In ancient times - say, before

the Arpanet - libraries were

occasionally destroyed along

with the rest of a building, a

city, a culture. These

charmingly unambiguous

accomplishments of our ancestors

provide a historical motif that

we can instantly recognize and

reflexively deplore. True, some

are probably grateful to have

been spared another semester's

worth of soporific classics, but

on average society still

considers the book-burners the

bad guys. In our postliterate

age, the main danger to

libraries is not marauding Huns

but perceived obsolescence, the

notion that libraries are quaint

and elitist, sort of like

classical music and the daily

paper. The Internet, in some

fat-pipe dream of universal

access, will relieve us of the

burden of research in the same

way that television relieved us

of reading for news and

entertainment.

The fantasy of being digital

insists that information is only

as good as the methods available

to search for it. Online

information rules here, since it

can be referenced in previously

unimaginable ways: by geography,

by heuristically determined

literary style, by quality of

pornographic content. The

reductio ad absurdum of this

idea is the so-called

intelligent agent, a vessel in

which, to paraphrase Jaron

Lanier, we can dilute our

personality, a device that will

ultimately allow the information

that we want to come looking for

us. In the plot of this just-so

story each host on the net is

another monkey typing, and at

the happy conclusion the agent

brings us Shakespeare.

It would be difficult to argue

that searching for information

on the Internet is inconvenient.

Already many people look to the

web before any other information

source, and therein lies the

problem. Even if you can reduce

the 10,000 hits from AltaVista

down to a handful, a really good

source must be both authentic

and permanent, or as computer

scientists would say,

"persistent." The design of the

Internet and the web is ideal

for exchanging ephemeral data

between trusted parties. Net

experts will argue that

authenticity can be ensured on

the Internet using PGP-like

schemes, but this provides only

technical trust, and doesn't

account for the ill-defined

shared confidence that most

people mean by "authenticity."

On the Internet, it seems that

consensus has been found by the

infra-left and the ultra-right

("We are all puppets controlled

by unseen masters"); but the

visible portion of the political

spectrum only manages a

confusing pluralism that mirrors

society at large. Which isn't

surprising - this is a culture

where we can see Beavis and

Butthead and Portrait of a Lady

at the same cineplex, and no one

is too surprised if it's the

Henry James fans who wear

Metallica T-shirts and slurp

jumbo Cokes. God only knows what

might get by the Internet's

drain trap and settle into

respectable thought.

Libraries at least serve as a

cultural band-pass filter, an

imperfect but workable way to

salvage some signal from a flood

of noise. Traditional libraries,

though they may be inefficient

and bureaucratic by the

standards of the Internet, are

more than book repositories.

They are a naive sanctification

of knowledge, manifesting the

hopeful belief that high

ceilings and artificial

solemnity will bestow

credibility. As for permanence,

books may be stolen, destroyed,

lost, or misfiled; but they

seldom change their own content

and they're usually stored in a

format that people can still

read 10 years after publication.

When information does become

mostly electronic, maintaining

it will require elaborately

conceived digital libraries.

David Levy and Catherine

Marshall of Xerox PARC, who

design such things, understated

things nicely: The Internet's

infrastructure "lacks the

crucial institutional services,

such as collection development

and cataloging, by which

collections are stabilized for

ongoing use." Even a

congressional subcommittee could

determine that current libraries

do this job pretty well.

If all knowledge really were

power, or even a direct

antecedent, no one would

question the relevance of

libraries. Rather, we'd insist

upon - and probably pay for -

these sources of information.

But much of what libraries hold,

especially basic research, has a

deferred value that engenders a

justifiable, if unfortunate,

lack of exploitation. The

Internet, though, has already

passed the peak of its scruples,

when the government discovered

some time back that it had

accidentally created something

useful. When the well-heeled

invent a mechanism for buying

preferred packets like they do

first-class airline seats, the

economy section will get

light-headed from sucking bits

through a narrow straw. It will

be a bad time to find that the

city library has been converted

to a parking garage.