and the truth will get you a B. Know what the teacher
> wants, and you get an A."

I was, of course, ignoring the teacher's intent, since the
section of the textbook we had read that day was about FOR loops.
However, the assignment did *not* say to use a FOR loop.

[Dan Espen]

In grade school we had a homework assignment to draw
a bunch of triangles with a compass based on their dimensions.

One of the "triangles" had dimensions 1, 2, 3.

Took me at least 15 minutes and another teacher coming into
the room to explain to the teacher that although she could
draw a triangle with those dimensions on the blackboard
using her chalk compass, there was no such triangle.

The whole time the teacher and the class are looking at
me like I'm an idiot.

--
Dan Espen

[Charlie Gibbs]

In article <1bvciuj...@pfeifferfamily.net>, pfei...@cs.nmsu.edu
(Joe Pfeiffer) writes:

> jmfbahciv <See....@aol.com> writes:
>
>> I never understood the enamour for hex.
>

> The crucial thing is the width of the fields you're looking at. For
> a machine like a -11 with lots of three bit fields, thinking in octal
> was appropriate. For a machine like a VAX with lots of four bit
> fields, hex was what worked better. For hex instruction notation
> to seemingly be popular with x86 seems just perverse to me.
>
> Hex has additional smaller benefits: representing a 16 bit quantity
> in octal takes five digits and an extra bit, which isn't really a
> problem but is a bit annoying to the OCD. You can write a value in
> fewer characters with hex.
>
> The drawback to hex is you have to memorize twice as many binary bit
> patterns.

You run into this problem with octal too, if you're dealing with
values whose length is a multiple of 4 bits. What's easier:
0xFFFF or 377377? I could never be bothered worrying about
skipping bits here and there.

It's all a matter of what you're used to. I started with the IBM 360
architecture, and cut my teeth on hex. (Once, in a fit of nerdiness,
I did a long division in hex.) When Canada went metric, I had no
problem with my speedometer - I converted its readings from miles
per hour to km/h by simply reading the numbers in hex.

> But of course, the fact that you can fill unused memory with
> 0xDEADBEEF is enough reason to prefer hex all by itself!

Sounds good. Let's go for 0xC0FFEE.

--
/~\ cgi...@kltpzyxm.invalid (Charlie Gibbs)
\ / I'm really at ac.dekanfrus if you read it the right way.
X Top-posted messages will probably be ignored. See RFC1855.
/ \ HTML will DEFINITELY be ignored. Join the ASCII ribbon campaign!

[Charlie Gibbs]

In article <jrdm2p$tl0$1...@dont-email.me>, Peter...@Yahoo.com

(Peter Flass) writes:

> For some reason debugging from the front panel seemed easier than
> running a debugger today, although they each have their uses.
> Maybe it was because the dials and switches were better designed
> for ham-fingered hands than mice and keyboards,

Or maybe because the bureaucracy of your typical IDE makes it
just not worth the effort. I do everything with makefiles, but
once while trying to track down a nasty bug I thought I'd load
my program into the debugger that came with Borland BCB4. But
it wanted me to set up a "project", re-define all my dependencies,
and basically create several megabytes of crap just to compile my
100K program. I realized that it would be faster to fall back on
that old standby of sprinkling printf()s through my code.

I miss the way it was done on my old MS-DOS Lattice compiler -
add an option flag and the compiler would include debugging
information in the object file, which would be picked up by
their freestanding debugger. Simple, effective... and totally
boring to today's gadget freaks.

[Charlie Gibbs]

In article <ylfkhaud...@dd-b.net>, dd...@dd-b.net
(David Dyer-Bennet) writes:

> Walter Banks <wal...@bytecraft.com> writes:
>
>> The lights on the IBM 1620 developed a very familiar pattern when it
>> executed a math instruction that couldn't find a end to the number
>> it was using in a computation. Time for a stop button and reload or
>> fix and reload.
>

> Yep! I think the rhythm we saw was it wrapping around memory (we got
> used to seeing this in the clear-core instructions we always started
> a session with, right?).

I didn't get to use a 1620 much, but I still remember 310000300002<R-S>.

I remember the sinking feeling I'd get upon seeing a similar pattern
on the front panel lights of my IMSAI. We called it a "stack loop",
because it was usually caused by some sort of recursion or stack
handling problem. The sinking feeling came from the knowledge that
all of memory - including all that code you had painstakingly toggled
in through the front panel - was now wiped.

[Ahem A Rivet's Shot]

On Fri, 15 Jun 2012 11:50:02 -0400
Dan Espen <des...@verizon.net> wrote:

> In grade school we had a homework assignment to draw
> a bunch of triangles with a compass based on their dimensions.
>
> One of the "triangles" had dimensions 1, 2, 3.

I don't know, bloody degenerates.

--
Steve O'Hara-Smith | Directable Mirror Arrays
C:>WIN | A better way to focus the sun
The computer obeys and wins. | licences available see
You lose and Bill collects. | http://www.sohara.org/

[Joe Pfeiffer]

"Charles Richmond" <nume...@aquaporin4.com> writes:

> "Peter Flass" <Peter...@Yahoo.com> wrote in message
> news:jrdmnd$1i3$1...@dont-email.me...
>> On 6/14/2012 8:45 AM, jmfbahciv wrote:
>>>
>>> But that's what was so great about octal. I used the octal as shorthand
>>> for binary.
>>>
>>> I never understood the enamour for hex.
>>>
>>
>> Hex is basically the same shorthand for binary, better suited to
>> 8-bit bytes, while octal is great for sixbit. You worked mostly
>> with PDP-10s where octal was a good fit. DEC also used it for the
>> -11, where it did not suit as well.
>>
>> The 16-bit value -16657 is hex 'BEEF'x, in octal it's '137357'B3.
>> If you wanted to know what the individual (8-bit) bytes contain it's
>> easy to see in hex, but somewhat harder to see in octal.
>>
>
> To help remember the hex bit patterns, I found it helpful to remember:
>
> A in hex is 10 decimal, and the binary is 1010. So 1011 would be B hex.
>
> F in hex is of course all one's in binary 1111. One less is E hex... 1110.
>
> C in hex is 12 decimal, and has a characteristic pattern of 1100. One
> more is D hex... 1101. That was easy for me to remember somehow.
>
> 9 in hex is like a split in bowling... 1001. And 8 in hex is one
> less... 1000.

That's more work than just remembering...

> And to convert from hex to octal, (and vice versa) the easiest way is
> go through binary.
>
> 0xDEAD => 1101 1110 1010 1101 => 1 101 111 010 101 101 => 0157255
>
> 0xBEEF => 1011 1110 1110 1111 => 1 011 111 011 101 111 => 0135357
>
> 0xFACE => 1111 1010 1100 1110 => 1 111 101 011 001 110 => 0175316
>
> Just re-arrange the bits in groups of three instead of four.

Yes.

[Joe Pfeiffer]

Dan Espen <des...@verizon.net> writes:
>
> In grade school we had a homework assignment to draw
> a bunch of triangles with a compass based on their dimensions.
>
> One of the "triangles" had dimensions 1, 2, 3.
>
> Took me at least 15 minutes and another teacher coming into
> the room to explain to the teacher that although she could
> draw a triangle with those dimensions on the blackboard
> using her chalk compass, there was no such triangle.

Do you mean she could draw a triangle at label it as having dimensions
1, 2, 3 (as opposed to actually dawing a triangle with those
dimensions)? Because I don't see any way using a chalk compass is going
to make it any more possible...

[Joe Pfeiffer]

"Charlie Gibbs" <cgi...@kltpzyxm.invalid> writes:
>
>> But of course, the fact that you can fill unused memory with
>> 0xDEADBEEF is enough reason to prefer hex all by itself!
>
> Sounds good. Let's go for 0xC0FFEE.

I'll have to remember that one.

[Peter Flass]

On 6/15/2012 10:52 AM, Charles Richmond wrote:
>
>
> And to convert from hex to octal, (and vice versa) the easiest way is go
> through binary.

Actually, the *easiest* way is to use a calculator, but otherwise you're
right. I suppose you could find a method of converting directly, but
the patterns would only repeat every twelve bits.

>
> numerist at aquaporin4 dot com
>

--
Pete

[Peter Flass]

On 6/15/2012 12:13 PM, Charlie Gibbs wrote:
> In article<jrdm2p$tl0$1...@dont-email.me>, Peter...@Yahoo.com
> (Peter Flass) writes:
>
>> For some reason debugging from the front panel seemed easier than
>> running a debugger today, although they each have their uses.
>> Maybe it was because the dials and switches were better designed
>> for ham-fingered hands than mice and keyboards,
>
> Or maybe because the bureaucracy of your typical IDE makes it
> just not worth the effort. I do everything with makefiles, but
> once while trying to track down a nasty bug I thought I'd load
> my program into the debugger that came with Borland BCB4. But
> it wanted me to set up a "project", re-define all my dependencies,
> and basically create several megabytes of crap just to compile my
> 100K program. I realized that it would be faster to fall back on
> that old standby of sprinkling printf()s through my code.
>
> I miss the way it was done on my old MS-DOS Lattice compiler -
> add an option flag and the compiler would include debugging
> information in the object file, which would be picked up by
> their freestanding debugger. Simple, effective... and totally
> boring to today's gadget freaks.
>

I use gdb and its front-end DDD. Not my favorite, but it works.


--
Pete

[Joe Pfeiffer]

I use gdb from inside emacs. I find emacs combined with makefiles and
gdb is the best IDE I've ever seen.

[Dan Espen]

Joe Pfeiffer <pfei...@cs.nmsu.edu> writes:

> Dan Espen <des...@verizon.net> writes:
>>
>> In grade school we had a homework assignment to draw
>> a bunch of triangles with a compass based on their dimensions.
>>
>> One of the "triangles" had dimensions 1, 2, 3.
>>
>> Took me at least 15 minutes and another teacher coming into
>> the room to explain to the teacher that although she could
>> draw a triangle with those dimensions on the blackboard
>> using her chalk compass, there was no such triangle.
>
> Do you mean she could draw a triangle at label it as having dimensions
> 1, 2, 3 (as opposed to actually dawing a triangle with those
> dimensions)? Because I don't see any way using a chalk compass is going
> to make it any more possible...

You need to try it.

I remember expecting to see the 2 curves meet right on the line,
but the inherit inaccuracy of the real world let the teacher and
the students that did the homework point at their squatty little
triangles and say, "can't you see?".

--
Dan Espen

[Ibmekon]

On Fri, 15 Jun 2012 14:33:59 -0400, Dan Espen <des...@verizon.net>
wrote:

Or you can cheat...

http://ostermiller.org/calc/triangle.html

says..
-----------------------------------------------------------------------------
The values entered cannot make a triangle. The sum of the length of
every two sides must exceed the length of the third side.
-----------------------------------------------------------------------------

Reminds me of a pub question - that I puzzled over, and gave up on.

Image a nice high wall.
Image a box - a 1 meter cube.
Image a ladder 10 meters long.

Push the box against the wall
Put the ladder against the wall, touching the box.

Now give a calculation showing the exact height the ladder reaches on
the wall.

Carl Goldsworthy






wall.

[Joe Pfeiffer]

Ibmekon writes:
>
> Reminds me of a pub question - that I puzzled over, and gave up on.
>
> Image a nice high wall.
> Image a box - a 1 meter cube.
> Image a ladder 10 meters long.
>
> Push the box against the wall
> Put the ladder against the wall, touching the box.
>
> Now give a calculation showing the exact height the ladder reaches on
> the wall.

Gave up on, or saw that it couldn't be answered?

[Peter Flass]

It's a right triangle: 1**2 + x**2 = 10**2; x = sqrt(99), whatever that is.


--
Pete

[Joe Pfeiffer]

Oh, wait. Misread it... I'm ending up with a fourth degree polynomial,
and my algebra is too rusty to remember where to go next (I do remember
that they don't in general have a closed form solution, but I don't
remember how to spot the ones that do or what to do with them).

[Ibmekon]

Just to be clear.. the right angle triangle has sides of -

The height at which the ladder touches the wall - unknown
The hypotenuse ie ladder - known.
The base - the box plus a bit extra of ground - unknown

Carl Goldsworthy

[Joe Pfeiffer]

I think we're interpreting the question differently. I've got the box
on the ground up against the wall; the ladder's legs are on the ground
and it's going up at such an angle that it touches the corner of the box
(one meter up from the ground, one meter out from the wall) and then
goes on up to rest against the wall at the top.

So I've got two right triangles: one from the ground to the corner of
the box (x1 unknown, y1=1), and one from the corner of the box to the wall
(y2 unknown, x2=1). There is a third right triangle constituting the
entire ladder, with y=y2+1 and x=x1+1. x**2 + y**2 = 100.

I'm not seeing how this simplifies to your equation...

[Ahem A Rivet's Shot]

It must be possible to work it out, there's only one way the ladder
can touch the wall, the floor and the edge of the cube, but it's not
obvious how to pin it down.

[Rod Speed]

"Ahem A Rivet's Shot" <ste...@eircom.net> wrote in message
news:20120615235521....@eircom.net...

> On Fri, 15 Jun 2012 23:26:18 +0100
> Ibmekon wrote:
>
>> On Fri, 15 Jun 2012 18:08:29 -0400, Peter Flass
>> <Peter...@Yahoo.com> wrote:
>>
>> >On 6/15/2012 5:56 PM, Joe Pfeiffer wrote:
>> >> Ibmekon writes:
>> >>>
>> >>> Reminds me of a pub question - that I puzzled over, and gave up on.
>> >>>
>> >>> Image a nice high wall.
>> >>> Image a box - a 1 meter cube.
>> >>> Image a ladder 10 meters long.
>> >>>
>> >>> Push the box against the wall
>> >>> Put the ladder against the wall, touching the box.
>> >>>
>> >>> Now give a calculation showing the exact height the ladder reaches on
>> >>> the wall.
>> >>
>> >> Gave up on, or saw that it couldn't be answered?
>> >
>> >It's a right triangle: 1**2 + x**2 = 10**2; x = sqrt(99), whatever that
>> >is.
>>
>> Just to be clear.. the right angle triangle has sides of -
>>
>> The height at which the ladder touches the wall - unknown
>> The hypotenuse ie ladder - known.
>> The base - the box plus a bit extra of ground - unknown
>
> It must be possible to work it out, there's only one way the ladder
> can touch the wall, the floor and the edge of the cube, but it's not
> obvious how to pin it down.

Depends on what you mean by 'touches the box'

If the side of the ladder touches the box and the ladder
is flat on the wall, the answer is obviously 10m.

This one must be googlable.

[Joe Pfeiffer]

The problem statement said "exact", which I've interpreted as meaning
closed form. I've gotten it to a form that a Newtonian approximation
could get arbitrarily close, but I don't have a closed-form solution.

And a bit of reflection shows that if the angle is anything other than
45 degrees (which it trivially is), there are two solutions.

[Ahem A Rivet's Shot]

OK let x be the intercept on the x-axis and y be the intercept on
the y-axis. Then x^2+y^2 = 100 and x = y/(y-1) - so

y^2/(y-1)^2 + y^2 = 100

y^2 + y^2(y-1)^2 = 100(y-1)^2

y^2 + y^4 - 2y^3 + y^2 = 100y^2 - 200y + 100

y^4 - 2y^3 - 98y^2 + 200y - 100 = 0

Same fourth order polynomial Joe ? I also cannot recall how to get
out of this mess. ISTM there must be a simpler way.

[Ahem A Rivet's Shot]

Nope, the ladder touches the cube at the top, not the bottom, so
it's more than 1m out from the wall at the bottom.

[David Dyer-Bennet]

Walter Banks <wal...@bytecraft.com> writes:

> David Dyer-Bennet wrote:
>
>> Walter Banks <wal...@bytecraft.com> writes:
>>
>> > The lights on the IBM 1620 developed a very familiar pattern when it
>> > executed a math instruction that couldn't find a end to the number it was
>> > using in a computation. Time for a stop button and reload or fix and
>> > reload.
>>
>> Yep! I think the rhythm we saw was it wrapping around memory (we got
>> used to seeing this in the clear-core instructions we always started a
>> session with, right?).
>

> It was a strange almost IBM1620 unique, every session was always
> started by typing 260000800009 clearing memory and then loading
> the add and multiply math tables. I had to dig out a IBM instruction
> book to remember what the 26 opcode was actually called. I knew
> it was some sort of move or copy. The IBM mnemonic was TF for
> Transmit Field.
>
> 26 TF
> 00008 LS BCD digit of destination
> 00009 LS BCD digit of source
>
> Copy source into destination and if the field flag was zero decrement
> the source and destination addresses and do it again checking the
> field flag each time. This cleared memory by copying a 0 from the
> instruction each digit destroying any possibly of the field flag being
> for the next digit transfer.

Yep. You could also do it with 16 TFM (the "immediate" form of Transmit
Field) and with Transmit Record, which *think* had op-code 31...yep.

> The IBM addressed memory a digit at a time. Instructions took
> either 7 or 12 addresses. Memory fetch time was on the order of
> 20 micro seconds per digit pair. Yes a two operand fetch took
> 120 microseconds.

Um, one or two addresses. Instructions were always 12 digits long; two
for the opcode and two 5-digit fields (two addresses, an address plus a
5-digit immediate operand, and a few weird cases I'm sure).

Bitsavers.org has the systems reference manual for the model 1 1620; TF
takes 160 + 40 D[q] microseconds (D[q] being the number of digits in the
field being transmitted) to execute.

> There were a number of interesting tricks that could be done with
> the IBM 1620 just by changing the math tables. The math tables
> were in unprotected data memory

Although one fairly common RFQ was a memory-protect switch for the add
tables.

> As primitive as the IBM 1620 was data and executable code
> was typed in from the console typewriter or loaded through
> a card reader.

They could be equipped with paper tape readers, or with disk drives,
though the ones I used had neither. Cards weren't "primitive" at the
time :-).
--
David Dyer-Bennet, dd...@dd-b.net; http://dd-b.net/
Snapshots: http://dd-b.net/dd-b/SnapshotAlbum/data/
Photos: http://dd-b.net/photography/gallery/
Dragaera: http://dragaera.info

[David Dyer-Bennet]

"Charlie Gibbs" <cgi...@kltpzyxm.invalid> writes:

> In article <ylfkhaud...@dd-b.net>, dd...@dd-b.net
> (David Dyer-Bennet) writes:
>
>> Walter Banks <wal...@bytecraft.com> writes:
>>
>>> The lights on the IBM 1620 developed a very familiar pattern when it
>>> executed a math instruction that couldn't find a end to the number
>>> it was using in a computation. Time for a stop button and reload or
>>> fix and reload.
>>
>> Yep! I think the rhythm we saw was it wrapping around memory (we got
>> used to seeing this in the clear-core instructions we always started
>> a session with, right?).
>
> I didn't get to use a 1620 much, but I still remember 310000300002<R-S>.

Where I was, that was one of the "unusual" ones (all the rest I know had
*increasing* addresses in the two operatnds, but TR works backwards
through memory compared to TF, so you needed decreasing addresses
instead).

> I remember the sinking feeling I'd get upon seeing a similar pattern
> on the front panel lights of my IMSAI. We called it a "stack loop",
> because it was usually caused by some sort of recursion or stack
> handling problem. The sinking feeling came from the knowledge that
> all of memory - including all that code you had painstakingly toggled
> in through the front panel - was now wiped.

I never did work with a computer so primitive I had to enter large
amounts of code from the front panel. I avoided the nasty new-fangled
micro-computers for years because they were so immensely inferior to
what I was used to at work; they weren't tempting at all.

[David Dyer-Bennet]

Ben Pfaff <b...@cs.stanford.edu> writes:

> "Charles Richmond" <nume...@aquaporin4.com> writes:
>
>>> I had to take a BASIC programming class in grade school. We were
>>> assigned to write a program to print the sum of 1 through N, with
>>> N an integer input by the user. I turned in the program:
>>>
>>> 10 INPUT N
>>> 20 PRINT N*(N+1)/2
>>>
>>> and received a failing grade.
>>
>> There was a book written in the 1970's and the author said:
>>
>> "Know the truth and the truth will get you a B. Know what the teacher
>> wants, and you get an A."
>

> I was, of course, ignoring the teacher's intent, since the
> section of the textbook we had read that day was about FOR loops.
> However, the assignment did *not* say to use a FOR loop.

I'm really not sure why I didn't have more trouble with this in school.
I think I *mostly* gave them what they wanted unless it was actively
wrong, maybe.

Well, not with programming; that wasn't offered until 10th grade, at
which point I was professionally employed writing assembly code, so
taking their introductory FORTRAN course wasn't very tempting.

[David Dyer-Bennet]

Dan Espen <des...@verizon.net> writes:

> In grade school we had a homework assignment to draw
> a bunch of triangles with a compass based on their dimensions.

And here I thought compasses were for drawing circles! :-)

> One of the "triangles" had dimensions 1, 2, 3.
>
> Took me at least 15 minutes and another teacher coming into
> the room to explain to the teacher that although she could
> draw a triangle with those dimensions on the blackboard
> using her chalk compass, there was no such triangle.

One of the problems with teachers requiring the exact answer they're
expecting, I think, is that in grade school they're often made to teach
stuff they don't understand. That makes it harder to figure out that an
off-beat answer is correct.

> The whole time the teacher and the class are looking at
> me like I'm an idiot.

I think they'd all learned better by then, in my case.

[David Dyer-Bennet]

Joe Pfeiffer <pfei...@cs.nmsu.edu> writes:

> jmfbahciv <See....@aol.com> writes:
>
>> Joe Pfeiffer wrote:
>>> jmfbahciv <See....@aol.com> writes:
>>>
>>>> Michael Black wrote:
>>>>> On Mon, 11 Jun 2012, jmfbahciv wrote:
>>>>>
>>>>>
>>>>>> Kewl. Let us know how it worked out. If the lights are realistic,
>>>>>> I'd like to know how easy/hard it was for the kids of this
>>>>>> pointy/clicky generation to use. oh, yea..and how much they
>>>>>> liked it.
>>>>>>
>>>>> You move the mosue over the switch, and click the mouse button. How hard
>>>>> can that be? Of course, if they don't know binary (it wasn't octal,
>>>>> right?) and they don't know the op-codes, or know how to program the
>>>>> tbing, they can't get very far.
>>>>
>>>> It was octal the last time I looked at the lights. I didn't read binary;
>>>> I would think in octal.
>>>
>>> The few -11 front panels I saw were binary, but with the lights in
>>> groups of three (there weren't spacing differences, just markings in the
>>> silkscreen) so reading them off in octal was automatic.

>>
>> But that's what was so great about octal. I used the octal as shorthand
>> for binary.
>>
>> I never understood the enamour for hex.
>

> The crucial thing is the width of the fields you're looking at. For a
> machine like a -11 with lots of three bit fields, thinking in octal was
> appropriate. For a machine like a VAX with lots of four bit fields, hex
> was what worked better. For hex instruction notation to seemingly be
> popular with x86 seems just perverse to me.
>
> Hex has additional smaller benefits: representing a 16 bit quantity in
> octal takes five digits and an extra bit, which isn't really a problem
> but is a bit annoying to the OCD. You can write a value in fewer
> characters with hex.
>
> The drawback to hex is you have to memorize twice as many binary bit
> patterns.

Yes, that's one drawback. If I'd used it at all often, I'm sure I would
have gotten them memorized.

You ALSO need to memorize extra addition tables; for octal, I just need
add two digits as if they were decimal, then convert the decimal result
to octal.

> But of course, the fact that you can fill unused memory with 0xDEADBEEF
> is enough reason to prefer hex all by itself!

Has it's points, yeah.

[Andrew Swallow]

I assume the hypotenuses was sqrt(5) = ~2.236

Andrew Swallow

[Andrew Swallow]

On 15/06/2012 22:39, Ibmekon wrote:
{snip}

> Reminds me of a pub question - that I puzzled over, and gave up on.
>
> Image a nice high wall.
> Image a box - a 1 meter cube.
> Image a ladder 10 meters long.
>
> Push the box against the wall
> Put the ladder against the wall, touching the box.
>
> Now give a calculation showing the exact height the ladder reaches on
> the wall.
>
> Carl Goldsworthy

You need to know the angle between the ladder and the floor to solve the
problem. A simple solution is a height of 0 - the ladder is lying along
the floor.

Andrew Swallow

[Joy Beeson]

We have three similar right triangles:

Using y for the height which is to be determined
x for the distance of the foot of the ladder from the wall
and a for the hypotenuse of the lower small triangle:

x^2 + y^2 = (10)^2

(x - 1)^2 + 1^2 = a^2

1^2 + (y - 1)^2 = (10 - a)^2

x/y = (x-1)/1 = 1/(y-1)

(10 - a)/a = (y - 1)/1 = 1/(x-1)

That's five equations with only three unknowns; it should be possible
to work something out, but it's been fifty years since I studied
quadratic equations and I'm getting a headache.

I think I'd start by determining a in terms of x, a in terms of y,
then y in terms of x. Um, no, x in terms of y, then you can plug that
in for x in the first equation and solve for y.
--
Joy Beeson
joy beeson at comcast dot net

[Ahem A Rivet's Shot]

Nope, that won't be touching the cube.

[Rod Speed]

Ahem A Rivet's Shot <ste...@eircom.net> wrote
> Andrew Swallow <am.sw...@btinternet.com> wrote
>> Ibmekon wrote

>>> Reminds me of a pub question - that I puzzled over, and gave up on.

>>> Image a nice high wall.
>>> Image a box - a 1 meter cube.
>>> Image a ladder 10 meters long.

>>> Push the box against the wall
>>> Put the ladder against the wall, touching the box.

>>> Now give a calculation showing the exact
>>> height the ladder reaches on the wall.

>> You need to know the angle between the ladder
>> and the floor to solve the problem. A simple solution
>> is a height of 0 - the ladder is lying along the floor.

> Nope, that won't be touching the cube.

The end of it can touch the cube, and the side of it the wall.

[Joe Pfeiffer]

Ahem A Rivet's Shot <ste...@eircom.net> writes:

> On Sat, 16 Jun 2012 03:09:51 +0100
> Andrew Swallow <am.sw...@btinternet.com> wrote:
>
>> On 15/06/2012 22:39, Ibmekon wrote:
>> {snip}
>> > Reminds me of a pub question - that I puzzled over, and gave up on.
>> >
>> > Image a nice high wall.
>> > Image a box - a 1 meter cube.
>> > Image a ladder 10 meters long.
>> >
>> > Push the box against the wall
>> > Put the ladder against the wall, touching the box.
>> >
>> > Now give a calculation showing the exact height the ladder reaches on
>> > the wall.
>> >
>> > Carl Goldsworthy
>>
>> You need to know the angle between the ladder and the floor to solve the
>> problem. A simple solution is a height of 0 - the ladder is lying along
>> the floor.
>
> Nope, that won't be touching the cube.

It'll be touching the cube, but not the wall.

[Rod Speed]

"Joe Pfeiffer" <pfei...@cs.nmsu.edu> wrote in message
news:1bmx439...@pfeifferfamily.net...

It will if the ladder is on the floor with the side touching the cube
and the end of the ladder touching the wall.

[Andrew Swallow]

On 16/06/2012 05:49, Ahem A Rivet's Shot wrote:
> On Sat, 16 Jun 2012 03:09:51 +0100
> Andrew Swallow<am.sw...@btinternet.com> wrote:
>
>> On 15/06/2012 22:39, Ibmekon wrote:
>> {snip}
>>> Reminds me of a pub question - that I puzzled over, and gave up on.
>>>
>>> Image a nice high wall.
>>> Image a box - a 1 meter cube.
>>> Image a ladder 10 meters long.
>>>
>>> Push the box against the wall
>>> Put the ladder against the wall, touching the box.
>>>
>>> Now give a calculation showing the exact height the ladder reaches on
>>> the wall.
>>>
>>> Carl Goldsworthy
>>
>> You need to know the angle between the ladder and the floor to solve the
>> problem. A simple solution is a height of 0 - the ladder is lying along
>> the floor.
>
> Nope, that won't be touching the cube.
>

Yes it does. The top of the ladder touches the bottom of the cube.

Andrew Swallow

[Andrew Swallow]

The top of the ladder touches the side of the cube, at the bottom.
The side of the ladder touches the wall.
Laying on the floor the latter touches the floor.

Andrew Swallow

[gareth]

"Ahem A Rivet's Shot" <ste...@eircom.net> wrote in message
news:20120615172540....@eircom.net...

> On Fri, 15 Jun 2012 11:50:02 -0400
> Dan Espen <des...@verizon.net> wrote:
>
>> In grade school we had a homework assignment to draw
>> a bunch of triangles with a compass based on their dimensions.
>>
>> One of the "triangles" had dimensions 1, 2, 3.
>

> I don't know, bloody degenerates.
>

Unless, of course, the blackboard was spherical in construction

[Peter Flass]

Ah, I was ignoring the "bit extra" - the heck with it, did a hole and
bury the box until its top is level with the ground ;-) So you have two
congruent right triangles. One side of each is a meter and the sum of
the hypotenuses is ten meters.
a**2 + 1**2 = c**
a'** + 1**2 = c'**
c**2 + c'**2 = 100
We need the sum of a and a'.
Now what?


--
Pete

[Ibmekon]

On Sat, 16 Jun 2012 08:22:43 -0400, Peter Flass
<Peter...@Yahoo.com> wrote:

<snippy>

>Ah, I was ignoring the "bit extra" - the heck with it, did a hole and
>bury the box until its top is level with the ground ;-) So you have two
>congruent right triangles. One side of each is a meter and the sum of
>the hypotenuses is ten meters.
> a**2 + 1**2 = c**
> a'** + 1**2 = c'**
> c**2 + c'**2 = 100
>We need the sum of a and a'.
>Now what?

If only I could remember who passed on this infection !
Hanging would be too good for them :-)


Carl Goldsworhyt

[jmfbahciv]

Ahem A Rivet's Shot wrote:
> On Fri, 15 Jun 2012 11:50:02 -0400
> Dan Espen <des...@verizon.net> wrote:
>
>> In grade school we had a homework assignment to draw
>> a bunch of triangles with a compass based on their dimensions.
>>
>> One of the "triangles" had dimensions 1, 2, 3.
>
> I don't know, bloody degenerates.
>

<GRIN> My thought exactly.


He was in grade school. He'd have had his mouth washed out
with soap for uttering a swear word.

/BAH

[jmfbahciv]

Put the ladder on the floor on its edge. Have top of ladder
touching the wall with the "top" and the outer edge of the cube
with its "bottom".

/BAH

[Waldek Hebisch]

AFAICS the expected answer is:

+--------------+ +------------+
| +--+ +--+ | +--+
(\|- 10\|23 + 52 + \|23 - 5)\|10\|23 + 52
+
+--------------+
+--+ | +--+
(\|23 + 5)\|- 10\|23 + 52 + 2
/
4

That is what computer gave me, I am too lazy to derive and solve
quartic by hand.

--
Waldek Hebisch
heb...@math.uni.wroc.pl

[Charlie Gibbs]

In article <k1ant79gj9hfipfo5...@4ax.com>, Ibmekon

(Ibmekon) writes:

> Reminds me of a pub question - that I puzzled over, and gave up on.
>
> Image a nice high wall.
> Image a box - a 1 meter cube.
> Image a ladder 10 meters long.
>
> Push the box against the wall
> Put the ladder against the wall, touching the box.
>
> Now give a calculation showing the exact height the ladder reaches
> on the wall.

10 + 1 = 11 meters - if you set the ladder on top of the box,
flat against the wall.

You didn't say the ladder had to touch the ground...

--
/~\ cgi...@kltpzyxm.invalid (Charlie Gibbs)
\ / I'm really at ac.dekanfrus if you read it the right way.
X Top-posted messages will probably be ignored. See RFC1855.
/ \ HTML will DEFINITELY be ignored. Join the ASCII ribbon campaign!

[Charlie Gibbs]

In article <ylfkaa04...@dd-b.net>, dd...@dd-b.net

(David Dyer-Bennet) writes:

> I never did work with a computer so primitive I had to enter large
> amounts of code from the front panel. I avoided the nasty new-fangled
> micro-computers for years because they were so immensely inferior to
> what I was used to at work; they weren't tempting at all.

While conceding your point on the one hand, on the other hand it was
fun to have a machine I could play with at home. When the IBM PC
came out, I wasn't in the least tempted - early models had nothing
that my IMSAI didn't. And the IMS 5000 boxes I used in a work
application, with their hard drives and 6-MHz Z80Bs, blew the
doors off an XT - just as the TurboDOS they ran was so much nicer
than MS-DOS. It took the advent of the Amiga to provide a home
machine that excited me enough to make a switch.

[Ibmekon]

On 16 Jun 12 10:53:00 -0800, "Charlie Gibbs" <cgi...@kltpzyxm.invalid>
wrote:

>In article <k1ant79gj9hfipfo5...@4ax.com>, Ibmekon
>(Ibmekon) writes:
>
>> Reminds me of a pub question - that I puzzled over, and gave up on.
>>
>> Image a nice high wall.
>> Image a box - a 1 meter cube.
>> Image a ladder 10 meters long.
>>
>> Push the box against the wall
>> Put the ladder against the wall, touching the box.
>>
>> Now give a calculation showing the exact height the ladder reaches
>> on the wall.
>
>10 + 1 = 11 meters - if you set the ladder on top of the box,
>flat against the wall.
>
>You didn't say the ladder had to touch the ground...

Well it was not given to me as a riddle.

Common sense says enough information is given, once the ladder is
pushed up against the box and wall - so it is just a matter of
developing a formula.
But the triangles seem to drive you in circles, so to speak.

For some reason, things I cannot understand, but think I should, stick
around in my head.
The rest goes over my head !

Carl Goldsworthy