Skip to main content


"It's OK if the number of secondary fediverse services¹ is much less than the number of accounts or the number of primary account services², as long as it's much more than one."

#Poll #EvanPoll

¹ for example: people search, text search, tag search, bridges, groups, reposters
² for example: Mastodon, Friendica, Pixelfed, Pleroma instances

  • Strong agree (43%, 51 votes)
  • Qualified agree (42%, 50 votes)
  • Qualified disagree (9%, 11 votes)
  • Strong disagree (5%, 6 votes)
118 voters. Poll end: 1 year ago

Evan Prodromou reshared this.

I think freely indexing posts on the fediverse is maybe not a good thing!
@shoofle totally unrelated to this poll!
@shoofle well, let's say, only tangentially related.
oh i thought that's what text search was!
number of individual instances?
@edythemighty what's your question? Could you be clearer?
nevermind, your followup post cleared that up. Also, hello back!
I'll take for instance the number of services that help you find your Twitter social graph on the fediverse. I know of Debirdify, Fedifinder, and Movetodon.

There are probably a few more I don't know of, but I doubt there are more than about 10 publicly available.

There are about 10-20K primary account servers on the fediverse.

There are close to 10M accounts on the fediverse.

So roughly 1::1000::1000000 proportions there.
qualified agree in that I can't really imagine how else federation would look? Do you have any other "disagree" viewpoints in mind?
need an option:

WTF I have no clue but want to learn, ;)
@Ronkjeffries

https://prodromou.pub/@evan/109598310242408830
It seems like it would be totally bonkers if there were more secondary services than individual accounts on the Social Web, or even an equal number. Does that make me a Strong Agree? Or have I missed a key distinction?
@splicer@Ronkjeffries nope, you sound like a strong agree.

One way that the numbers would be equal is if all the services were only implemented in the primary account services.
Same goes for Kernels and Operating Systems.
If you're opposed to text/people/tag search of any kind, even those that are opt-in and respect noindex tags, please just ignore those kinds of services for this poll, and consider other kinds of services.

If you're opposed to any secondary services of any kind, well, I think you're strong disagree, since 0 is not greater than 1.
@brion OK!

What does that mean for services like debirdify or fedifinder or Movetodon?
@brion I don't think all secondary services have to expose people's personal data. An RSS reposter, shared blocklist, or a social game would be examples that don't.
I don't think I understand the meaning of the question. The phrasing is extremely convoluted.
Your explanation also made less understandable the results, since disagree would include people strongly want secondary services AND people strongly disagree with them. 🤷
@grin that's OK! Feel free to skip this question if it doesn't make sense to you.
@grin
Thank you, I will utilise this offer. 😉

Regardless, I am for having as much auxilary services as possible, I support search, directories, hashtag following and search, whatevers. But that poll maybe isn't about that at all.
@grin it kind of is.

Basically, if you imagine a service type like a shared blocklist, how many instances of that service are necessary for it to be a healthy part of the fediverse?

Is one enough?

Or one per Mastodon instance?

Or one per person with an account?

Or somewhere in between?

The poll proposes one heuristic (1 << N << number of Mastodon instances) and asks whether that's good enough.
@grin
NB, this explanation is far better than the ones I've seen yet (including our OOB chat). And has absolutely no resemblance to what little clarity I was taking from the original question.

No need to reply, it can wait until after the poll closes, but just thought I'd share this.

@grin
@grin
@dredmorbius
indeed, this is clear explanation. Still the results sum the "I want none" and "I want much more" opposite cases. Sometimes this agree/disagree options aren't working well.
@dredmorbius yes, provided I would need the results! 😉
I am having trouble parsing. It's ok if the quantity of index/search services (eg 17) is less than the quantity of author-accounts (eg 30M) or the quantity of publishing-instances (eg 1k) as long as ... one of those is more than 1?
I'm not understanding the question here, and feel as if I'm missing critical context.
@dredmorbius does this help?

https://prodromou.pub/@evan/109598310242408830
A little bit?

"Some service accounts are OK, so long as they exist at all" being my attempted rephrase.

I'm ... still not sure what issue / problem this is addressing, if any.
@dredmorbius great! Feel free to skip the question and just go live your life. There will be another poll tomorrow!
My life goes on.

I figure you're aiming for something here, and in this case missing. But likewise.
@dredmorbius maybe! I'm talking about decentralisation theory and practice, and I'm not sure how important that is to you. It's really a specialist poll.
@dredmorbius I'll be happy to explain it further after the poll is over.

I try not to go too deep on the explanations of my polls, because I don't want to sway people one way or another.
Sure, I get that and run my own polls that way myself.

Talk after is good.
@dredmorbius I would say I am also definitely interested in the topic but didn't find the poll clear enough to vote.
@ocdtrekkie@dredmorbius sorry to hear that! I'll give a fuller explanation after the poll is over.
Although equality of the first two categories is probably something like siloed accounts, it can be interesting to figuring out what could be a world with more services than accounts... Isn't something like merchants vs paiements services? Or more sadly, OpenID consumers versus OpenID providers?
@nus this *was* me taking it up with the founder 😛 and it turns out that there are enough bad actors on the fediverse that it's scrambling for more safety features, which some people refuse to take seriously, which is why i'm talking about it when someone asks for opinions. if you meant bring it up with gargron, good luck getting him to listen to anyone's suggestions about making this a technology which can be used to build healthy spaces!!
@nus the fact that privacy is easy to violate in theory doesn't mean we should like making it easier.
I don't get this question. Can you repeat please? 🤔
@amicalmanthttps://prodromou.pub/@evan/109600185530138387
So, this was my least enjoyable poll so far.

We don't have great terminology on the fediverse for account services and other kinds of services. So a lot of people didn't get the question.

But also, the question really is, how many of a particular kind of service do we need for a healthy fediverse?

Four options might be:

- 0
- 1
- much more than 1 and much less than the number of Mastodon instances
- the number of Mastodon instances

Most folks don't think about this kind of question, though.
I did not see the original poll but my answer is much more than one and much less than the number of ~Mastodon instances
It's that 3rd one, IMHO, though it may depend on the service.

It would be awkward if every instance were running different software. But having only one dominate is bad.
True. I honestly didn't get that question. And I'd be happy to help if I could.

(Ordinary old guy here. Used to do desk jobs. Dealt with legal wording, and occasional clever stuff. Could wrangle Excel; but Word formatting of stats tables nearly made me cry.)

Following you via something boosted by @aendra
Suggestions for further polls:
* Which secondary fediverse services do you use?
* Which secondary fediverse services are you missing?
I still have no idea what this is about. I didn't say anything the first time tho.

Maybe try it another way - why is the question important?
@Shan_Ye the question is important in how we think about services working on the fediverse.

When someone launches a good¹ full-text search service for the fediverse, that meets our community's requirements, we might ask, is that enough? Do we need more? How many more?

We might also design those services differently if we think there may need to be 10, 1000, or 1000000 of them.

If none of those seem important to you, maybe you're not the audience for the poll.
My least enjoyable poll ever was this one on Twitter, where I had lots of people telling me you can't learn from books.

https://mobile.twitter.com/evanpro/status/1574057757442920454
I remember that poll!!!
I did terribly at school. It wasn’t an environment I felt good in, and didn’t learn well in groups. In my late teens I suddenly realised I could learn so much more from books, so that’s what I did. And I didn’t feel so dumb anymore, which is what school made me feel.

But anyway, doesn’t the answer to that question very much depend on what domain of expertise you are talking about?
Ai yi yi. The list is long, but I learned to knit from books, and here I am, countless sweaters, hats, socks, and shawls later:
the yokes of two matching ski sweaters
@rrix so, I think for a lot of types of services, there is a benefit to having a small number of instances of that type of service.

If those instances need to share data, like for a shared blocklist service, they can share with a limited number of well-known peers. This is harder with thousands of instances.
@rrix In terms of numbers of implementations, having people split their effort across a lot of implementations means getting a lot of half-baked cakes instead of a few fully baked ones.