[AI text deleted, with questions about hardware, mesh, and uplinks]
In my opinion, this is NOT a question of technology, but of keeping a community of volunteers engaged over a period of at least 5 years and being able to repeatedly compensate for turnover within the team.
For that, you need
a) enough socializing that also reaches nerds who tick in completely different ways — multiple „channels“.
b) enough budget for a playground and perks to keep things interesting. That means hardware to play around with (that isn’t tracked too closely), plus enough VMs and rack space for private projects.
c) Separate teams (or at least separate people) for the roles of „accounting,“ „club administration,“ „press/social media,“ and „project management.“ (As soon as just two of these hats fall on one person, it gets REALLY difficult.)
As for the technical questions: that’s actually not important at all. The people who have to do it need to feel comfortable with it. Everyone who’s been doing Freifunk for more than 5 years has experience and usually strong opinions. And „failing together + starting over with a different concept“ is also part of the shared learning process.
Suggesting even just ANYTHING from outside leads to the „not-invented-here“ effects, of which we already have plenty in Freifunk.
Far too often, FF communities have lacked shared goals in the past.
If the only metric left is „lots of points on a map“ and „lots of green lines = good,“ then the end is foreseeable.
Then there’s the danger that a larger number of highly motivated node installers promise the moon in order to get nodes set up at people’s places, hoping that this Wifimesh thing with „as many nodes as possible“ will just sort itself out.
And a few people in the backend who take the frustration when that doesn’t happen, and quietly bail out one after another. And then there are no new people for that area either.
So you should ask yourself early on what the long-term goal actually is:
Being a community around the hardware and having fun together, with the occasional PR-stunt press release now and then. So basically in the style of a private/civilian rescue dog squad.
Or you see it as an educational mission — and not as educating „others,“ but as further training for the active members. Self-learning within the club. If a bit of internet falls out here and there as a bonus, all the better.
Or you put the SLA service-provider character front and center: found a WISP/WLL cooperative for internet supply. Pay people for the unpleasant (but important) things. Procure hardware that’s at least entry-level „carrier grade,“ don’t just treat BG/VDE/lightning-protection topics as a derailing subject at regulars’ meetups, but simply bring in people who handle it pragmatically without endless discussion. And respect the paying customers as customers. Because — lacking the beaten-down, wage-dumping sub-subcontractors like the big providers have — they’ll even pay more than with the red, blue, or magenta-colored giant.
There are of course surely a dozen other models. My simple recommendation is to question your own approach and, if necessary, carry out a realignment.
In my opinion, this is AI-generated spam. 
(See the link in the name at the end)
It’s ‚unfortunately‘ no longer accessible, but the title alone made me expect link spam. Despite the attempt to turn this into a thread, I’m in favor of deleting the whole thread.
If these questions really are from an AI, then one can only congratulate it on the choice of topics.
Which hardware is best suited for long-range connections: ??
Are there any specific firmware optimizations for improving stability and range: ??
Any experiences with mesh networks over large open areas: ??
In other words: if this is how AI scrapers perceive what the focus of Freifunk is (i.e., among other things, the discussions here), then it says something about how, in my opinion, we’re not setting the relevant priorities. At least not the ones that would put a stop to the community dying out (see Helgoland a few months ago, or currently Freifunk-Magdeburg).