Given you didn't even ask me for information about my estimate, and simply said what you said — it feels like you aren't having a discussion in good faith.

Based on previous surveys I've run for similar types of things, probably something like 0.001%.

a 0.001℅ response rate means 1 person out of 100 000. citing made up statistics to support an argument is a quick way to loose credibility

Then I don't understand why you replied to my thread.

There are parts of what you wrote that don't make any sense.

(Which is why I was trying to get you to think-through the implications of what you wrote.)

I've said my piece. I have no interest in debate.

When I tried to understand this 0.001% before — to me it seemed like Mastodon was the wrong tool for them.

To me it seemed like they should instead be using something more like a private Signal DM group.

...

One thing to perhaps be aware of is — if a person uses (any) Fediverse software (as a regular user), then information will be recorded/collected about them.

This is not just how Fediverse software works, but how almost all software with user accounts works.

If this 0.001% wants to avoid data being collected on them, then they should avoid most software in existence.

Don't use e-mail. Don't use chat apps. Don't use Windows or MacOS. Don't use Android or iOS phones. Etc.

Having said that — I would like to understand the concern more, as perhaps there is some way I could help them (without negatively impacting everyone else).

...

Regarding:

"""
However, people on here tend to get *very* twitchy when somebody starts collecting data about them.
"""

Let's be clear that these "people" are a very, very small fraction of Fediverse users. Based on previous surveys I've run for similar types of things, probably something like 0.001%.

Which again doesn't make the concern invalid.

But, I think it is important to point out that this doesn't seem to be a concern of the vast majority of people on the Fediverse.

…

Given that, the onus is on you as the developer to demonstrate that you are not endangering people by collecting this data. You need to take this very seriously and both ensure and demonstrate that you've done the work.

I'm not an expert in this beyond a general understanding of how collected data can be used in unintended ways, so I don't have any real concrete threat models I could provide. There are probably others here who could.

However, people on here tend to get *very* twitchy when somebody starts collecting data about them. This is often for good reason--there are people whose literal real-world safety depends on certain things being hard to find.

(ctd)

Could you talk more about how you feel follow date-times would lead to harassment, so I can better understand the concern?

.

AFAICT (when I have run surveys about this type of thing), there is a very, very small fraction of the Fediverse (something like 0.001%) that feel the way your suggested.

While it doesn't make the concern invalid— Mastodon and other public-first social-media software are probably the wrong tool for them.

Something more like a private Signal group would probably better fit what the use-case they are hoping for.

.

Storing info about when a DB record is created is common

Probably all Fediverse software record this — because almost all software stores this information

...

The latter would be fine for the use-cases I had in mind

.

Tangential to the main discussion, I'd like to point out that folks here get *very* nervous when this kind of data gets collected and you need to show that you've very carefully thought through how to keep from enabling harassment.

Having that logic implemented in every client feels absurd.

But having logic to compute the time for "follower since" from an outbox/inbox in every client that wants to display it seems less absurd (to me). I can understand if you don't.

BTW, I'm not saying that having that information in the followers/following collection is not worth it, but that the change required to have that happen, moves you away from ActivityPub into maybe ActivityPub v2, or something else all together.

The information in the followers and followings collections is already in the outbox.

Why bother with the followers and followings collection if they are not the source-of-truth, and we could just iterate through the outbox instead?

To me they (the followers and followings collections) seem like a "projection" of (some of) the data in the outbox — to make things more tractable and developer-friendly.

I think we all need to start thinking of the social API of ActivityPub, not as a graphql correspondent, where API callers get all their information in one request, but as one (of many) source of truth, on top of which we layer some other abstraction that offers the elements that we want to show.

So it's less about presenting users the followers collections directly, but more about extracting meaningful information from that collection and collate it with other elements form their inbox/outbox.

I meant to use "url" rather than "id" in that sample code I wrote.

But, your broader point is still correct. I do more or less want a collection of Follow and Accept[Follow] activities.