Is gnutella community still alive?
 

I don't see much activity here.

Is there someone still interested?
Is there any p2p client developer?

The Gnutella Forums (GF) is still actively used as a reference site but there's minimal (almost rare) posting from users. For example, if I post a link to a program patch or a connection installer, it does not take long (a day or within days) before there's lots of downloads. I suspect people are a little shy about the GF & think it's somehow linked to LimeWire.org (LW) & the authorities which of course is not the case because it's independently owned & no information about GF members has ever been compromised. If it had, I'd have left the site a long time ago, and the other moderators feel exactly the same. Our privacy and the protection of the privacy of members is of the utmost importance to us.

7 to 8 years ago there might have been anywhere up to 20-odd new posts a day. Now it's lucky if there's one new post in a month or two.

As for Developer Discussion, well the GDF became even quieter several years ago. The old developers kept a low profile & still now, some are no longer committing to their program. GTK-Gnutella is an exception & came back out of the woods a few years ago but all their developer discussion is (& always has been) limited to their GTK SourceForge site.

Not sure how often any GTK devs pass through the GF, once in a blue moon I guess. :D

There just does not appear to be the interest in open source p2p projects like there was last decade in regards to offering some kind of input, testing or developing.

Active programs? GTK-Gnutella. WireShare (Bigjx is the only actual coder).
I've noticed a couple of LW4-clones silently get updated now & again though I've no idea what changes they are making because they still do not support DHT or TTL. I also never see them (LW4 clones) on GWeb sites. Their default gwc sites are long out of date.

I'm actively developing my GWC in this period so if any p2p client developer need a particular feature it is time to say it :)

I had an interesting idea that come from a document found over the internet, a sort of prioritization of IP near the IP of the requesting client, teorically near IPs are also near locations (not all cases but many since they are probably assigned in groups) and this probably speed-up connections since they trying to connect to near hosts.

For the ones that create LimeWire mods it would be good if the switch vendor codes from LIME and provide a correct User-Agent otherwise they may be blocked.
Many LimeWire mods use LIME vendor code and this is confusing, if a small mod doesn't want to create a specific vendor code I suggest LIMM as generic LimeWire mod vendor.

I noticed LIMM listed in the fake-clients list, so is that wise? I don't know where LIMM originated from.

Several LW-4 Clone programs use LIME as their vendor code. At least one or more don't which probably causes problems for them connecting (such as to GWC's.) Since LimeWire, FrostWire, Cabos, & WireShare are the most popular clients on the network and reserve a certain number of connections for those with LIME vendor code (however most of the LW4 clones either include Wire or Lime within their name.) Definition of LW4 Clones is here. The LW 4 Clones are all GWC based programs, having never upgraded to UHC, DHT or TTL abilities AFAIK. Their connection & other data files are usually the tell-tale sign of this.

According to the LW/WS code, locale is usually prioritized where it's applicable.
//move the connections with the locale pref to the head of the lists
//we prioritize these disregarding the other criteria (such as isGoodUltrapeer, etc.)

My code automatically switch vendor from LIME to LIMM when it detect a mod in some ways, they aren't necessarily bad; then I block bad mods while allowing other mods.

If you look in the logs then you can see that in the query there is "&client=LIME" but in my detection I set it to LIMM to make clear it isn't the original LimeWire.
My detection isn't perfect but it should be enough accurate in most cases.

PS: I was talking about prioritization in the hosts returned by the GWC.

GWC should have been optional along with a built in UI for adding and removing GWC hosts.

I like both those points you made.

In Shareaza there is already.

Haven't used Shareaza since I stopped using Windows (2004). I bet the app has come a long way since then.

Yes there was much changes, now it also support the DC++ network.