A modest technological proposal in praise of imperialism

[andrewducker]

Having seen the news this morning concerning the failure to kill off imperial measurements I hereby propose an Imperial measurement for data.

No longer will we be under the sway of these new-fangled kilo, mega, giga and terabytes. Instead we shall refer to our technological achievements by the old-fashioned imperial measure - the inch, foot and yard. How will we measure data in yards, you ask? Simple - we will return to the ancient days of computing and revive the first portable standard for data transport - the punch card!

By stacking an imperial standard punch card of data on top of its noble brethren one will construct a stack. Said stack's height can then be measured, producing a result both memorable and patriotic. No longer will a coder be forced to report their output in foreign units - they will be able to report exactly how many feet of code they have produced! I feel
sure that within the month computer users acros the nation will be proudly reporting how many miles their hard drives can contain, yards of bandwidth they have used and leagues of porn they have downloaded.

(with thanks to octopoid_horror, who said we needed a new measurement to replace the kilobyte)

Comments

[chillies.livejournal.com]

yeah ... and one punch card == 72 columns of data in my FORTRAN world, so get practising those 72 time tables.

[octopoid-horror.livejournal.com]

HERETIC!

I outed you earlier, by thunder! You're a "stacker" when the only RIGHT measurement is to lay said cards end-to-end.

End-to-Enders shall triumph over the nonsensical "stackers"!

DEATH TO STACKERS!

Also, my hard drive holds three hands and two rills of pr0n. Woo.

[dapperscavenger.livejournal.com]

I, for one, am glad. They really needn't have been so heavy handed about the change-over to begin with. Anyone with half a brain can see that by teaching kids metric we can phase out imperial gradually. Though, I still buy my sweets in quarters and my mince in pounds. The portions seem more sensible.

[andrewducker]

A pound of mince is 454g. So 500g or half a kilo is close enough.
And a quarter is 113g. So 100g is probably close enough (give or take a couple of murray mints).

Should you ever need to know :->

[octopoid-horror.livejournal.com]

There was a funny Somethingawful article a while back about units of descriptive measurement ie football pitches and Everests....

I think I mentioned about the stupid ned on the deli counter in Morrisons who, when I asked for 200 grams of something, put the requisite item on the scales and when it said 0.200kg, she had to go ask her supervisor how to find out how many grams it was...

[flickgc.livejournal.com]

I once asked in a shop for half a pound of something or other. After a slightly funny look, I was handed a very very small bag and asked for 50p....

[surliminal.livejournal.com]

Pints are the problem : litres and half litres don't map. AS is in fact mentioned at length by a prole in 1984...

[dapperscavenger.livejournal.com]

I'm not asking for 100g of sweeties! Thats LESS than what I'd usually get! I'd die without that extra 13 grams, I'm certain. :p

[cybik.livejournal.com]

100g of sweets is LESS than a quarter; THEREFORE IT IS NOT ENOUGH SWEETIES!

I want my sweeties, you meanie :(

[0olong.livejournal.com]

Thing is, we've been teaching kids metric measurements for quite a few decades already. Maybe having that combined with dual measurements printed on everything will eventually make the difference, but I dunno...

[dapperscavenger.livejournal.com]

I've found that most kids younger than me tend to ask for things in metric and don't have a clue what a yard is, so I'm basing my assumption of a gradual phasing on that.

[surliminal.livejournal.com]

You mean the punch-yard?

[0olong.livejournal.com]

I had a tutor once who frighteningly talked about the number of miles of code in projects he'd been involved in...

[the-magician.livejournal.com]

Well we measure code in KLOCs, but without a sensible length of a line, how do we know how big that is?

At least with 79 column punch cards, we know how long a line is ... and if I could figure out the weight of the card in a punchcard I could come up with an estimate of thickness, and then figure out how thick 240 punchcards would be (which of course would be broken down into 20 sets of 12 for ease of calculation!) and then we could figure out how many cards in 14 inches (a light-femto-fortnight) and we'd be all set ...

[the-magician.livejournal.com]

The IBM card format, designed in 1928, had rectangular holes, 80 columns with 12 punch locations each, one character to each column. Card size was exactly 7-3/8 inch by 3-1/4 inch (187.325 by 82.55 mm. The cards were made of smooth stock, 0.007 inch (0.178 mm) thick. There are about 143 cards to the inch.
So 143 cards in a block 7-3/8 x 3-1/4 x 1 (142.86 actually if the cards are *exactly* 0.007inches thick)

That makes 10299.12 cards per cubic foot (within 1% of 10K) or 278076.4 cards per cubic yard. At 80 characters per card that makes a cubic yard of data 22,246,109 bytes.

So one gigabyte = 48.266 cubic yards of data. Which is a cube roughly 3.6yards along each side (just under 11 feet).

For the end-to-endians ... one byte is 0.092188 inches, so one gigabyte (1024^3) is 1562.272 miles. Which makes the Earth just over five gigabytes in circumference (5.0736 to be reasonably precise).

Bytes per light femto fortnight ...

[the-magician.livejournal.com]

since a card is 0.007 inches thick, there are almost precisely 2000 cards per light femto-fortnight! Or 160,000 bytes per light femtofortnight (based on card thickness)

For the end-to-endians, one card is pretty close to being half a light femtofortnight (or as some would call it, a light femto-week!) so 160 bytes per light femto-fortnight ... so, rule of thumb, translating between card depth and end-to-endian is roughly a 1000:1 translation.

[whumpdotcom.livejournal.com]

Well, speaking for my Second Life character, who lives in a steampunk sim: "Capital idea, Mr. Ducker. Here, here!"

[azalemeth.livejournal.com]

If we assume that a 'computer punchcard' holds 128 punches, multiplied by six bits a punch (http://www.cs.uiowa.edu/~jones/cards/history.html) (near the bottom of the page) for a total of 768 bits/card, or 96 bytes a card, and each card is 0.007 inches thick, (and 2 5/8ths in high by 3 1/4 in wide), then 1 gigabyte is approximately 1 099 511 627 776/96, or 1.14532461 × 10¹⁰ punchcards.

This corresponds to a stack of paper 6 681 060.24 feet - 1 265.35232 miles - high, all in order, and all about 6.67cm by 8.25cm.

Now, being a naturally red-blooded eighteen-year-old male, I can say hence that my vice collection stretches about 250, 000 miles, or, nearasdamnit, the best part of ten circumference-of-the-earths.

[azalemeth.livejournal.com]

Incidentally, this is roughly one moonlength, 4.25 x10^-8 lightyears, 1.3 x10^-8 parsecs, 100 584 000 000 termite lengths (african soldier), or, in much more familiar terms, 72 414.6868 leagues.

Any questions?

[call-waiting.livejournal.com]

The metric system has been taught to the exclusion of the Imperial system for ages now. I was always taught in metric, and until I was about 17 I wouldn't have known a 'foot' if it had hit me in the face.

It's only since then that, in the process of getting along in the "real" world and interacting with the people that inhabit it, I've learned, mostly through osmosis, the imperial measures, and in many circumstances have come to use them by 'default'. If I'm measuring something, I'll measure it in metres. But if I'm reckoning something, I'll reckon it in feet. Don't ask me how this is, but the Imperial measures just make some sort of intuitive 'sense'.

And so it is with kilobytes, megabytes, etc. Because these are the Imperial units of data, don't let those metric prefixes fool you. If a kilobyte was metric, it would be 24 bytes shorter. In fact the metric equivalent, is precisely 1000 bytes. Technically we're supposed to call our 1024-byte units 'kibibytes', because supposedly 'kilobyte' really is the metric name for it. But nobody does this, because it's intuitively wrong (what's 1000 in hex? I have no idea off the top of my head, but I can tell you that 1024 is 0x400 without blinking, and also just sounds fucking stupid).

And let's not forget the somewhat arbitrary selection of 8 as the base multiple of the real atomic unit of data.

Yep, my computer gets 32 bits to the word, and that's the way I likes it.

[andrewducker]

8 seems like a reasonable base - fairly arbitrary, but not entirely. You'd want a number that was itself a power of 2, for referencing purposes. 4 is definitely too small for much to be useful. 16 was probably beyond what was feasible at the time that computers first became popular.

[kelly-lesbo.livejournal.com]

This post reminds me that I'm really just a nerd pretender..I shall retreat to the comfort of my videogames and sci-fi tv shows