Maybe fun, a blog post of sorts…
Oh yikes - glad to say I’ve read this one before, and feel no great need to read it this way! It does perhaps concentrate the mind on the text as written. I think we should call this art, and feel free to spend as long (or as little) contemplating it as we like… Bravo.
Yes, it’s certainly slightly off-topic and a bit artsy, as well. However, so many punch cards can’t be wrong!
Are there any stories about reading lengthy texts and/or sources directly from punch cards, e.g., because no printer had been available?
For the more practical side, are there any metrics for a stack of 10,700 cards in terms of weight and height? The only thing I found was this table for calculating card stock by IBM http://ibm-1401.info/CrownPaperIBM/TabCardStockManSpec.pdf
(A Quora answer works out 12.5 billion cards to hold a terrabyte, which translates to “1,205,467 cubic feet of space and weigh 34,097.5 tons” – no idea, how accurate this is. But I’m rather sure there would have been some formula ready to estimate storage requirements of data on card stock, back in the day.)
A stack of this size would take an IBM 802 about 10 3/4 minutes to sort it (this was rather fast at 1000 cards/minute), but it would take an IBM 548 Alphabetic Interpreter about a week to label the cards. (This one labels two rows of 60 columns for 60 cards a minute, so it’s a bit unclear how it would perform, if it could actually work on 80 columns. You probably wanted to print the labels directly on the card punch.)
A bit of calculation could tell us what we need to know, then!
Here’s a photo of 62,500 punched cards, being the SAGE software, 5MByte apparently:
While messing about with my own card project, I came up with this, based on the IBM stock calculator, some old shipping weight invoices for card boxes, and physically scaling from stock photos of cards on palettes, plus a whole bunch of hand-waving:
- A single box of 2000 punched cards was about 370 × 200 × 95 mm and weighed roughly 5 kg
So your “novel application” would be a stack of six boxes totalling 370 × 200 × 570 mm and weighing about 29 kg. The top one wouldn’t be completely full. I didn’t count the loss of mass from the chad, but you can — each hole loses a mass of 0.000715 g, working from 161.1 gsm (“99 pound basis weight”, a maddening unit) and my always-to-hand copy of Federal Information Processing Standards Publication 13: rectangular holes in twelve-row punched cards.
A friend has recently confided that he has 2–3 unopened boxes of punched cards, but I suspect he won’t let me near them as I might steal them. But it would be theft for science …
This reminded me of this story told by Butler Lampson, recipient of the Turing Award for his contributions to computer science:
So it has negative weight? What does the software revolution mean for planetary revolution (and, esp the Moon?) Is there danger?
On the more serious side, if we ignore the border of a card for a moment, a possible punch position occupies a about a fourth of that particular card area. Provided that we use this to encode text, there are zone punches, meaning there are about two punches per code in average. Meaning, the chad is about 2/48 of the punchable area, or a bit more than 4%. Adding the vertical edge area to this, it’s 2/52 or a bit less than 4%. The horizontal edges of a card are about 4 columns worth of material, so let’s subtract 4/84 from this, which gives about 3.75% of the total weight being chad. For 29 kg, this is about 1 kg (even a bit more) – which is not insignificant.
Jane Austen is wanted for 1 kg of worthy card stock!
(What’s now about “no material damage was done”, the ending of about every Austen novel?)
I can remember that around 1980 we drove to a customer with our latest product, about 12,000 lines PL/I, in the form of punched cards in the trunk. The stern was almost hanging on the road.