The Bayeux Tapestry archiving model

In the town of Bayeux in northern France you can see the world's oldest information archive based on a long ribbon of material, a very early example of what was to become tape media. The 1000-year-old Bayeux Tapestry, actually an embroidery on a fine linen background, depicts a sequence of scenes telling the story of the …

[Anonymous Coward]

bit slow...

"2.168 bytes/per hour (17.344 bits/min)", surely that is kilobytes per hour.

[Anonymous Coward]

This article is slightly silly and a little bit pointless

and I enjoyed it very much indeed - thank you.

[Anonymous Coward]

Do the math?

51,678.72 square inches at 47dpi (which is 2,209 dots per SQUARE inch) with 3 bits per dot (because there are 8 colours) is 342,474,877 bits, or approximately 40.826MB. However, this figure fails to account for the fact that the Bayeux Tapestry is not a tapestry but an embroidery. So it's really only the embroidered area which counts, not the linen ground, which will reduce the information content by maybe 50%. Even so, that still comes out rather higher than the author's figure of 2.429MB.

[Matthew Malthouse]

Depends...

You only get a saving on the "empty" areas if you use compression.

[system]

Colour

There are 8 colours used in the tapestry, making it 3 bit colour. For 8 bit colour depth, you'd need 256 colours.

So, the capacity calculation becomes (47 * 51,678.72 * 3) / 8 = 910.84 KB. Or 7.29 megabit.

[Anonymous Coward]

In my experience...

Any sales department would love an archive like this - telling just one side of the story and with a wonderfully revisionist version of events.

[Matt Bucknall]

Um..

Maybe I'm being really thick (which wouldn't surprise me given how ill I am at the moment), but I'm not sure your figures are right.

Your calculation appears to be: 51678.72 * 47 / 1000000 = 2.429MB

So this is in fact assuming 1 byte per colour, not one bit - Which makes more sense given you previously stated a colour depth of 8. Although it's not clear if that is a depth of 8 bits or a depth of 8 colours.

That aside, shouldn't the calculation be: 51678.72 * 47 * 47 / 1000000 = 114.16MB?

The resolution is 47 dots per inch and therefore 2209 dots per square inch.

[bitten]

Article length 4.8K

"10.84 bytes/hour; a spectacularly slow data rate in information transfer terms."

I hope this article was written in much less than 30', otherwise the the writing was "spectacularly slow" in information transfer terms.

[Anonymous Coward]

wrong!

Many scrolls http://en.wikipedia.org/wiki/Scroll pre-date the tapestry, and still exist. Scrolls are also a long ribbon of material containing information.

[MacroRodent]

Not to mention clay tablets

There are any number of records from the antiquity that beat the age of the Bayeux tapestry by thousands of years. Clay tablets are probably the record holders (the Wikipedia entry shows a picture of one from circa 2270 BC). OK, they are not tapes, more like memory cards (the precursor of punched cards?).

As recording devices, they have modern aspects, like being of more or less standard size and using a standard data encoding (at least within a given culture). The Bayeux tapestry was a one-off, but the clay tablets reflect an ongoing culture of long-term archival.

[Marvin the Martian]

"clay tablets"??

They're not really "ribbon based", eh?

[fch]

think of cylinder seals ...

... and you've got a tape write head. The information written is limited and repetitive, but there's nothing stopping you from creating nice ribbons with those.

The ancients had it all !

[John Sturdy]

Identifiying clay tablets

If you write the clay tablets in cuniform, you can identify them using cuniform resource locators!

[gerryg]

@Identifiying clay tablets

using cURL?

[Michael Wojcik]

scrolls and other long flat formats

Yes, Chinese paper scrolls with writing are around 2000 years old, making them comfortably older than Bayeux. And Egyptian papyrus is about 5000 years old. Textiles with ideographic information may be older. The "oldest in the world" claim is unsupportable.

As for the rest of the article, it's cute; but people do these sorts of calculations all the time (often along the lines of Tannenbaum's famous "bandwidth of a station wagon full of tapes" remark, though he doesn't actually provide the numbers). Even the serious data-comm texts I have on my shelves indulge in it once in a while, and you'll see it pop up from time to time on Usenet. A standard geek genre, in other words.

[Anonymous Coward]

Excellent article

A good reminder of the transience of our current information storage methods.

Today it's fast, but ephemeral. If you want longevity, then the data has to be encoded into the fabric of the substrate and the read heads need to be biological. It also helps that the biological processor hasn't changed much over the years.

So what's today's equivalent for long-term archiving?

[Thecowking]

Etch it into metal

and then strap said metal to a space ship and blast it off out of the solar system.

Voyager will most likely still be whizzing through the stars when the Earth is a blackened cinder. Of course it's mainly a write once read never mechanic, so you can achieve the same end by just melting the metal after you've written to it.

I think it's beta.

[Richard 12]

Microfilm and Microfiche

Microfilm is the 'tape' version, while Microfiche is the 'disk' version.

Microfilm is a lot more widespread.

To read it you just need a person and a magnifying glass. And patience, yes, lots of patience.

It's also machine-readable and relatively simple to copy, so you've got the best of both worlds.

[cosymart]

Paper

Preferably archival paper using suitable inks. So how much data is going to be lost forever with the move to kindle type machines and books that eventually will never get printed :-(

[Allan George Dyer]

So true, paper is indestructable...

it's not like it can burn.

Library of Alexandria anyone?

[MacroRodent]

Fires? Not a problem with clay tablets

Here clay tablets again show their superiority: Fire just makes them stronger.

[Allan George Dyer]

But

they don't do so well on the drop test.

[longbeast]

Why not ask for fast and long lasting?

Researchers and engineers aware of this need for high capacity, ultra-high reliability long term storage. The most promising approach that I've heard of involved electrically switched mechanical bits.

If you've etched a mechanical switch that is big enough to ignore radiation and thermodynamic effects, and you keep it somewhere sensible, it'll store data pretty much indefinitely. If you've etched several billion of them... problem solved.

I'm assuming that you do need some kind of locking mechanism to prevent vibration screwing up your data. Not really a problem if it's a write once, read forever type of system. The switch needs to be ultra reliable, the lock only needs to change state once.

[Chris Harden]

Awesome!

This has to be the geekiest and most awesome article I've read in quite a while.

(Almost as geeky as reading El Reg on a Sunday (aka unbillable hours))

[A 11]

Reading version

There's a copy in Reading Museum. I don't know how long it took them to make but they censored it as they went along, putting clothes on the naked men.

[Cliff]

Perfect place for it

Reading is home to Oracle, Microsoft and a bunch of other significant IT companies - so a wonderful quirk of fate.

[Anonymous Coward]

Fun article, I enjoyed that

now do the Lascaux cave paintings

[vonBureck]

Oh do come on

You mean you went on holiday to France and all you could think of was storage performance? Now THAT'S one dedicated tech hack.

Mind you, could be the start of an all new section on El Reg: a geek's-eye view of the world's tourist attractions, or how to make something commonly held as interesting mind-numbingly boring to all but the geekiest techs. Could even be made into a TV series, with James May to host it... Can't think of a suitably pun-laden title, though. Ideas, anyone?

Paris, obviously, as that's one tourist attraction all techs would like to cover.

[lglethal]

Making contractors proud...

You read it as Chris going on holiday and thinking purely of storage performance (extremely geeky i would have to admit). However, I read the article as Chris going on holiday and then thinking of an easy way to get the company to pay for the whole trip!

"Yep boss, i was investigating old archiving mechanisms and discovered this really amazing old method called the Bayeux Tapestry. I think it would make a really interesting story... Oh and heres my expenses for the holid.. i mean business trip..."

Good work that man!

[Big Al]

French? Mais non!

"a just war in which the French regained what was theirs by right"

'Normandy' was created as a fief by the Treaty of Saint-Clair-sur-Epte in 911, based on concessions made by the French King to the Viking leader Rollo - the word Norman being related to 'Northmen' (Normanni in Latin).

Normandy was not France, and the Normans were not really French - rather, they were Scandinavians ruling natives of Frankish and Gallo-Roman stock. Vassals of the French King they were, but that was far from being French!

[Archie The Albatross]

Quite so

The Normans despised the French.

A fine tradition which we continue to this day!

[Steve Carr 1]

And in the Kiwi corner.....

There is an awesome rendition of the Bayeux Tapestry in small town New Zealand.

http://www.1066.co.nz/mosaic/mosaic.htm

1.5 million "bits" of spring steel, individually coloured, so maybe say one byte per "bit". This is a scaled down rendition, not at the original physical size, but there has been no information lost, per se.

[Angus 2]

Love it..

Totally bonkers!

[whbjr]

Yes, yes, but it's really...

the first comic strip. Well, perhaps not the *first* comic strip, but one of the more significant ones. There's a direct artistic and storytelling link from the Bayeux Tapestry (which I've loved ever since National Geographic printed the whole thing in '66) to Peanuts, Doonesbury and the like.

But (said in my most condescending voice) if you IT folk would like a small claim to it, then we'll not object.

[Adrian Midgley 1]

Ridden down, not arrow

I understood (from an antiquarian and reliable source) that it shows Harold being killed by a mounted knight, with a sword. The guy with the arrow would be soemone else.

I like the article.

[tony trolle]

mmmmm

I remember something like that also, on TV years....ago

[Richard Cassidy]

Arrow or sword?

The illustration in the article shows the crucial scene, with the caption "Hic Harold rex interfectus est" positioned above both a figure with an arrow in the face, and a figure on the ground, who is being slashed with a sword to the thigh by a mounted knight.

R. Allen Brown, in his article 'The Battle of Hastings', in Anglo-Norman Warfare, adopts the view that the tapestry shows consecutive scenes - the figure with the arrow and the figure on the ground are both Harold. This interpretation has Harold mortally wounded by a chance arrow, then slashed as he lay prostrate.

[Graham Marsden]

Data Loss / Data Recovery

As mentioned most recently in a documentary hosted by Tony Robinson, on close inspection there appears to be a line of stitch holes indicating that there was once thread depicting an arrow in the eye of the figure on the ground who is being hit in the leg by the mounted Knight but the thread had somehow been lost.

If so, it's rather akin to recovering lost data using more sensitive tools than a standard read-head :-)

[Liam Johnson]

Missing arrow

But why did he change his socks in the meantime?

This is what happened. Harold was hit by the arrow, but not in the eye. The arrow simply grazed the side of his head and got trapped in the padding of his helmet. Bleeding heavily from the scratch to his head, he decided to leave the field and send in his double to take his place. However, due to last minute confusion at the laundry, the double was wearing the wrong socks! Nobody noticed at the time, but it is plain to see in the documents. Harold was officially dead, so he couldn’t just walk back into the battle. Superstitious people on either side would have simply killed the “ghost”. He was never seen again. Rumours of his location will be revealed in next weeks episode.

[Dave Bell]

With 'is 'awk in 'is 'and?

Yes, there's an argument for a different part of the picture to be the death of Harold. The Tapestry has a pattern linking pictures and text, and if Harold got it in the eye, that pattern is broken. it is, as I recall, to do with where the verbs are,

The Tapestry isn't the only source, and the eye version doesn't get mentioned elsewhere.

[lglethal]

@ Adrian

Im not sure anything can really be classified as a reliable source on matters as trivial as exactly how someone died in a battle over a thousand years ago (well it probably wasnt trivial to the person who died, but its certainly a detail guaranteed to be forgotten, reimagined, reportrayed and forgotten again many times over the course of 1000 years).

[The Nameless Mist]

1000 years old ??? someone can't do subtraction

Current Year = 2011

Year of the Invasion = 1066

Difference 945 years.

IN fact the item (its not strictly a tapestry) was likely commissioned several YEARS after the event so its closer to 935-940 years old.

[Tom 38]

To the victor, the history

The 'perfidity of Harold' is widely thought to be an invention of William in order to secure Church blessing for the invasion. Apparently it was going nowhere till he 'suddenly remembered' that Harold had sworn an oath on relics - pull the other one mate, it hath bells on it.

Plus, anything you read about the English that has been written by the French (whether 1000 years ago or yesterday) has to be taken with a smidgen of suspicion

[cloudberry]

could we possibly get over silly national sterotyping, already?

"Plus, anything you read about the English that has been written by the French (whether 1000 years ago or yesterday) has to be taken with a smidgen of suspicion"

How utterly refreshing.

[Neoc]

Calculations mayhem

"a rate of 2.168 bytes/per hour (17.344 bits/min)"

Er, no. 2.168 bytes/hour = 17.344 bits/hour = 0.289 bits/min ... or ~3.4 minutes/bit

[Chris Girocco]

Data Integrity

You are ignoring the fact that colors may have faded over time and therefore the integrity of each and every bit, except for the zero bits, is in question.

[cloudberry]

re: Data Integrity

But the author assures us that they haven't: "The vegetable dyes used to colour the threads have kept their colour for nearly one hundred decades". How can he know this?

[Anonymous Coward]

Or you could make your own in ten minutes...

http://www.adgame-wonderland.de/type/bayeux.php

(As seen all over Encyclopedia Dramatica...)

BTW, are these read-write heads bidirectional or uni-? And should we really consider the information content in it as a bitmap, or would a vector graphic description be a more realistic depiction of what was actually in the heads of the authors?

[Will Godfrey]

Accuracy

Nice effort, but I think the reporting accuracy is about as good as the 'document'.

Hmmm. I wonder what the equivalent of the BSOD is?

Nice Sunday evening reading anyway

[Chemist]

"I wonder what the equivalent of the BSOD is?"

Moths !

[Yet Another Anonymous coward]

@Ridden down

No that's what the ruling norman elite want you to think - he was actually hit from a grassy knoll not shown in the tapestry

[Anonymous Coward]

Reminded me of a question

in a text book years ago, to work out the data transfer rate of a St. Bernard with a 330K (shows age) floppy around it's neck, compared with a 300 baud (shows age again) dial-up connection.

[John Dougald McCallum]

Oor Wullie....

Duke William The Bastard "and his defeat of the perfidious British ruler Harold Godwinson"

UMMM Harold Goodwinson was not King of Britain But he was King of the Anglo-Saxon Kingdom of England.

[VictorVonZeppelin]

Squiffy Calculations

I think the calculations are a tad squiffy.

When you say "8 colours, at 1 bit per colour", isn't that not how binary bits work? 8 bits is 256 possible colours, while 8 colours is therefore 3 bits.

So, my maths (which admittedly may not be following the right route) would be:

(47*51678.72*3) to get bits. Then of course divide by 8 for bytes, and divide by 1024 for KB, which would be 889.4 KB

I hope my maths is on the right track there.