What happened to finding things
Why the web feels smaller than it used to, what the discovery layer actually was before search consolidated, and why every attempt to rebuild it stays small. Sourced throughout, rather than nostalgic.
What this section is
Reported and analytical writing about how web discovery worked, why it stopped working that way, and what that means for anything too small to be a brand. It is the part of the site that explains why a register like this needs to exist at all.
It is also the part with the heaviest sourcing rules. Every historical claim is cited to primary or contemporaneous material recorded in the ledger with the date it was retrieved. Widely-repeated figures about the early web usually have no source behind them, so where we could not find one, the article says so instead of repeating the number.
Start here
What we lost when the blogroll died is the argument in one piece: human-curated outbound links were the discovery layer, what replaced them optimises for something else entirely, and any honest alternative has to accept a constraint the replacement does not.
It includes the unromantic reasons blogrolls disappeared — maintenance burden, link rot, and the sheer labour of curation — because a story where they were simply taken away is a more comfortable story and a less true one.
The history
A short, sourced history of serendipitous discovery
Human-edited directories, webrings, blogrolls, shared queues and the random button — organised by mechanism rather than by company, so the analysis outlives the brands. Read it.
What a directory was supposed to be
The original editorial-directory idea, what it got right, the four ways it failed, and which of those failures we have not solved either. Read it.
What a webring was
A ring of sites that agreed to point at each other: how it worked, why it collapsed, and what the current versions do differently. Read it.
One thing this section will never do: claim continuity with, or succession from, any earlier discovery service. Those companies existed, some of the people who built them are still working, and we are not their heirs. We reference them factually and stop there.
The analysis
Why every search result looks the same now
The same brands, the same page shapes, the same eight answers — the structural reasons, without attributing intent to anyone. Read it.
Why the long tail got harder to find, not easier
More small businesses online than ever, and less findable than ever. The paradox, and the four causes behind it. Read it.
Every revival stays small, and why that's the point
A fair survey of the discovery tools people actually use, and the constraint they all share. Read it.
The cost of intake: why almost nobody prunes
Adding a listing is a one-off cost; keeping it true is recurring. Follow the money and dead listings explain themselves. Read it.
Where to look instead of here
Where people still find independent work is a linked reading list, most of which is not us, with a stated check date and a refresh cadence. It includes routes that have nothing to do with the web — markets, noticeboards, and people telling each other things.
We name and link the other tools in this space generously and describe them accurately, including where they do something better than we do. Ranking them would be a strange thing for a site that refuses to rank anything else.
How this connects to the register
The argument of this section is that curation does not scale, that nobody funds accuracy, and that the discovery layer disappeared because maintaining it was unpaid work. The register is a small attempt to do the maintaining anyway, with the dates published so you can see whether we are keeping it up.
If we stop, the check dates will show it before we admit it. That is the point of publishing them.
Questions about this section
Is this just nostalgia for the old web?
We try hard for it not to be, which is why every historical claim needs a contemporaneous source and why the pieces include the boring reasons things ended. Nostalgia makes bad arguments because it only remembers the good half.
Are you claiming to be the successor to anything?
No, and that is a rule rather than a preference. Earlier discovery services were built by people who are largely still around; describing ourselves as their continuation would be untrue and unfair to them.
Why no statistics about how much the web has consolidated?
Because the figures that circulate mostly trace back to each other rather than to a source. Where we could not find a primary one, the article says the claim is unsourced rather than repeating a number that looks authoritative.
Do you criticise search engines here?
We describe mechanisms and avoid attributing intent to any company. The convergence of results has structural explanations that do not require anyone to have decided on it.
Why link to competing discovery tools?
Because a register this small cannot be anybody's only source, and pretending otherwise would be the first dishonest thing on the site. Several of them are good and do things we do not.
Will this section ever be finished?
No — the reading list needs refreshing and the analysis dates. Where a piece has a check date on it, that date is the honest statement of how current it is.
Every claim in this section resolves to a source in the ledger with a retrieval date, or is marked as unsourced.