Rare Apple I computer sells for $216,000 in London

[Joe Pfeiffer]

Doesn't sound like line printer output, that was usually 135 column
sprocketed paper (a few were 80 columns)

Doesn't sound like US line printer output. I'd sort of expect line
printers in europe to use A4 in exactly the same circumstances where the
US would use 8 1/2 x 11 (though I don't have any experience with
european computer installations).

 

[Ahem A Rivet's Shot]

Doesn't sound like US line printer output. I'd sort of expect line
printers in europe to use A4 in exactly the same circumstances where the
US would use 8 1/2 x 11 (though I don't have any experience with
european computer installations).

All the line printers I saw used sprocketed paper that was either
80 or 132 columns (10 cpi) and usually 66 lines (6 lpi). A4 was usually
used in daisy wheel printers, although I do recall paper for "NLQ" dot
matrix and early inkjet printers (built like dot matrix printers with an
inkjet head instead of hammers and ribbon) that was sprocketed with "micro
perforations" such that when you (carefully) stripped the sprocket holes
and separated the sheets the result was A4 sized with smooth (ish) edges -
except where it tore while pulling the sprocket strips off.

 

[Charlie Gibbs]

What did you do, replace their sequential search of the symbol
table with a hash or an AVL tree???

I've always been a big fan of insertion sorts for such things -
then I can use a binary search. But I don't think that was the
main reason my assembler was so much faster. Univac's assembler
was heavily I/O-bound; it must have been doing a lot of passes,
or was keeping too much stuff on disk. The binary had 40 overlays
or so. Mine consisted of 4 phases, with one overlay each for the
second, third and fourth phases. The first phase read the source
file and expanded macros, the second one read the expanded code and
built the symbol table, the third read the expanded code again and
generated the listing, object code, and cross-reference data, and
the fourth sorted and printed the cross-reference data.

If I was assembling a program from a card deck, the first pass
worked directly from the card input without reading it into a
disk file first. You could hear the card reader stutter or pause
at each macro. The second phase ran in about the time it took the
card reader motor to time out and shut down, at which point the
printer fired up and started cranking out the listing.

This was on the Univac 9300, although I used the same principles
when I wrote my OS/3 assembler. Rumour has it that Univac's OS/3
assembler was a hacked version of IBM's DOS/360 assembler, whose
source code somebody found in the trunk of a car. Another rumour
states that IBM wanted it found.

 

[Charlie Gibbs]

Curious (20 min vs 40 min). Could make a guess of how many minutes
were the threshold? 30 mins?

I don't think there was a threshold so much as programmers had the
lowest priority for machine time, and the quicker I could get in and
out, the more likely I could wheedle access.

 

[Charlie Gibbs]

The program is already developed, and the output is pretty good, so
far. It's just that it is S L O W.

Slow is no problem - get a scanner with a sheet feeder, fire it up,
and go to bed. It's the preuf reeding that's the killer.

 

[Charlie Gibbs]

That's very odd. My shop had a dozen 26s and one or two 29s. It was
very rare to have one with a down IBM card on it. Those keypunches
were used by students who had minimal training about the care and
feeding of keypunches.

In our student area, there were usually one or two punches down
due to jams. For me, this was a benefit, for I knew how to
clear them. In a student area with long lineups for a punch,
it was as if one had been reserved for me; I'd walk past the
lineups, sit down at the jammed punch, clear it, punch my cards,
and leave the punch available for someone else.

The most common mainentance problem was when a student forgot to flip
the switch which lifted the wheels off the drumcard. But that was fixed
in a couple of minutes.

If you cared to dig down inside the machine and retrieve the missing
star wheels, and knew how to re-attach them...

 

[Rod Speed]

keithr wrote

Rod Speed wrote

Hmm I don't remember anybody called rod speed at DCR,

Your problem.

so how would you know?

I did anyway.

I was at ABS from '73 to '76 and from '80 to '86. ANU, in my
experience were pretty anti-IBM, the academics preferred Univac.

Have fun explaining the hardware they had in the late 60s.

 

[Rod Speed]

keithr wrote

Rod Speed wrote

That shows you more as an old fart than a geek.

Nope, just shows they are convenient to write on.

 

[Rod Speed]

jmfbahciv wrote

Rod Speed wrote

That's very odd. My shop had a dozen 26s and one or two 29s.

I was talking about just 029s.

It was very rare to have one with a down IBM card on it. Those
keypunches were used by students who had minimal training about
the care and feeding of keypunches. The most common mainentance
problem was when a student forgot to flip the switch which lifted the
wheels off the drumcard. But that was fixed in a couple of minutes.

How many did you have?

Rather more than you were tallking about.

 

[[email protected]]

Slow is no problem - get a scanner with a sheet feeder, fire it up,
and go to bed. It's the preuf reeding that's the killer.

The program isn't that smart. It uses TIF files for input. It has
a numer of other drawbacks one of which is that it does a poor job
of deskewing the image.

Besides, scanners that can read B sized paper that have feeders
aren't all that cheap.

 

[Rod Speed]

jmfbahciv wrote

Rod Speed wrote

Sheesh! Can't you answer a question with a fact?

Everyone can see you are lying, as always.

 

[Rod Speed]

Sure. Neither of the above two breakages required a field service call...usually.

Irrelevant to whether they had to be on maintenance contract anyway.

 

[Rod Speed]

You seem to be having a conception problem.

You clearly need to retake Bullshitting 101.

A simple, "I can't remember" answer would have sufficed.

It would not have been accurate, fool.

<reams of your puerile shit any 2 year old could leave for dead flushed where it belongs>

 

[Rod Speed]

jmfbahciv wrote

Rod Speed wrote

They didn't have to be on a service contract

Depends on the capabilitys of the individuals involved.

but it was wise to have a contract early on.

It's not irrelevant when you have a dozen keypunches which are used
12-14 hours/day. One down keypunch can create student unrest.

What I said in more words.

 

[Charlie Gibbs]

Wheedling worked?

Sometimes. At other times a more direct approach was more effective:
"Do you want this fix or not?"

 

[Frank Slootweg]

Frank Slootweg wrote:
Frank Slootweg wrote:
Frank Slootweg wrote: [...]
Scanning it with OCR sofware is much easier and has
remarkably good/reliable results.

Sometimes. Think about the listings generated by ribbon ink.

I have had some remarkable results with my lowly EPSON
(Perfection V30) scanner. Also on very old policies which were
printed by a (ribbon ink) line-printer,

Do you know which one? How did you eliminate the lines?

No, I don't know which line printer, I only have the printouts.

OK. Some were pretty good at printing and others were awful. I was
curious. [...]
FWIW, the policies were printed on - somewhat translucent - 'white'
paper.

How wide was the paper? Or was it TTY paper?

It was A4 format, i.e. about the US 8.5x11" format.

Doesn't sound like line printer output, that was usually 135 column
sprocketed paper (a few were 80 columns)

It was probably the other way around. We computer geeks needed 13X
column width for our silly source code, etc., but the real/business
world was happily printing documents for their customers, and that was
US-letter/A4.

Anyway, all the 13X column line printers I remember, could also do 80
column by just moving the right-hand sprocket.

80 column is 8.5 inch wide US letter format. A4 is a little bit less
wide, so no problem. For US-only line/sprocket-feed printers, where the
right-hand sprocket could not be moved far enough left to accomodate
normal A4 (with tear-offs) paper, there was paper where the left or
right tear-off was a little wider, resulting in A4 width between
tear-offs.

To those trying to remember names of printers, IIRC one popular brand
was Data Products. IIRC my (ex) employer - HP - sold rebadged Data
Products printers.

 

[Joe Thompson]

Anyway, all the 13X column line printers I remember, could also do 80
column by just moving the right-hand sprocket.

I had a Seikosha 9-pin dot-matrix printer (ca. 1990) that worked this
way. If your fanfold was 8.5 inches wide, you would set it left; if it
was 11 inches wide, set it right. You could set it anywhere in between
but there would be no reason to.

After a few false starts I developed the habit of rough-aligning by
hand, threading the paper onto the sprocket, then using the holes
themselves as a guide for proper alignment.

I remember one of the ads for that printer in _Computer Shopper_ showed
a sheet of aluminum with the logo printed on it, allegedly by the bare
pins of the printer in question (the idea was to show that the printer
was tough). -- Joe

 

[Charles Richmond]

[snip...] [snip...] [snip...]

To those trying to remember names of printers, IIRC one popular brand
was Data Products. IIRC my (ex) employer - HP - sold rebadged Data
Products printers.

I guess Data Products was a common third-party printer that was
"re-badged" by different companies. At a PPoE, I worked in a lab
with a Harris 800 (later a Harris 1200 with ECL) that had a Data
Products line printer re-badged as a Harris. I do *not* remember
us ever having any trouble with our Data Products printer.

You *could* buy the same printer from Data Products directly for
*less* money, but then you could *not* get Harris to cover the
thing on their service agreement. So you were constrained to buy
the higher priced re-badged model.

Later, Harris started re-badging those cheap Wyse terminals. One
of the labs at my PPoE just bought *twice* as many of the cheapest
Wyse terminals directly from Wyse. (I think they were about $50
each.) Harris would *not* cover these under their service
agreement, but if one broke... the lab just threw it away and got
another one out of storage. Cheaper to replace than pay the
service agreement for the re-badged version.

--
+----------------------------------------+
| Charles and Francis Richmond |
| |
| plano dot net at aquaporin4 dot com |
+----------------------------------------+

 

[Charles Richmond]

Sometimes. At other times a more direct approach was more effective:
"Do you want this fix or not?"

Yeah... You could say: "Okay, I don't really need to put in this
software fix. Only 40% of the paychecks will come out wrong, and
we can fix all that later..."

--
+----------------------------------------+
| Charles and Francis Richmond |
| |
| plano dot net at aquaporin4 dot com |
+----------------------------------------+

 

[Joe Pfeiffer]

So you are one of those people who believe that withholding data
and information gives you more power.

where it belongs>

Not cartons of Kleenix but railroad cars full.

Barb, really -- he just isn't worth the effort.