Time |
Nick |
Message |
00:25 |
MTDiscord |
<GoodClover> What's the shortest-lived issue/PR Minetest has? I'm in the competition now… |
00:45 |
|
Soni joined #minetest-dev |
01:21 |
|
Soni joined #minetest-dev |
04:00 |
|
MTDiscord joined #minetest-dev |
04:57 |
|
calcul0n_ joined #minetest-dev |
05:41 |
|
Fixer joined #minetest-dev |
06:07 |
|
Fixer joined #minetest-dev |
07:46 |
sfan5 |
@x2048 I guess 1.0 is a better default then |
08:21 |
|
appguru joined #minetest-dev |
08:27 |
|
Alias joined #minetest-dev |
08:54 |
|
appguru joined #minetest-dev |
10:09 |
|
calcul0n_ joined #minetest-dev |
10:13 |
|
proller joined #minetest-dev |
10:30 |
|
appguru joined #minetest-dev |
11:08 |
|
appguru joined #minetest-dev |
11:35 |
|
cranezhou joined #minetest-dev |
11:51 |
|
appguru joined #minetest-dev |
12:29 |
Zughy[m] |
Can we ban servers revolving around crypto from the serverlist? In these days there is a new one called "PAW Crypto Mine Server" stating in the description is about rewarding people with cryptocurrency. And of course, it constantly has 30+ players online |
12:42 |
MTDiscord |
<ROllerozxa> I'd say punish its score rather than ban. the server gives players an incentive to AFK in order to earn money so while the player count is pretty inflated it's not obviously botted, just an AFK farm |
12:42 |
|
Taoki joined #minetest-dev |
13:12 |
Zughy[m] |
I mean, 99% is scam |
14:28 |
|
appguru joined #minetest-dev |
14:40 |
|
olliy joined #minetest-dev |
15:23 |
MisterE[m] |
I think part of the spirit of minetest is that you are free to do with it as you like. If that involves using crypto when minecraft bans it, that's ok. If we have evidence that they are scamming people, that is a different matter. |
15:30 |
MTDiscord |
<Warr1024> You are free to do whatever you like with it, but you are not free to exploit shared platforms to advertise your projects if they are not in the spirit of those platforms. Minetest does and should allow crypto projects, but the serverlist is a different matter. |
15:32 |
MTDiscord |
<Warr1024> The scoring mechanism in the serverlist is intended to help players find an experience that the average player is likely to enjoy. The reason why player count is included is because it's assumed that an "average" player would like to play socially, and that other people remaining on the server are implicitly endorsing that it's fun enough to stay on. |
15:34 |
MTDiscord |
<Warr1024> Any server creating other incentives for players to hang around, e.g. promise of payment, is effectively misrepresenting that aspect of the score and should have their score adjusted, regardless. If they're doing it on purpose to knowingly exploit the scoring system just to draw in users, knowing the general purpose of the serverlist, then they can be subjected to more severe remedies. |
15:35 |
MTDiscord |
<Warr1024> I don't think they should be banned outright, in this case, but certainly should have their score capped. |
16:08 |
sfan5 |
I don't think either should happen |
16:20 |
MTDiscord |
<Warr1024> So wait, you mean you're fine with this? |
16:28 |
MTDiscord |
<Benrob0329> @x2048 Can games control the tone map that tonemapping influences? The whole point of TM is being able to adjust the coloring of the final render result, but if all games have the same tone map then it doesn't really make any sense |
16:28 |
sfan5 |
i dont think we as a project should do something about it |
16:29 |
sfan5 |
the question is kind of whether the server list is supposed to be curated list of acceptable servers or if it's just a list of servers |
16:30 |
sfan5 |
the current "rules" are a bit fuzzy regarding that, firstly you could argue the ranking it definitely curation but ranking != removal |
16:30 |
MTDiscord |
<Warr1024> If it's just a list of servers then it probably shouldn't have a single centralized "score" and instead should just dump the info to the clients and allow the clients to choose their own sorting and filtering methods. |
16:30 |
sfan5 |
except the informal rule is also that server names / descriptions should be family friendly |
16:31 |
MTDiscord |
<Warr1024> Hmm, "cleanliness" rules are not necessarily a thing we have full discretion on anyway; sometimes they're dictated by the platforms we want to make MT available on. |
16:31 |
sfan5 |
(which I don't think has led to any complaints or enforcement action in the last two years) |
16:31 |
MTDiscord |
<Warr1024> Have we ever had to, or attempted to, actually "ban" anyone or anything from the server list? |
16:32 |
MTDiscord |
<Warr1024> I do remember we changed scoring a few times in the past, and some of those changes may have led to certain groups ragequitting the serverlist... |
16:32 |
sfan5 |
pragmatically speaking it's just one server who cares, I don't want to start micro-managing what servers are allowed to do just for that reason |
16:33 |
MTDiscord |
<Warr1024> Right now we do have an incentive for people to add bots or encourage idlers in order to draw traffic to their servers. Ideally, we shouldn't have this incentive, or it should be limited to reduce the impact. |
16:33 |
sfan5 |
"no bot players" is also an informal rules but it's not like anyone goes around checking |
16:34 |
sfan5 |
rule* |
16:34 |
MTDiscord |
<Warr1024> Well, nobody officially goes around checking, but we have enough "busybodies" in the community that nothing goes unnoticed very long :-) |
16:35 |
sfan5 |
> Have we ever had to, or attempted to, actually "ban" anyone or anything from the server list? |
16:35 |
sfan5 |
aside from the obvious case of the person who shall not be named: maybe once or twice for non-family friendly descriptions or blatant botting |
16:35 |
MTDiscord |
<Warr1024> Hmm, okay. |
16:36 |
MTDiscord |
<Warr1024> I suppose ideally what we should do is find ways to change the scoring calculation such that it removes perverse incentives, but affects all servers equally. |
16:36 |
sfan5 |
for the record there are currently servers being prevented from announcing to the list |
16:36 |
sfan5 |
no servers |
16:36 |
sfan5 |
* |
16:36 |
sfan5 |
as in zero, 0 |
16:37 |
MTDiscord |
<Warr1024> If we just had the server report per-client idle time, and used that to change the weight of each player in how much they boosted the score, with active players counting for more than idle players, then that would make botting at least trickier. |
16:37 |
MTDiscord |
<Warr1024> and it would make paying players to idle more or less a non-issue. |
16:39 |
sfan5 |
that's the same micro-managing but as a technical solution |
16:40 |
sfan5 |
the current line of thinking is a "an active player contributes to the server's playability" which is very approximate but also reasonably correct |
16:40 |
MTDiscord |
<Warr1024> Right, that's what I'm proposing. |
16:41 |
MTDiscord |
<Warr1024> The current line of thinking is that a connected player contributes to playability. |
16:42 |
|
Desour joined #minetest-dev |
16:42 |
sfan5 |
when you add finer measurements to this a la "less idle time = better" you quickly move into the 'eh dunno' territory of correctness |
16:43 |
|
vampirefrog joined #minetest-dev |
16:43 |
sfan5 |
because nodes places, idle time, amount of regulars, ... may or may not correlate with playability depending on the game, what the player does or other factors |
16:44 |
sfan5 |
placed* |
16:44 |
MTDiscord |
<Warr1024> Eh, I was just thinking at the level of "how long has it been since either the last control input or lookdir change, or chat message" or something. |
16:47 |
MTDiscord |
<Warr1024> Really coarse-grained would serve the purpose, even. Like, if it's been more than 5 minutes since the last input, they're idle, otherwise, they're not. We don't need like a sliding scale or anything complicated. Could just cap the player count, for scoring purposes, at twice the number of non-idle players. |
16:47 |
MTDiscord |
<Warr1024> It'd require server owners to go to at least some trouble to setup bots, and that forces them to actually cheat the system in a more blatant way at that point. |
16:50 |
MTDiscord |
<Warr1024> If we really want the server list to be less "curated" then we should probably actually generally cap most score factors severely, so that there's less of a score gap between high-scoring servers and low-scoring ones, making the rankings less crystalized. With a system that's more prone to runaway positive feedback, you have to do more manual monitoring and curation to prevent it from getting stuck in a pathological state. |
16:50 |
sfan5 |
I believe all of the scores are capped |
16:50 |
sfan5 |
just not as low as you might be thinking |
16:50 |
sfan5 |
score factors* sorry |
16:51 |
sfan5 |
anyway #12790 will mix up the rankings probably soon |
16:51 |
ShadowBot |
https://github.com/minetest/minetest/issues/12790 -- Take geographic distance into account for server list ordering by sfan5 |
16:51 |
MTDiscord |
<Warr1024> oh, nice |
16:51 |
MTDiscord |
<Warr1024> I like mix-ups |
16:54 |
MTDiscord |
<Warr1024> You want to give players the best experience possible by default. One way to do this is to look for evidence of which experiences are better and direct them there. However, if you do this too strongly, then you end up unbalancing where they're sent to, which can be bad. It can overload a few servers while leaving many empty. It can also make it very difficult for novel experiences to get anyone to try them and gain any visibility. |
16:54 |
MTDiscord |
It also doesn't take into account the discrepancy between what a player asks for and what they'll actually enjoy; in this case, neither you nor them are actually good judges of that a priori, so the best approach is just to ensure there's at least a certain amount of noise in the predictions you offer. |
16:55 |
MTDiscord |
<Warr1024> Like, I would seriously consider adding a noise value to every score each time the list is served, so some things shuffle around randomly (though I'm not sure exactly how I'd decide how much noise). |
16:57 |
|
appguru joined #minetest-dev |
17:03 |
Desour |
it would be nice if we had all the info from the serverlist in the maintest main menu available, and if one could sort and filter by all factors (including serverlist score, but also server version (to exclude old and inofficial versions)). we could even define a score or sort function as lua code that takes all the parameters. also, it'd be nice to know if a server has a topic on the forums, or similar things that indicate that the owner has put some |
17:03 |
Desour |
work into the server setup |
17:11 |
|
appguru joined #minetest-dev |
17:14 |
MTDiscord |
<Warr1024> It'd be nice to have that kind of flexibility. Unfortunately, it's the kind of things where if you give users the flexibility they desire, then it's too complex for them to actually use. What a player wants is something like "I want to play on a server with lots of players, but only cool players, and a server with lots of mods, but only the fun ones, and not too old of a version but also not restarting every 5 minutes for updates". |
17:14 |
MTDiscord |
If you ask them what any of those mean, they basically tell you "well you're the programmer, why can't you figure it out?" |
17:16 |
MTDiscord |
<Warr1024> A Lua expression parser for score calc would never get used except by someone already savvy enough to just run their own serverlist proxy. Individual sortable columns would never really capture the combinations that players actually want. It'd be nice to have and might work in a few cases, but overall I don't think it'd improve usability enough to be worth the work. |
17:20 |
Desour |
I guess you're right. but I'd at least like to be able to see all the info from the serverlist in mainmenu without joining or looking in a webbrowser, to get some more info on what the server offers |
17:22 |
MTDiscord |
<Warr1024> It's a bit crowded on there right now, but configurable columns would be awesome :-) |
17:24 |
Desour |
I've rather thought of an interface where you select a server and on the side a tab (with tabs "join" "generic server info" "players" "mods") and then see the respective info on the side |
17:25 |
Desour |
(i.e. the player list for the selected server) |
17:27 |
MTDiscord |
<Warr1024> I've actually considered the idea of setting up my own proxy for the serverlist. It'd just scrape the main one, cache for a little bit, and do some custom sorting and filtering. People can already choose an alternate server list URL, and it would take a lot of the "curation" work off of the central one if more people just did their own curation and players chose list providers based on their own tastes. |
17:27 |
MTDiscord |
<Warr1024> (Assuming of course that curators are honest and transparent about what they're doing) |
17:41 |
|
ely-the-kitsune joined #minetest-dev |
17:46 |
|
ely-the-kitsune joined #minetest-dev |
18:38 |
|
ely-the-kitsune joined #minetest-dev |
18:44 |
MTDiscord |
<x2048> @Benrob0329 What type of control do you mean? |
18:45 |
Zughy[m] |
So title must be family friendly but I can create a crypto farm and go unnoticed? On a platform used by a lot of kids, considering that on Android there is a filter for non violent content only on CDB? |
18:46 |
Zughy[m] |
Like, what? |
18:47 |
Zughy[m] |
Also I thought MT was against crypto in general |
18:48 |
Krock |
crypto 🤮 |
18:50 |
ROllerozxa |
family friendly would most likely mean as in the server metadata shown on the serverlist, because I'm pretty sure there's some minetest servers that in-game have a lot of mature content like for anarchy servers |
18:50 |
ROllerozxa |
also that crypto server has about 60MB media worth of music, I think that stops most mobile players even if accidentally :P |
18:53 |
rubenwardy |
luckily, we don't have SSCSM - you can't do crypto farm |
19:00 |
MTDiscord |
<Warr1024> I don't think MT is "against crypto in general." I think we're against scams where users are tricked into doing something harmful. I think in this case though it's not necessarily something harmful, but possibly just something dumb and pointless. |
19:01 |
MTDiscord |
<Warr1024> Right now, about the only thing that tricking players to come onto your server really lets you do is use them to trick more players to come onto your server. Other than increasing your hosting costs, I don't know that there's really any value you can "exploit" them for. |
19:01 |
MTDiscord |
<Warr1024> I mean, you probably can't even sell advertising on an MT server. Who would buy it? How would they trust that you actually created the ad impressions you say you do? Are there even enough players to make it an audience worth targeting? |
19:04 |
MTDiscord |
<Warr1024> Re: the family-friendly thing, I think the major point of that is that any content that's displayed anywhere in MT other than stuff that happens after you join an online game has to meet Google Play Store's content guidelines, or else we get more "mature" ratings and lose access to a big chunk of audience. |
19:11 |
|
Desour joined #minetest-dev |
19:13 |
|
ely-the-kitsune joined #minetest-dev |
19:28 |
|
ely-the-kitsune joined #minetest-dev |
19:28 |
|
ely-the-kitsune joined #minetest-dev |
19:32 |
|
Noisytoot joined #minetest-dev |
19:38 |
|
ely-the-kitsune joined #minetest-dev |
20:13 |
MTDiscord |
<Benrob0329> @x2048 Control the coloring and strength, iirc originally MT just copied another game's tonemap and it was mostly pointless |
20:27 |
|
Noisytoot joined #minetest-dev |
20:35 |
|
Noisytoot joined #minetest-dev |
21:07 |
|
proller joined #minetest-dev |
21:21 |
|
ely-the-kitsune joined #minetest-dev |
22:36 |
|
panwolfram joined #minetest-dev |
23:04 |
Zughy[m] |
give me one example of crypto not turning into speculative assets/scam |
23:06 |
MTDiscord |
<Warr1024> "speculative assets" and "scam" ... these two are very different things |
23:07 |
Zughy[m] |
hence the slash |
23:08 |
MTDiscord |
<Warr1024> It's like asking "give me an example of any time a pixel artist isn't criminal/niche" |
23:08 |
MTDiscord |
<Warr1024> Yes, it's a fairly niche interest, and sure, some of them become criminals |
23:09 |
Zughy[m] |
yes, the ones doing NFTs, speaking about web3 |
23:14 |
Zughy[m] |
that's why we shouldn't support that shit. Exactly like Krita declared they're against NFTs and they're not gonna implement any export tool (contrary to what Photoshop did). Deciding to ignore such servers mean to be ok with them. Personally I'm not, Krock doesn't like so either, I wonder the rest of the team |
23:29 |
|
kilbith joined #minetest-dev |
23:48 |
|
AliasAlreadyTake joined #minetest-dev |
23:49 |
|
kilbith joined #minetest-dev |