The internet archaeologist does not wear a little Indiana Jones hat.
Usually.
Mostly he sits in front of a computer looking at something named:
manual_final2_REVISED_old.zipAnd wondering what chain of human suffering produced it.
His job is looking for the old internet.
Not the current internet. The current internet is very easy to find. It comes looking for you.
The current internet wants you to subscribe to something.
It wants your email address.
It wants to send notifications.
It wants to know whether you accept cookies.
It wants to know whether you would like to sign in with Google.
The internet archaeologist is looking for the other internet.
The one that didn’t ask.
There used to be stuff everywhere
The old internet was built by people who apparently had no concern whatsoever about branding.
Some guy repaired oscilloscopes.
So he made a website.
Gray background. Blue links. Photograph of his cat for reasons never explained.
Buried four directories deep was a collection of service bulletins covering equipment manufactured between 1978 and 1991.
The guy died.
His ISP disappeared.
Google forgot.
Twenty-five years later, some idiot is staring at the remains at 1:13 in the morning because one of those service bulletins contains a capacitor value somebody needs.
Welcome to archaeology.
The artifacts are not pottery.
They’re PDFs.
Sometimes TIFFs.
Sometimes text files.
Sometimes a 640×480 JPEG containing the only surviving diagram of something extremely specific.
Sometimes you find a ZIP file containing another ZIP file containing a WordPerfect document created before several current programming languages existed.
You don’t open these files.
You negotiate with them.
The dead internet is remarkably talkative
People call it the Dead Internet Theory.
The stronger version says much of the modern internet has become bots talking to bots, automated content, generated pages, synthetic engagement and huge platforms recycling the same material until nobody is entirely sure where anything started.
I don’t need the grand conspiracy version.
I can see enough from the front porch.
Search for something obscure now.
You may get six websites explaining what the thing is.
Four buying guides.
Three videos with somebody pointing at the thing in the thumbnail.
A Reddit result quoting another Reddit result.
A page apparently written by an organism whose natural diet consists entirely of affiliate links.
And somewhere around page seven:
A 74-year-old retired engineer’s website.
Last updated 2004.
Bingo.
Unfortunately half the images are gone.
Sometimes you find people
This is the strange part.
You start out looking for a file.
You find a person.
Not literally. That would complicate the job considerably.
You find the outline they left behind.
Personal homepages are full of this stuff.
Someone documented every modification they made to a Commodore 64.
Someone maintained an enormous list of railroad frequencies.
Someone scanned the manual for a machine because the manufacturer wanted $38 for a photocopy.
Someone cataloged every pinball machine they encountered between 1996 and 2003.
Someone wrote eleven pages about getting a particular SCSI controller working under Linux.
Nobody asked them to.
There was no content strategy.
No conversion funnel.
No newsletter.
No “LIKE AND SUBSCRIBE.”
The page existed because some bastard cared about SCSI controllers.
This was apparently enough justification to publish something once.
Civilization peaked.
Sometimes you find arguments
Old forums are wonderful.
A technical discussion from 2001 begins:
Does anyone know the correct voltage for this board?
By message six, someone has the schematic.
By message twelve, the problem is solved.
By message seventeen, two men named Dave are accusing each other of destroying the electronics industry.
Perfect.
Modern community platforms occasionally preserve useful knowledge.
Old forums preserved grudges.
That gives them texture.
You can practically smell the cigarette smoke coming out of the HTML.
Sometimes you find a manual nobody remembers
These are my favorite.
An old piece of equipment exists.
Maybe hundreds still exist.
People search for the manual.
The manufacturer has merged four times, changed names twice and now produces software for hospital parking garages.
Its support page begins at products manufactured in 2018.
The machine is from 1983.
Good luck.
Yet somewhere out there is the manual.
Thirty-eight megabytes.
Crooked scan.
Page 47 has somebody’s coffee ring on it.
Page 83 was apparently fed briefly to a dog.
Still readable.
Still useful.
Still alive.
That’s the important distinction.
The web page may be dead.
The information isn’t necessarily dead.
Sometimes it has simply fallen behind the furniture.
You find abandoned neighborhoods
Entire sections of the old web feel like walking through an empty shopping mall.
The directory structure is still there.
The sign is still there.
The fountain has stopped running.
/~mike/
/~susan/
/~jhenderson/Every door used to lead somewhere.
Astronomy.
Recipes.
Linux.
Motorcycles.
Poetry.
A man’s exhaustive documentation of every Pontiac he ever owned.
Then everybody left.
No announcement.
No closing ceremony.
Nobody stood outside Geocities playing taps.
One day the pages just stopped changing.
That’s how most of the internet dies.
Not with deletion.
With the final update.
Last modified: November 18, 2002.
That’s the headstone.
And then there are the weird things
You find software written for companies that no longer exist.
Catalogs full of products nobody remembers manufacturing.
University project pages whose students are now approaching retirement.
MIRROR directories copied from servers that vanished decades ago.
Shareware registration instructions asking you to mail twelve dollars to Nebraska.
Documentation mentioning operating systems you haven’t heard named aloud since the Clinton administration.
Hardware compatibility lists.
BBS files.
Tiny utilities.
Driver disks.
Firmware.
Readme files containing one absolutely critical sentence buried under 900 words of angry developer commentary.
Sometimes you find a file that has been copied so many times nobody seems to know where it originally came from.
Sometimes three copies exist and all three are different.
One is missing pages.
One is compressed into mashed potatoes.
The ugly one is complete.
Of course the ugly one is complete.
Archaeology has changed
The old problem was finding information.
The new problem is finding the original information underneath ten thousand pages discussing the information.
There are moments when the modern web feels less like a library and more like five million people standing outside the library explaining what they think might be inside.
Search results increasingly favor what is recent, popular, optimized, commercial, socially active or machine-readable.
Fair enough.
Those things pay the electric bill.
Unfortunately a 1994 service manual has terrible engagement metrics.
Nobody has posted a reaction video.
The manufacturer has not partnered with an influencer.
The manual has zero followers.
Yet when your 32-year-old machine throws error code 7B, that ugly old PDF suddenly becomes the most important document on Earth.
For about fifteen minutes.
Then it goes back to being worthless.
That’s my neighborhood.
The bitter part
You find good things.
Then you realize how easily they could have disappeared.
A person scanned something.
Another person copied it.
Someone else mirrored it.
A forgotten server survived long enough for another system to notice it.
Twenty years later the file is still here.
Barely.
Meanwhile modern websites containing absolutely nothing appear immortal.
SEO slurry reproduces like algae.
A useful FTP directory dies because nobody paid the hosting bill.
A 3,800-word article answering no identifiable human question gets replicated across thirty websites before lunch.
Nature is healing incorrectly.
The archaeologist develops opinions.
The good part
Every once in a while you find it.
The actual thing.
No commentary.
No rewritten version.
No twenty-minute video.
No “Top 10 Things You Need To Know About Hewlett-Packard Test Equipment.”
Just the document.
There it is.
Somebody made it.
Somebody kept it.
Somebody forgot it.
You found it.
For a few seconds the internet works exactly the way you remember it working.
Then you notice page 63 is missing.
Son of a bitch.