XSF Discussion - 2024-02-21

does anyone have an idea for something interesting they'd like to do with the domain xmpp.beer ? otherwise I'll probably just let it expire

Hm. It will be immediately be registered and then sold for 3000$ I bet

I run an instance that strictly loses money (I do not accept donations). Part of what is great about the Fediverse is that it does not have to make money to be successful. The costs are distributed among so many people, that it's a reasonable expense to deal with in many cases.

> Unfortunately instances need to make money to be sustainable to cover the hosting costs as they get more users. That's true, but instances do not need to continuously get more users for the Fediverse to be successful. In fact, that would lead to an unhealthy federated network. See what happened to XMPP with Google Talk for example. For the Fediverse to be successful, the cost of running it must be distributed. > In the end, the only instances that will be around are the largest and most centralized instances that have the most users. I absolutely disagree. It's rare that small instances struggle to keep the lights on, because they are so cheap and easy to run. You can run Pleroma for a few people easily on a $5 VPS or a Raspberry Pi in your closet. Source: I did that. The only reason I don't anymore is because I migrated my instance to my Dell server, because I have one. Big instances are the ones that struggle the most, because they're the ones that the cost is real enough to be a problem that needs to be considered. See rglullis' comment for an example of this. See the mastodon.technology shutdown[0] for another example. > Thus, even if that happens, donations will not be a viable source of revenue to counter the increasing operational costs as an instance gets used by more users. I agree. [0] https://news.ycombinator.com/item?id=33120136

topgun, basically true, however in particular with the activitypub-based fediverse there are issues around small instances. it's much harder to reach out and meet and hear about people. It's possible, but it's harder; you need to resort to relays and such, which are options many don't even know about.

The two posts above I took from hacker news and it's a very good insight on how to know how to analyze a Federated communities sucesso based on it's financial aspects

Astro did great work there with relay.fedi.buzz, but still.

Stepping Back from the Tusky Project - https://news.ycombinator.com/item?id=37286543

Above is the original article

MSavoritias fae.ve: I think you will relate to that article a lot

> topgun, basically true, however in particular with the activitypub-based fediverse there are issues around small instances. it's much harder to reach out and meet and hear about people. It's possible, but it's harder; you need to resort to relays and such, which are options many don't even know about. Yes, it only works if it maximizes discoverability of smaller instances (provided they want to be discovered) and minimizes barries to entry like cost of running an instance and technically knowledge required to run it ↺

i agree that instances need to have a cap on how many people they have. and also i am against open registrations. i think xmpp could learn a lot by fedi too also. especially the thematic instances concept which i think makes the server more of a community and the operator less likely to have a burnout and more likely to think they do something meaningful

That article seems to make a distinction between operators (running instances) and developers (coding stuff). I think it's probably reasonable to further distinguish between server and client developers. I believe that the latter have it slightly easier to generate a revenue stream.

"dump relays" doesnt work anymore

and i doubt it ever worked

sorry, the _former_ (server devs) have it slightly easier to generate a revenue stream, in my experience.

right i remember the mastodon.technology shut down. we had something similar happen twice in xmpp in the last year

at least they gave some time to migrate

There's a model to get an intuition about which forces shape an industry called Porter 5 forces. It's generic to any industry, but it provides also Insights for our case ( social Network/chat). I'll post a link to the generic model and one applied to Facebook, which can be useful for us:

General model: https://www.investopedia.com/terms/p/porter.asp

Applied to social media: https://www.forbes.com/sites/greatspeculations/2014/11/28/facebook-through-the-lens-of-porters-five-forces/

trying to apply capitalistic concepts to social networks that is ultimately about people talking to each other just devalues said social networks. people in fedi have been talking for a while now about how growth and user base and income all dont matter if we want to make healthy social networks. essentially a post-capitalist model

I'd offer that without income, there's no room for developers to contribute enough to make a social network in the first place.

> trying to apply capitalistic concepts to social networks that is ultimately about people talking to each other just devalues said social networks. > people in fedi have been talking for a while now about how growth and user base and income all dont matter if we want to make healthy social networks. > essentially a post-capitalist model It's not about capitalism, it's about factors that shape any project. For instance, running alternative social network has to be cheap, as it's done usually by volunteers. It has also to somehow attract Facebook users etc. So they it's good to know which incentives Facebook users needs to try alternatives. ↺

topgun, why does it "have to" attract people that use facebook?

> I'd offer that without income, there's no room for developers to contribute enough to make a social network in the first place. i agree. i should have phrase it more as maximazation of profit under hierarchical structures. but it sounded too much of a mouthful :P a social network needs to be sustainable in income and manpower to be viable

> I'd offer that without income, there's no room for developers to contribute enough to make a social network in the first place. The developer, specially the ones working for free, should have low Costs to develop. This means it should use a programming language which most of them know. The volunteer person running the server needs a software which can handle the most users for his low Budget. So by this you see most of developers would benefit from JavaScript or Python, but for server runners it's best something resource efficiently like Erlang/elixir or rust. ↺

> topgun, why does it "have to" attract people that use facebook? You have to attract people, and given that twitter has 75% of market share ans Facebook has 15%, your user base will probably be someone who before was either in Twitter or Facebook ↺

i would argue we need more server software that is not designed to scale. like the kitten project. > Will this scale? > *le sigh!* > (Yes, it will scale for the purposes it was designed for. It will not scale for the purposes of farming the population for their data and destroying our human rights and democracy in the process. That’s a feature, not a bug.)

I can't see how the choice of a programming language is even relevant, to be honest. WIth roughly (same order of magnitude) amount of training, a given developer is roughly as efficient (again same order of magnitude).

same

Guus, [citation needed]. Extreme example: x86 assembly ;P

sure, you can easily find outliers

I *thought* I'd seen different numbers, but I could misremember. Got the source?

I do not have numbers to back this up - purely based on personal experience and estimation.

> I can't see how the choice of a programming language is even relevant, to be honest. WIth roughly (same order of magnitude) amount of training, a given developer is roughly as efficient (again same order of magnitude). How much % of developers would do a free project I'd they don't have to learn a new language? How much % would do the same project I'd they had to learn a new language? ↺

> > topgun, why does it "have to" attract people that use facebook? > You have to attract people, and given that twitter has 75% of market share ans Facebook has 15%, your user base will probably be someone who before was either in Twitter or Facebook why do you need market share? and who cares what my user base is? Let me put it more concretely: Social Networks matter for the people that want to use them and need them. And their friends of course. If xmpp has 5 people using it and their happy it has achieved its purpose. like activitypub or xmpp or matrix now. As long as people enjoy using them they have fullfilled their purpose. Market share doesnt matter. because by thinking about market share you dont create social networks but social media. Also personally i dont like "user base". People use facebook. its not a "user base".

If one programming language was an order of magnitude more efficient (that is: programmers are at least 10 times as fast to do any given task in that language than any other), then I'd wager that no other programming languages would exist beyond the scope of hobby projects.

projects are supposed to be fun imo above all. (this as a person writing software in guile /scheme :P)

>> You have to attract people, and given that twitter has 75% of market share ans Facebook has 15%, your user base will probably be someone who before was either in Twitter or Facebook > why do you need market share? and who cares what my user base is? > > Let me put it more concretely: > Social Networks matter for the people that want to use them and need them. And their friends of course. If xmpp has 5 people using it and their happy it has achieved its purpose. like activitypub or xmpp or matrix now. As long as people enjoy using them they have fullfilled their purpose. Market share doesnt matter. because by thinking about market share you dont create social networks but social media. > > Also personally i dont like "user base". People use facebook. its not a "user base". Well, I'd like to talk to my friends using XMPP, but they won't install it because they already have WhatsApp or Facebook. So in my case, yes, for me to able to use XMPP, I need it to be able of competing against big tech ↺

(As someone who doesn't use a client with quotation support, the usage of quoted quotes of large texts kind of annoys me)

topgun, xmpp can never compete with big tech. so you are going to wait a long time if you are waiting for that. matrix tried and is already almost bankrupt because of that with millions more

xmpp has benefits over big tech. but its not in the budget or "competition"

> topgun, xmpp can never compete with big tech. so you are going to wait a long time if you are waiting for that. > matrix tried and is already almost bankrupt because of that with millions more Hmm thanks for giving me realistic expectations, that I'll keep having to use WhatsApp ↺

> xmpp has benefits over big tech. but its not in the budget or "competition" It's in what? ↺

i didnt say that :) i said that for people to use xmpp you cant convince them by having the better client or the most stickers. or being on top of the latest fad like ai

> > xmpp has benefits over big tech. but its not in the budget or "competition" > It's in what? privacy, democracy, community governed, no companies, developed collectively. its some of the things that big tech can never copy. its harder to convince people with these i know. but personally i dont want to move people to xmpp by just having them be "consumers". because then they will leave to the next shiny thing. like it already happened with matrix. (which now has its own problems so people arn jumping ship again)

> i didnt say that :) i said that for people to use xmpp you cant convince them by having the better client or the most stickers. or being on top of the latest fad like ai So how can you convince them? This stuff works by network effect. They will probably only use if lots of people are using anyway. And that's why I sent the article ↺

we need a long term strategy. that will teach people to pick tech consiously. instead of mindlessly going for whatever you tell them to

and i realize this goes against snikket and jj probably. but opinions and all that :P

i agree it works by network effect. but i would add the other stuff too in there

topgun: for instance discoverability and avoiding user overload, I like the sound of this trick: https://github.com/benbusby/farside > A user navigates to /nitter and is redirected to nitter.net. The next user to request /nitter will be guaranteed to not be directed to nitter.net, and will instead be redirected to a separate (random) working instance. That instance will now take the place of nitter.net as the "reserved" instance, and nitter.net will be returned to the list of available Nitter instances.

> we need a long term strategy. that will teach people to pick tech consiously. instead of mindlessly going for whatever you tell them to On that porter 5 forces model, one Thing that impact the usage of a product is the barries to entry, switching costs. A barriers to to entry for people running servers: - cost of running instance - ease of usage - how easy it is to make the server discoverable by others Barries to People developing clients/servers: - effort to learn the programming language - organization of the project Barriers to user: - how easy it is to crate an account - how easy it is to find relevant chat groups - how easy it is to find people they already know ↺

> topgun: for instance discoverability and avoiding user overload, I like the sound of this trick: https://github.com/benbusby/farside >> A user navigates to /nitter and is redirected to nitter.net. The next user to request /nitter will be guaranteed to not be directed to nitter.net, and will instead be redirected to a separate (random) working instance. That instance will now take the place of nitter.net as the "reserved" instance, and nitter.net will be returned to the list of available Nitter instances. I use farside.link. it's great!!! ↺

as i said this seems to be looking it from a capitalistic point of view that doesn't make any sense for community projects

and some of them are directly harmful to building any social network to begin with

> as i said this seems to be looking it from a capitalistic point of view that doesn't make any sense for community projects I don't see anything bad In making the life of volunteer developers and server runners as well as users easy ↺

i dont either :)

> does anyone have an idea for something interesting they'd like to do with the domain xmpp.beer ? otherwise I'll probably just let it expire Yeah I can make something up

> (As someone who doesn't use a client with quotation support, the usage of quoted quotes of large texts kind of annoys me) Guus: have you considered fixing your client :P ↺

moparisthebest: yes. I only have to find ~1 year of time to do so.

To add quote support? You don't have to give me estimates like that, I'm not your boss lol

Well, it's a redesign of the chat layout (as the current layout doesn't align multiple lines of text to the left hand side, making adding | symbols in front of a quote ... difficult. At that stage, we might as well redesign the entire UI ... at that stage, we might as well ... you'll get my drift.

as per summit discussion: https://gultsch.de/files/xep-mds.html

Nice

🎉

🥳

Another topic: Im odering new stickers. this is their cutting proposal. Shall I print it?

https://jabbers.one:5281/file_share/NSOXSSCXv9RtURxTv7ljYx-O/zb2rhgv2e15mraG5GGqHozTkFhDKQn1QK1wnWZqzi1upsZSrN.jpg

please vote 👍👎

Daniel: I suppose we can't re-use the chat marker displayed element because it is specced to always be stanza id and never mam id?

singpolyma: yes

emus: is the different from what we had before?

the edges yes ↺

https://jabbers.one:5281/file_share/1V8D-t23HbAljgedtqs0Et18/zb2rhifFn4EQY3PTGBN3BEysRsSxLBL3dyXcSheTYecwAn7pe.jpg

I can ask to make it thicker, but I got told its not nice

I guess I like the old one a bit better?

Yeah I guess me too, but it doesn't really matter

How does server injected stanza-id work in 1:1?

Also should there be any assumption mise on whether sent messages have been displayed? I'm thinking maybe this is left implémentation defined?

Maybe it should be specified though that it's up to the sending client? And not those receiving carbons or MAM to update the state right away? ✏

Ah you do already :)

I also liked the old one better

> For all other chats the stanz-id attribute Typo, stanza

Is "home server" a term we actually use now?

Dunno. Do you have a different suggestion?

I use the term "instance". My instance, your instance, that guys instance.

Daniel: server?

I’m fine with „home server” term, but I’ve always used „server of the user” or „server hosting user account"

"a user" belongs to a unique server right?

> user’s server. You actually use this later on :)

Watch as some madman makes servers that split the existence of the user off onto several instances.

Ah, you see. There is me using the term instance again.

Dear all, as you might have notices we have been accepted as GSoC organisation 2024 again! 🎉️ https://summerofcode.withgoogle.com/programs/2024/organizations/xmpp-standards-foundation Now its time to spread the word and the mentors needs to sign up to the platform. It's also important now to take action on advertising in your communities. We have a list of interesting projects open: https://wiki.xmpp.org/web/Google_Summer_of_Code_2024#Project_Ideas Looking forward, XSF GSoC Org Admin 2024

Unfortunately, we are not able to participate the Outreachy program this year as Umbrella organisations are not allowed and I was informed one day before the deadline 😕 Sorry, my bad also. I will try next year again.

emus: thanks for the work and congrats, too bad for outreachy, but at least we will know for next year!

👍️