Yes, I would think 5. goes in the right direction.
I don’t think that my idea with Blockchains was very good either.
But maybe, there is a way of doing this, that isn’t adding in a lot of junk to the INDEX and also doesn’t make the world too complicated.
We need a way of establishing consent.
We have two classes of peers in the Search Engine:
- Crawlers
- Searchers
A Crawler is a peer, that adds to INDEX and verifies other who add to the INDEX.
The crawlers use Public Key Cryptography to sign off on each others updates and to the INDEX.
Before a crawler signs off on an INDEX File Change,
they verify the changes.
Every searcher keeps a list of peers they accept as crawlers.
This they can do, first of all, because they can decide to trust somebody or because they have a version of INDEX already
and thus the Identity of the crawlers, that signed off on those changes.
This they do, both on the INDEX File and by putting their currently most recent and supported version of INDEX on their IPNS.
And when a searcher wants to fetch the new version of INDEX, they just look at the IPNS records of these trusted crawlers and fetch the version, that the majority of those currently has put there.
The advantage of this is, although a group of crawlers could potentially split of the other crawlers, they will still mostly have an eventually consistent global INDEX.
Those splits would be very much intentional though.
To start a new INDEX File, you would create:
- An INDEX File Format / Database.
- Implement the crawler.
- Implement the searcher.
But the question is, would there be a use for such a search engine?