Ghostbyte DevelopmentWe deployed the current version of FediSea. It is now available on https://fedisea.surf/
This is a very early version and is nothing more than the current development state.
The source code of all three repositorys is public now:
Crawler: https://github.com/ghostbyte-dev/fedisea-crawler
API: https://github.com/ghostbyte-dev/fedisea-api
To initially fill the software table of the database we used the data in the "software.json" file of communityDB (FediDB).
Gustavo.
Krafting@ghostbyte a better alternative to fedidb! Yes! Thank you for your work, Will be using it from time to time đ
the_rebel@krafting The figures are so completely wrong that you really shouldnât even look at them
Deviations of more than 20% are simply unacceptable
@ghostbyte
Deviations of more than 20% are simply unacceptable
@ghostbyte
@the_rebel @ghostbyte tbh I never trusted Fedidb numbers.... But I might be wrong
Edit: for example fedidb doesn't index pawoo.net... A 1M user instance...
Edit: for example fedidb doesn't index pawoo.net... A 1M user instance...
@krafting
FediDB does not index Misskey or similar services in full, which introduces further distortions.
However, there is not just FediDB, but other services as
https://fedi.wrm.sr/
or
https://fediverse.observer/list
The figures are quite similar, if you take certain fluctuations into account. Or bear in mind that some services are missing entirely.
A difference of just 20% is too stark. Perhaps the criteria have been set incorrectly, or certain areas are being overestimated, which then arenât included in the count
@ghostbyte
FediDB does not index Misskey or similar services in full, which introduces further distortions.
However, there is not just FediDB, but other services as
https://fedi.wrm.sr/
or
https://fediverse.observer/list
The figures are quite similar, if you take certain fluctuations into account. Or bear in mind that some services are missing entirely.
A difference of just 20% is too stark. Perhaps the criteria have been set incorrectly, or certain areas are being overestimated, which then arenât included in the count
@ghostbyte
@krafting Thank you very much!
But as other people already mentioned, please don't take the numbers too seriously at this state. It is a very early state of development.
But as other people already mentioned, please don't take the numbers too seriously at this state. It is a very early state of development.
crossgolf_rebel - kostenlose Kwalitätsposts@ghostbyte Sorry, but the figures differ massively from those of other statistical tools in the Fediverse, even when you factor in inaccuracies.
Discrepancies of more than 20% are simply unacceptable
Discrepancies of more than 20% are simply unacceptable
@crossgolf_rebel As we said in the post, this is not a finished product. The crawler isn't perfect yet. There are many edge cases the crawler has to take care of. Troll instances with fake numbers, robots.txt policies, handling DNS alias/redirects, ...