This release saw contributions from some new people, which is great to see.
In this release you can now follow other people, not just join communities. When you follow someone their posts will show up in your ‘Subscribed’ feed, regardless of whether you have joined the community it was posted in or not.
You can also follow Mastodon+ accounts. Getting Mastodon integration working really well is going to be a long journey but this release gets us headed in that direction. Eventually I hope that this integration will provide a steady source of timely content to cross-post into threadiverse communities as well as being a peer citizen in the Mastodon-flavored parts of the fediverse.
New:
- You can follow other users and their posts will always show up in your ‘Subscribed’ feed, regardless of the community they were posted into. Following Mastodon users works also but should be considered as very experimental. This will improve soon.
In the future you will be able to upload a list of people that you exported from your mastodon profile and follow them in one go. And add people that your instance doesn’t already know about - but not yet. These features will be rolled out in future versions once the dust settles on this version.
-
Mastodon posts have a new design that is distinct from threadiverse posts - there is no post title and larger author profile pic.
-
Communities can be marked as being a favorite, which makes them more prominent in the ‘Communities’ menu.
-
Clicking anywhere in a post teaser takes you to the teaser - you don’t need to aim for the title
-
Dramatically improved page load times on Chrome-based browsers (Speculation Rules API).
-
Little spinner thing while a vote is being cast
-
Updated translations
-
Preview markdown when editing a wiki page
-
RSS feeds on the home page, including a feed of posts from your subscribed communities
-
Topic selection during onboarding looks way nicer now.
-
Accessibility improvements to headers.
For Developers:
-
Massive CSS reorganisation by travis-jeans. Lots more usage of CSS variables. Improves maintenance and accessibility.
-
API: Users can now log out
-
API: Follow and unfollow users
-
API: video file upload works now
For Instance Admins:
-
Communities where every moderator is a bot or inactive are flagged as un-moderated. Reports about posts in un-moderated local communities go to admins.
-
Warnings on domains can be 3 types instead of only a warning - the other two are ‘helpful context’ and ‘recommendation’
-
Send a DM to a user asking them if they are a bot, with 1 click. If they do not respond their account is automatically flagged as a bot
-
Email bounce inbox can now be accessed through POP3, not just IMAP
-
Tokens used by mobile apps can be set to expire, using the JWT_EXPIRY_DAYS environment variable
-
More accurate tracking of when a user was last active. This will significantly increase your MAU stats.
-
Strict allowlist mode to reduce attack surface exposed by federation
-
/admin/federation form split up into 3 separate pages rather than one long confusing page.
-
Option to disable DM sending for individual users if they lack the self-control to use it responsibly.
-
compose.yaml has changed a bit: For most containers the build target is now "runtime’ instead of ‘builder’.
For example
celery:
build:
context: .
target: runtime
-
Topics can be associated with countries, for more targeted onboarding suggestions based on the viewer’s location.
-
To reduce server load and network traffic, a maximum number of votes that can be cast each day has been set at a level that will only affect the top 2% of voters, leaving the bottom 98% unaffected. This quota can be altered with an environment variable. View a profile to see how much of their daily quota has been used.
To upgrade from 1.6.x
git pull
git checkout v1.7.x
At this point you might see an error message about a merge conflict with compose.yaml. To preserve your custom compose.yaml you will need to copy it somewhere else, then git checkout compose.yaml then git pull again (or use git’s stash feature). This time the pull will succeed so after that copy your custom compose.yaml it back, overwriting the one from git.
Then,
./deploy.sh or ./deploy-docker.sh
If you had to do the compose.yaml fix up earlier then you might want to compare what you have with https://codeberg.org/rimu/pyfedi/src/branch/main/compose.yaml and manually copy and paste some improvements in particular the target: part of the db container.
Donations
PieFed is free and open-source software while operating without any advertising, monetization, or reliance on venture capital. Your donations are vital in supporting the PieFed development effort, allowing us to expand and enhance PieFed with new features.
I cannot recall the last time that I downvoted something - I do it exceedingly rarely, and only as you said to indicate irrelevance; although technically for rule-breaking content, shouldn’t reporting+removal be more of a consideration? (for e.g. disinformation removal would seem preferable, whereas for misinformation mere downvoting may be sufficient)
To give a specific example, I was looking recently at the posts in !lotrmemes@piefed.social and noticed how over the last week several posts have <10-20 upvotes (I can’t easily tell how downvotes factor into this), and keep in mind this community has existed for 11 months already. New posters will stop bothering to contribute content unless they receive positive feedback, which puts the burden of content generation upon the most prolific posters who are able to specialize in doing so, having optimized their procedures. This leads not only to lesser activity overall, but a heavily concentrated set of posters and seems naively to me like it would trend towards encouraging people to lurk more rather than engage, both by sending the subtle message that voting is not desirable while also causing content to receive fewer upvotes than it otherwise would.
If true, then this affects the entire network - even Lemmy + Mbin & now Mastodon (+nodeBB, etc.). Of course I could be very wrong: perhaps votes being a limited resource could encourage people to vote MORE than they otherwise would, or to vote differently, like with greater intentionality (rather than merely smashing a “like” button, sending a signal that it would be desirable to see more of similar content in the future in that community). Though… I would have thought the opposite actually? In the past, in communities that I help moderate or want to see grow, I have upvoted every single comment within them, not to signal “agreement” in that case but “this comment is welcome, it is good for it to have been added here”. If it makes people feel good to feel welcomed, then why not upvote MORE?
But perhaps this is a different model for upvoting - a more "feels good’ social media one than a news aggregation one? I actually have quite a few thoughts about that!! One being that you can’t really trust votes from unvetted accounts in the first place, so for certain situations where votes actually matter beyond merely sorting within the various Feeds (Subscribed, Local, All, Topics, etc.) then we would really need a whole new type of vote class than a mere “standard upvote”, which can be botted and also influenced by drive-by non-community members who browse there by All. I am so enthused to see Piefed moving in what I consider great directions there!! Allowing a community to restrict voting to solely members helps with both of those aforementioned effects (nothing will ever entirely be a defense, but by placing a barrier it does help, and the barrier could always be raised further if the need arises, e.g. introducing a delay of one week before someone who signals intent to join a community before they can begin voting in it, to prevent drive-by All browsers from joining for the explicit purpose of influencing a specific vote than they would lack the proper context for anyway).
So here I just meant the standard, social media-esque upvote, of posts and comments. Though without numbers or the context to compare them, I am just thinking in terms of general principles here.
It helps that this will be able to be set by an environmental variable, so that different instances can tune it towards however they want, although then my concern turns towards transparency - how could someone know what this variable is set to, as they consider which instance they want to join? Or if the thought is that this number is SO exceedingly high that a normal human would not likely reach it, then… well no, because it’s already been stated that this definitively overlaps with the range of human interactions right now, and will most definitely impact some people - 2% of them, to be precise.
Did I miss something - is this not an instance setting but something that the users can opt-out of by changing their personal account settings, hence is a measure to help curb someone’s social media addictive behaviors?
Thank you for your feedback.
We have rate limits (a quota by another name) on posting and commenting. Why not on votes? Casting a vote causes as much network activity as posting a comment.
On everyone’s profile is a bar showing how much of their quota has been used. I’ll be monitoring this to see what happens. If it’s cramping everyone’s style too much then it’ll be removed.
I wouldn’t mind a per-minute rate limitation of anything - posts, comments, even voted, though a per-day one feels weird. Without that immediate feedback it is something that someone impacted by it would need to expend effort towards tracking, as if those additional votes were arguably somehow “bad” (otherwise why limit them at all? I mean for a normal human, even the top 2% of us, but obviously a superhuman bot level of activity would be something else entirely - and in fact in that case, could rather be handled by banning the account rather than merely limiting its contributions?) rather than contributions made to enliven a social media network that by most accounts (e.g. posts to various Fediverse communities) already lies more on the spectrum towards the fairly dead side, rather than being too active.
That said, I recognize that this is an incomplete, even arguably naive perspective since I am unaware of the technical hurdles being faced. Which raises difficult questions: if Lemmy is able to handle not only the current scale of tens of thousands of users but another order of magnitude or so beyond that, then is Piefed unable to do so? (Again, I do not know the answer here nor would be much help in finding one, just stating the questions that will arise from this, especially from the ML crowd looking for any excuse to criticize what they stubbornly see as solely a competitor rather than ally against corporate interests.)
A per-minute rate limit wouldn’t work here because people go through bursts of activity - I might do 50 votes in 3 minutes, then none for 10 hours. If the rate limit was 50 votes per 3 minutes then people would still be able to cast thousands of votes per day.
PieFed is able to do so, and has been doing so.
But I don’t think think it should have to and I don’t think we should let a handful of people have the amount of influence they have been having. Did you see the graphs I shared in the matrix room?
No I haven’t been to the Matrix room.
I don’t deny that a limit is desirable in the code - even if some instance admins prefer to tune the level up- or down-wards, or turn it off entirely, having choices is unquestionably a good thing?
What I question is why the default limit was chosen specifically to overlap with what some current humans are doing. This is not a bot-detection metric, nor a limiter on downvoting to influence algorithmic sorting, this seems a limiter rather on total votes, both up and down… which is a form of “contribution”.
Otherwise, “people would still be able to
castmake thousands ofvotescontributions per day.”Personally, I wish people would upvote MORE often - it makes the place feel more active and encourages people to contribute more in the other forms, both posting and commenting. e.g. a comment that receives zero upvotes or downvotes “feels like” - to me and I presume everyone - a wasted effort, as if it was never even seen to begin with. (If a tree falls in a forest and noone is around to hear it, would it make a sound?) I’ve even specifically tried to make it a point to upvote more often myself, and now I wonder if I am unusual in that, perhaps a top 2% voter even, though I would have offered these comment-form contributions regardless.
It seems to me to boil down to: are the various forms of contributions “bad”? Posts are read by many and for the most part seem unquestionably a good thing, ignoring the content of such of course. Comments also offer a unique added benefit from a mere news article that people crave to experience - otherwise they can use an RSS reader, but that’s not what the Fediverse is. Here we are primarily social. Votes, outside of specific scenarios i.e. a questionaire poll (highly constrained and doesn’t seem all that relevant here? basically those are niche, corner cases), make a place feel more “alive” and “vibrant”, and upvotes in particular make the person who received them feel WELCOMED to this community space.
Otherwise why bother posting, if people did not bother to read what they offered? A lot of Lady Butterfly’s uplifting content receives barely any comments at all, hence I try to make one, whenever I can. Likewise, the tenforward community is a large part of what made people enjoy coming to Lemmy or Kbin (then later Mbin, and finally Piefed), during the main Rexodus, and a large part of what made it special was not only the true OC uploaded there, but the welcoming atmosphere - e.g. every comment made would somehow receive at least four upvotes. Comments made in other communities might receive none, perhaps one from the person responded to, but there people would actually SEE what you offered - be it written text, or a constructed graphic, maybe just a simple rehashed image.
The vast majority of the Threadiverse is dead on arrival. I’ve made posts that have fully zero comments or upvotes, to try to help a community flourish, but eventually gave up and just stopped doing so. If a year or so later after making it, still nobody had seen it, or indicated approval, then why continue to put forth efforts to make further such endeavors?
With the Threadiverse dying off as it currently is - trending downwards from >55k at our peak to currently a mere 35k - I don’t think that acting to CONSTRAIN any types of contributions is helping us to grow? Maybe if we were at 500k and worried about too many of such, then constraints would make more sense.
As for “unfairness”, I don’t think that logically holds up to scrutiny? Don’t we all have the identical ability to vote? Or post or comment for that matter? If people suck as ThePicardManeuver and PugJesus and LadyButterfly happen to do the heaviest lifting in terms of providing us all content to enjoy… then I don’t see a reason to limit their contributions? (Rather, good on them to do what really we all should, under ideal circumstances, but since circumstances are not ideal, then at least they are seeing to it that the work gets done!!) Nothing in life is ever going to be “equal” - e.g. men != women, though I would hope that things would be “fair”, as in everyone has the same access to do things as everyone else. Except that now, the most prolific voters among us no longer have the same access to vote as they wish, which the bottom 98% still do?
All that brings me back to: why are vote contributions suddenly considered “bad” - as in, shouldn’t people be upvoting more than we all currently do? This is the part that I still do not understand - I legit was thinking of it as an outright “virtue”!:-)
And worst of all, if people start with their subscribed content and then shift to more niche stuff, e.g. a lot of people say they do the former then switch to browse like All or perhaps New, then the top voters would run out of votes with the more popular stuff than doesn’t need it anymore, then not have any to deliver for the content that has none to few upvotes? If the top voters aren’t keeping the network active… then who will?
Here’s the data - https://piefed.social/c/fediverse@piefed.social/p/2190040/who-decides-what-you-see-on-the-fediverse-a-look-at-voting-patterns
I hope that more selective voting will surface higher quality content better than people spamming votes everywhere which averages out to nothing. Higher quality content being surfaced will result in more people voting, more people sticking around longer and more votes cast overall.
Personally, I’ve been voting MORE in the last couple of days, now that I’m aware how little of my quota I’ve been using in the past. Chances are I’ll return to my baseline behavior once the novelty wears off…
One thing this analysis ignores is that the power of an upvote RECEIVED (not GIVEN) diminished rapidly with each additional upvote. The most powerful vote is the very first one, changing a post or comment’s ratio from 1 to either DOUBLE that of 2, further improving its discoverability (far more than any subsequent upvote) or crossing into the null boundary at 0, altering its trajectory to actively appear lower in the sorting than it otherwise would have (a phase shift then, one of the most powerful effect there can be upon a content item). But the second upvote only adds +50%, and the third only +33%, and adding +1 to +1000 does little to nothing at all.
And yet to a content creator, seeing +100 upvotes means A LOT more than seeing merely +10 or +20 or even +50 - the former feels much more… “worthwhile”, out of proportion to its numeric effect.
Most votes across the Threadiverse are like background characters in a play - they don’t really do much, individually, but they collectively help sell the overall story effect of a larger universe.
So if you were really concerned about people having an outsized effect, you would need to count the POWER of each vote. Which has little to nothing whatsoever to do with their numerical superiority: someone browsing by All/New for just 10 minutes can impact a community far more greatly than someone else offering >1k upvotes spread over the course of an entire day. Though mind you, I still don’t see someone who does the former as having done anything BAD, yet I do feel that throttling the latter really will lead to an overall cooling effect upon the communities, as people must now consider an extra blocker to adding their thoughts.
It was an easy implementation of a thought, but it has far-reaching ramifications to alter the foundational nature of social media. If there were theory-crafting to back it up then I’d be all for it, but my own initial thoughts lead me to be against it. I hope sharing my reasoning has been helpful. :-)
Very helpful, thanks. I appreciate the effort you put in to explaining your position it gave me a lot to think about.
As with all software, nothing is written in stone and we can change it if it doesn’t work out or has unintended side-effects, etc.
This is what I half-suspected as the reason for why some feel so attacked by this change. What is disinformation is quite ephemeral in this post modern era and in the end it means that you’re downvoting things you disagree with. And unfortunately that usually means things that don’t match people’s pre-existing beliefs.
Hrm, personally I just unsubscribe or at worst even block a community or account rather than feel the need to downvote every single thing across the entire Internet. Though I suppose downvotes could play a role in combating it, it seems preferable to me to disengage entirely - e.g. when Hexbears refuse to play by the rules, even when pleaded to by their own admins (yet with zero enforcement enacted on their part), then defederate entirely, or if the instance admin refuses to - which is worth consideration as to whether you want to remain on such an instance - then ban the instance yourself personally.
It is called the Intolerance Paradox. There is no scenario in which feeding the trolls will ever work out well in the end. Appeasement never works.
All that said, I would have no problem with a quota specifically for downvotes, if the rationale is that people sometimes misuse them but that such a quota would help steer them towards a better understanding of alternative actions they could take to enjoy their time on the Fediverse without such tension.