Post by Neodome AdminPost by U.eeUsing only stored articles. I did little comparison and indeed: some
sites are indeed only represented in cleanfeed rejected feed.
I don't know right now, what to do with them, probably should be
included next time.
There’s a similar project, Top1000. Not sure if you are aware of it.
<http://top1000.anthologeek.net>
It collects data from participants and tries to create a graph of Usenet
distribution per site.
It is well known project, put I don't think there is overlapping.
Post by Neodome AdminIt’s interesting that I see some sites in your data that don’t appear in
their lists. I participate in that project and I remember seeing sites
popping up with only one or two messages sent from them. Perhaps some
glitch.
Post by U.eeThen again, my mai problem is huge node count, usenet is too big :)
Just need more collaboration between sites to be able to fold bunch of
nodes in one. Many of them are actually virtual nodes, i.e. just another
names for the same server.
If you follow some earlier messages, we had discussion with Grant Taylor
about that.
Post by Neodome AdminWhen you’re trying to create more readable map, is it possible to change
the thickness of connecting lines representing approximate amount of
articles flowing from one site to another?
That is interesting idea, but that is place, where Top1000 should be
relevant:)
I am afraid that this isn't viable with only one collecting site.
With peering relations, things are easier: I see one path containing
specific relations and that is good enough, even when 1000 other paths
are different.
With article count though, things are heavily depending on my peering
relations.
I can't use top1000 public data either, they show generic article count
per site, there isn't peering information.
Best regards
U.ee
