Thursday, December 15, 2011

You like day-ta, I like dah-ta, let's call the whole thing off

Complaints about service in the Riddled library are neither common nor wise. There can be discouraging consequences. Feral books are rumoured to have escaped into the Reference Section at inconvenient times for library users with a record of discontent. Not to mention chocolate digestive biscuits in the tea room that are so digestive they really belong in a D&D adventure.
Best not to speculate that important documents are sometimes tardy to reach the research staff because the library pixies have been hanging onto them and savouring that new-data smell, in the half-size fractal shelving they occupy between the second and third digits of Dewey classification. I accept the pixies' explanation that there must have been an obstruction or stricture along the series of tubes between the US and here.

Anyway, here is WIRED, only 3-½ years old, practically recent! What nostalgia the article evokes for those far-off halcyon days when the world was still fresh and Google still placed higher priority on web indexing than on social media!

Apparently Google provides the role-model for Science to follow [it is fairly clear where the author hoped to seek employment next]: "What can science learn from Google?" Scientists should be looking for correlations rather than causality, and stop worrying their pretty little heads about 'meaning' or 'explanation'. Explanatory theories are so much harder these days than the ones the author remembers from school, so obviously they will only become harder still.

I'm not sure quite how this post-theory science in the Laboratory of Babel will work. Evidently the sheer accumulation of data in "the Petabyte Age" -- orders of magnitude more data than in any library ever imagined, shut up Mr Borges -- will lead to some sort of Teilhard-du-Chardin noötic phase transition. Science will crystallise out of the supersaturated solution of Facts without human agency.

Then lightning bolts strike from a cloudless sky to fuse the power switch permanently on, and nuclear warheads detonate in their solos. This always happens. It must be an old tradition, or a charter or something.

This is all vaguely familiar from some old parable:

"Pugg meanwhile sat propped up against the barrel and read, as that diamond pen which the Demon employed to record everything it learned from the oscillating atoms squeaked on and on, and he read about how exactly Harlebardonian wrigglers wriggle, and that the daughter of King Petrolius of Labondia is named Humpinella, and what Frederick the Second, one of the paleface kings, had for lunch before he declared war against the Gwendoliths, and how many electron shells an atom of thermionolium would have, if such an element existed, and what is the cloacal diameter of a small bird called the tufted twit, which is painted by the Wabian Marchpanes on their sacrificial urns, and also of the tripartite taste of the oceanic ooze on Polypelagid Diaphana, and of the flower Dybbulyk, that beats the Lower Malfundican hunters black and blue whenever they waken it at dawn, and how to obtain the angle of the base of an irregular icosahedron, and who was the jeweler of Gufus, the left-handed butcher of the Bovants, and the number of volumes on philately to be published in the year seventy thousand on Marinautica, and where to find the tomb of Cybrinda the Red-toed, who was nailed to her bed by a certain Clamonder in a drunken fit, and how to tell the difference between a bindlesnurk and an ordinary trundlespiff, and also who has the smallest lateral wumpet in the Universe, and why fan-tailed fleas won't eat moss, and how to play the game of Fratcher-My-Pliss and win, and how many snapdragon seeds there were in the turd into which Abroquian Phylminides stepped, when he stumbled on the Great Albongean Road eight miles outside the Valley of Symphic Sighs—and little by little his hundred eyes began to swim, and it dawned on him that all this information, entirely true and meaningful in every particular, was absolutely useless, producing such an ungodly confusion that his head ached terribly and his legs trembled. [...]

"And it grew dark before his hundred eyes, and he cried out in a mighty voice that he'd had enough, but Information had so swathed and swaddled him in its three hundred thousand tangled paper miles, that he couldn't move and had to read on about how Kipling would have written the beginning to his Second Jungle Book if he had had indigestion just then, and what thoughts come to unmarried whales getting on in years, and all about the courtship of the carrion fly, and how to mend an old gunny sack, and what a sprothouse is, and why we don't capitalize paris in plaster of paris or turkish in turkish bath, and how many bruises one can have at a single time. And then a long list of the differences between fiddle and faddle, not to be confused with twiddle and twaddle or tittle and tattle, then all the words that rhyme with "spinach," and what were the insults which Pope Urn of Pendora heaped upon Antipope Mlum of Porking, and who plays the eight-tone autocomb."

This parable is not news to the Google programmers.

Hang on, there's a message from the Library; the book I ordered is ready to collect. BRB.

Left: high-density information access

Right: Hidden costs of data mining


Big Bad Bald Bastard said...

Feral books are rumoured to have escaped into the Reference Section at inconvenient times for library users with a record of discontent.

Keep a wary eye open for the dreaded Krankenhammer, by Stefan Schimpf, the Mad Cobbler of Mainz!

ifthethunderdontgetya™³²®© said...

and how to play the game of Fratcher-My-Pliss and win...

See, this is why Riddled is on me bloog-role.

Also, the cat pictures.

Substance McGravitas said...

Now a Demon of the Rule 34th kind is a necessary thing.

Smut Clyde said...

Link to the parable doesn't seem to be working, so here's an alternative.
Or complete text here.

Substance McGravitas said...

Love them blinking lights!

tigris said...

stop worrying their pretty little heads

Misspelled "pointy."

Rachel said...

You can bet that early feasibility studies of sub-prime mortgage crap involved 'correlating' abundant amts. of data using super duper maths/statistics formulae, that simply added as one parameter: plus, will make me rich. They forgot to add this parameter: plus, will not cause global financial collapse.

Also, sure, random gunfire eventually hits a target, but you don't learn anything about aim.

Also, my sister told me of an American system of introducing birth control to impoverished women in Latin America. The women were given the pack of pills, and also a beaded necklace on which a bead was to be pushed along each day, as a reminder to take a pill daily. So when they forgot to take a pill, they'd just move a few beads over and think "bring it on, Cesar".

Models must be employed, and must be understood, and must be humane. Our strange interest in being dis-interested is a lie... usually a self-serving one!!! It's Babel baybee.

Oh, and also too, Goedel's Incompleteness Theorem says no consistent system can ever attain complete self knowledge. So there! I got a C on the philosophy paper in which I noted this theorem, arguing NO re: "Can Machines Think?" I think the guy just didn't 'get' me.

Smut Clyde said...

Evidently 'Rachel' is really Sir Roger Penrose.

What I really enjoyed about the Wired article was the author's explicit distinction between old-fashioned scientists like Darwin and their tradition of collecting unstructured observations from which decades later a coherent theory would emerge, versus the new Google-style breed of scientists like Venter who are collecting unstructured observations.

Rachel said...

There, got you to spell it out for me. Phew.

mikey said...

I spent the first part of this year working with gigantic NoSQL databases and Map Reduce distributed processing tools, primarily Hadoop, which is somewhere between amazing and bloody miraculous.

Of course, after getting corporations to spend millions putting in the underpinnings, building the databases, writing the connector and business logic code and then developing the Hadoop queries, when the answer turned out to be obvious or gibberish we always just looked at them smugly and told them "Well, you know, you DO have to ask the right questions"...