[SIGCIS-Members] Candidate World Records -- Biggest Hard Drive, Tape and Portable Computer

[Ken Strauss]

See http://ibm-1401.info/CrownPaperIBM/B09V01S0201.pdf if you want to know the 
weight of IBM cards and http://ibm-1401.info/CardStockSpecifications.html for 
links to further data on the cards.

> -----Original Message-----
> From: Members [mailto:members-bounces at lists.sigcis.org] On Behalf Of Allan
> Olley
> Sent: Monday, September 21, 2020 2:13 PM
> To: thomas.haigh at gmail.com
> Cc: [SIGCIS]
> Subject: Re: [SIGCIS-Members] Candidate World Records -- Biggest Hard Drive,
> Tape and Portable Computer
>
> Hello,
>
> So just to attempt to clarify how big the SSEC tape drive was. I've 
> discussed this
> with Tom and admitted much confusion looked about started to see answers
> and then never nailed anything down, so let me just do this now.
>
> Anyway I am very sure the actual reels of paper tape used on the SSEC were
> much smaller than the 400 pounds of tape stock delivered fromt the factory. 
> A
> 400 pound paper tape would have stored I am guessing about 25 times as
> many numbers on the tape then the contemporary accounts refer to as stored
> by the machine in total. The standard size of the tapes was probably more 
> like
> 15 pounds of paper for all tapes on the machine and so slightly less than 5
> pounds for each large tape on the main readers. They were certainly still 
> very
> bulky and just looking at pictures of them I would guess they weighed
> somehting like 20 pounds, so I am still confused.
> Anyway my logic and evidence for these statements goes as follows (much
> unnecessary detail follows):
>
> So each number on the tape was coded in a row (78 potential punches) in
> binary-coded decimal (4 bits to the digit) 19 digits plus sign (so some
> descriptions of the SSEC describe it as using numbers of 19 digits, others 
> as 20).
>
> For comparison a standard IBM punch card was 80 columns, 10 rows of digits
> 0-9 and two invisible rows (A, B or X, Y).
>
> Contemporary accounts (Wallace Eckert, "Electrons and Computation" in The
> Scientific Monthly, Vol 67, No 5 (Nov) 1948 pp. 315-323) say the three main
> tape drives of the SSEC store in total 20 000 numbers or 6 666 numbers on
> each tape drive. 12 numbers are 12 rows and so constitue ~1 punch card, so
> each main tape drive has the equivalent paper of
> ~556 punch cards or ~1666 cards worth of paper in total. An IBM carrynig 
> case
> for 2000 cards apparently weighed 6.6 kg (14.52 lbs) according to this 
> webpage
> https://nuhc.ncl.ac.uk/moca/iomedia/pc.htm and that sounds about right for
> the weight of 2 000 cards but I wish I had a more definite benchmark to use 
> for
> the rest of my arithmetic then this one number pulled from a random webpage.
> Anyway, give that I don't think the tape on all three tape drives weighed 
> more
> than 15 pounds, the tape on each one would have weighed a little less than 5
> pounds.
>
> As mentioned there are 36 tape readers that are associated with 5 000 more
> stored numbers on tape. So each loop contained about 139 numbers (I've
> found archival sources that suggest they sometimes used loops of over 500
> numbers in these configurations but I guess that was non-standard?) tha is
> again rows of tape, so was equivalent to a little more than 10 punched 
> cards.
> So all 36 loops add less than the equivalent of 417 cards, so less than 
> another 5
> pounds, to the weight of the tapes on the SSEC.
>
> So the 25 000 numbers stored (as standard) on all the SSEC's paper tape is 
> just
> a tad more (1/24th more) paper than 2 000 punched cards in the carrying case 
> I
> am using as my weight unit (which would store 24 000 numbers by that
> method). Suggesting the paper tape of the machine amounted to about 7 kg of
> paper or about 15 pounds in total. I am not sure you really needed a chain 
> hoist
> for that, but depending on what you are doing with the original 400 poounds 
> of
> raw paper tape received from the factory to prepare it for use, the chain 
> hoist
> could have been useful.
>
> On some subsidiary issues:
>
> If you want to see what the short loops of paper tape looked like there is
> footage of the SSEC available on the Comptuer History Museum's website
> Youtube channel (see 49:50 of https://youtu.be/qundvme1Tik note this portion
> of the video is a lecture given by Herb Grosch). You can see a good image of 
> the
> main paper tape drives at the back of the SSEC in a famous of image of the 
> SSEC
> where the pesky support beams for the room it is in have been removed that
> shows up around 50:25 into the video. They are pretty big I don't think big
> enough to way 400 pounds (or even 133 pounds each) but big enough that I
> think they would weigh more than 5 pounds (they look like more than 500
> punched cards worth of paper), I would guess 15 or 20 pounds (so each tape 
> is
> more than 2000 cards worth of paper).
>
> I am pretty sure the reels on the main tape readers are glued together at 
> the
> end into a continous loop because John Backus describes an incident where 
> the
> SSEC started to act strangely and they realized they had glued the ends of 
> the
> tape together the wrong way around creating a giant mobius strip of tape 
> (and
> so reversing the numeric encoding of instructions and data at various 
> points). I
> can't imagine you would not notice if the short loops of tape were a mobius
> strip basically as you hung them up, so it must of have been one of the 
> three
> master tapes and so they must operate as continuous loops at least some of
> the time.
>
> (Backus's memory of this comes from page 126 of The History of Computing in
> the Twentieth Century, ed. Nicholas Metropolis, J. Howlett and Gian-Carlo
> Rota, 1980 Academic Press,
> https://books.google.ca/books?id=AsvSBQAAQBAJ&lpg=PA126&ots=Xgj6NX9t4
> _&dq=Mobius%20loop%20SSEC%20tape%20Backus&pg=PA126#v=onepage&q
> =Mobius%20loop%20SSEC%20tape%20Backus&f=false
> but I also have run across descriptions of this process, but not the Mobius 
> strip
> incident, in unpublished material)
>
> --
> Yours Truly,
> Allan Olley, PhD
>
> http://individual.utoronto.ca/fofound/
>
> On Sun, 20 Sep 2020, thomas.haigh at gmail.com wrote:
>
> >
> > Hello SIGCIS,
> >
> >
> >
> > I have agreed to a fun little consulting job of helping the Guinness
> > World Records people update and improve their computer-related
> > records. Although the Encyclopedia Britannica and other traditional
> > authorities have fallen by the wayside, the records book is still
> > going strong and is perhaps the last published authority standing.
> > Also, somebody once tried to prove me wrong by invoking it. So now I
> > can make sure that doesn't happen again.
> >
> >
> >
> > I'll be dribbling out a couple of additional requests over the next
> > week or so, not to overwhelm the list. I have some fun candidate
> > records, but would like to see if they hold up among this expert body.
> > BTW, If you have an idea for a good record feel free to pitch it to
> > me, but to avoid overwhelming the list better to send it directly. We
> > are trying to avoid firsts, which have traditionally accounted for the
> > majority of the computer-related records.
> >
> >
> >
> > BIGGEST HARD DISK: First up, the biggest hard disk. Not the largest
> > capacity, the BIGGEST. I figure records like that will illustrate
> > better than finding some boring box with a high capacity, and are less
> > likely to be out of date by the time the book is published. Plus where
> > do you draw the line between a drive and an array?
> >
> >
> >
> > Everyone knows RAMAC, which apparently had 24 inch platters. Platters
> > generally shrank over the years, as everything else being equal
> > smaller platters can be spun faster, so I believe later mainstream IBM
> > systems were smaller. But third parties offered higher performance,
> > higher capacity drives. The best candidate appears to be the Bryant
> > Model 2 Disk File from the early 1960s. That had 39 inch platters. The
> > image below is from a brochure online at
> > CHM:https://archive.computerhistory.org/resources/text/Bryant/Bryant.M
> > odel2.19
> > 65.102646212.pdf.
> >
> >
> >
> > [IMAGE]
> >
> >
> >
> > According to Wikipedia, "Also in 1961, Bryant Computer Products
> > introduced its 4000 series disk drives. These massive units stood 52
> > inches (1.3 m) tall, 70 inches (1.8 m) long, and 70 inches (1.8 m)
> > wide, and had up to 26 platters, each 39 inches (0.99 m) in diameter,
> > rotating at up to 1,200 rpm." This Computer History Museum page seems
> > to be hedging bets by calling the Bryant units "among the physically
> > largest drives everbuilt":
> > https://www.computerhistory.org/storageengine/hdd-competitors-ente
> > r-the-market/
> >
> >
> >
> > I'm fairly sure that the Bryant platters were the largest in a
> > standard commercial product. Can anyone prove differently? Or know of
> > a special-purpose system with even larger platters?
> >
> >
> >
> > BONUS - HEAVIEST TAPE REEL: I am pretty sure that the IBM SSEC had the
> > heaviest tape reel, at 400 pounds, which was a roll of the paper stock
> > cards were cut from. I believe this was run in an endless loop to used
> > as a high speed lookup table. It had to be lifted with a special
> > mechanism. But if anyone knows of a bulkier tape, let me know. The
> > SSEC tape is documented at
> > http://www.columbia.edu/cu/computinghistory/ssec-tape.html. That
> includes this description from Herb Grosch:
> >
> >
> >
> > "About those tapes: the card plant in Endicott got enormous rolls of
> > card stock from the paper mills. For regular card manufacturing they
> > slit the rolls to three-inch width (card height). For the SSEC they
> > furnished rolls eight inches wide (card length). The resulting rolls
> > weighed 400 pounds, and had to be hoisted onto the SSEC with a
> > thoughtfully-provided chain fall! For the Stallion, we pushed the rolls up 
> > a
> ramp.
> >
> > "The punch stations, slightly modified from standard IBM reproducer
> > components, punched two round sprocket holes at the edges, and 78(!)
> > regular IBM rectangular holes in between. The sprockets drove the tape
> > one line at a time, and drives under separate program control fed the
> > fresh or pre-punched tape under ten 78-brush reading stations. The
> > tapes hanging down could lengthen and shorten, and for program tapes
> > and the table lookup unit we cemented the tape end-to-end into short
> > loops (yes, someone had had to provide the jig). There were three of
> > these monsters at the end of the machine room. Up to 36 of the
> > fixed-length tape loops could be mounted on the separate table-lookup
> > unit, which in later years was also sometimes used for program
> > reading. For the lunar calculation, I used 24 loops to make lookup
> > time as short as possible, and we got programming from the main tape
> readers."
> >
> >
> >
> > BONUS - LARGEST PORTABLE COMPUTER
> >
> >
> >
> > That appears to be DYSEAC, which makes sense as AFAIK it was the only
> > portable full-scale vacuum tube computer. It apparently weighed 20
> > tons and took two 40 foot trailers to move. (Weight includes the
> > trailers, but not the tractors to pull them). I asked Evan Koblentz,
> > who wrote a book about portable computers, and he couldn't think of a
> > heavier/larger one. This makes even the Osbourne and IBM PC Portable
> > look extremely light. Read about the DYSEAC here:
> > http://ed-thelen.org/comp-hist/SEAC&DYSEAC-3-150.pdf, including this
> image:
> >
> >
> >
> > [IMAGE]
> >
> >
> >
> >
> >
> > It would be fun if official recognition of these records prompted some
> > enterprising teams to attempt to beat them.
> >
> >
> >
> > Best wishes,
> >
> >
> >
> > Tom
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >