I wanna post my idea for decentralised search again.
Uncategorized
15
Posts
2
Posters
3
Views
-
this can be enriched with better search tech as well such as document vectors.
How that would work is a search query has a bit of code that runs on it to convert it into a document vector; that is, a list of numbers that specify a location in topic space; this can be appended to a context vector your browser provides, which would work like a disambiguator; It would contain information about what topics you’re normally searching for:
vintage electronics, synthesizer modules, and sapphic erotica
that enables the search to rank those topics higher than others.
the index can include information about topic space and this enables sometimes higher quality search results than plain keyword search
that said, keyword search can be great, provided that you trust the index and it’s high quality
that’s where the web of trust comes in.
when you link out to a third party index, you can include a trust score. but bottom line is if it starts sucking, you stop linking out to it. the The decentralised search space is incentivised against scammy SEO practices. -
that said, keyword search can be great, provided that you trust the index and it’s high quality
that’s where the web of trust comes in.
when you link out to a third party index, you can include a trust score. but bottom line is if it starts sucking, you stop linking out to it. the The decentralised search space is incentivised against scammy SEO practices.if there’s a score then, when you have a transitive link- a link to an index contained in an index you link to…
if the score is between 0 and 1, you can multiply the scores out.
so a blog i trust 0.5 linked out to another blog it trusts 0.5; so i trust it 0.25-
this can be used in the ranking calculations
-
if there’s a score then, when you have a transitive link- a link to an index contained in an index you link to…
if the score is between 0 and 1, you can multiply the scores out.
so a blog i trust 0.5 linked out to another blog it trusts 0.5; so i trust it 0.25-
this can be used in the ranking calculations
it is then important that we begin treating search indexes as static assets, rather than an opaque api endpoint. something that is inspectable, scorable, filterable, cachable, archivable
don’t like that a search engine doesn’t show you exact phrases, or doesn’t index ‘{‘
frustrations aroumd being unable to find the music group “The The”?
you can fix it if you have the indexes and the indexers
sieze the means of discovery
-
it is then important that we begin treating search indexes as static assets, rather than an opaque api endpoint. something that is inspectable, scorable, filterable, cachable, archivable
don’t like that a search engine doesn’t show you exact phrases, or doesn’t index ‘{‘
frustrations aroumd being unable to find the music group “The The”?
you can fix it if you have the indexes and the indexers
sieze the means of discovery
Seize The Means Of Discovery
-
Seize The Means Of Discovery
let’s do for search indexes what RSS did for blogs, make site indexes statically findable, generatable, cachable and aggregatable
-
undefined oblomov@sociale.network shared this topic
-
let’s do for search indexes what RSS did for blogs, make site indexes statically findable, generatable, cachable and aggregatable
@bri7 you mean something like the sitemap.xml or something more sophisticated?
-
@bri7 you mean something like the sitemap.xml or something more sophisticated?
@oblomov go back to first post in the thread please *facepalm*
-
@oblomov go back to first post in the thread please *facepalm*
@bri7 oh damnit Tusky wasn't even showing me that this was a thread when I replied. I'll go read.
-
let’s do for search indexes what RSS did for blogs, make site indexes statically findable, generatable, cachable and aggregatable
@bri7 OK I've read the whole thread now 8-D
I wonder if a YaCy instance per website would do what you want.
-
@bri7 OK I've read the whole thread now 8-D
I wonder if a YaCy instance per website would do what you want.
-
@oblomov YaCy is just a completely different thing from what I am talking about. So is sitemap.xml.
And yes people do usually bring YaCy up when i talk about this idea. It’s kind of discouraging, makes me think I am not explaining myself well enough
-
@oblomov YaCy is just a completely different thing from what I am talking about. So is sitemap.xml.
And yes people do usually bring YaCy up when i talk about this idea. It’s kind of discouraging, makes me think I am not explaining myself well enough
@bri7 it's not the decentralized search per se, it's the search index, weighted web of trust, etc. AFAIK these are all things that YaCy provides (and ties together). Maybe you can clarify why something like a per-website YaCy instance does NOT fit the bill?
Gli ultimi otto messaggi ricevuti dalla Federazione
-
#Ucraina #quattroanni #guerra
Quattro anni, la guerra di #Putin
Basta.
Oggi per @repubblica
-
Trasferiti i colleghi indagati di Cinturrino. Assegnati a incarichi non operativi...tipo l'assalto di novaxsidvx alla CGIL prefetto piantedosi senza intervento di celerini poi buoni per askatasuna
-
Trasferiti i colleghi indagati di Cinturrino. Assegnati a incarichi non operativi
-
Non so se sono entrata in una sorta di spirale autodistruttiva, ma:
- Qualche giorno fa Vittorio è tornato a casa e ha trovato TUTTE le luci accese perché qualche ora prima mi ero messa a guardare al buio una serie di scene de L'Esorcista;
-
@Otttoz Qui nel quartiere, puoi chiederlo a chiunque, non si è mai vista (checché vada raccontando)
-
@WuMing2 Mestre fa parte del Comune di Venezia e comunque il circolo Uaar è su base provinciale 😅
-
@weirdocollector Grazie! Ci siamo formati mentre eravamo in fissa con i libri di Dune.
-
Fixing pollution should be paid for by those who profit off of it.
That means big plastic polluters, reckless industrial meat and dairy corporations, and climate‑wrecking oil and gas companies taking responsibility for the damage they’ve caused.
Add your name: https://act-int.greenpeace.org/polluters-pay-pact
