Expand the character limit for posts (toots) in our mastodon instance

Hi all, I posted a link to this from the social.coop admin account. For my part, I've grown accustomed to the limit and appreciate it for the reasons stated above. I wonder if the solution doesn't lie in adding tools/platforms, so that social.coop members also get a writefreely account (like we currently get a Meet.Coop account). Social.coop already is a bundle of platforms (Mastodon, Matrix, Git, Open Collective, Loomio...). I'd also like to see us collaborate on this with existing efforts like Co-op Cloud or MayFirst. (But I don't oppose the proposal.)

I follow some people who use long posts occasionally and actually prefer reading them instead of a fragmented chain. I don't miss the large amounts of comment threads on Twitter at all. I also think that having to be brief sometimes leads to cutting out things that might better explain the context of what you are trying to say, which can lead to less constructive discussion.

From an admin perspective is it really the case that you need to modify the Mastodon source to get longer posts, and that it's not an explicit config change?

https://github.com/mastodon/mastodon/issues/12265

While I'm sure it could be scripted in the deploy/update this might not be the best from a maintenance perspective.

that's a good point, when I joined on the tech team we had various customizations to mastodon, and we ripped them out as it is indeed hard to maintain... we only just manage to have it up and running at all :)

... there are indeed multiple places where it is hardcoded :/

• app/javascript/mastodon/features/compose/components/compose_form.js (twice, and also probably requires doing a webpack rebuild of the assets)

• app/validators/status_length_validator.rb (once)

Eugen (main developer of mastodon) seems very unsupportive of making this configurable... which is sad.

So, important thing, as well as deciding if we want to increase it, have to work out if we can increase it. This in practise means something for tech team to agree to make changes and document/support them. That's not me, and it's not something I'm personally up for.

Ping @Flancian ? ... I don't know how/where tech team is at now tbh...

Alternatively, switching to a fork that makes it configurable is another option, but that also needs capacity and interest within the tech team. It ain't me babe ;)

@Eduardo Mercovich, thank you for proposing this for us to discuss.

While I can sympathize with wanting a longer field, given the experiences with Twitter's original character maximum, and I understand that by expanding it to 5,000, it would still give people the freedom to be more expressive, my concern is that of those who do utilize this much space, it could wind up crowding out nearly most or all of a screen depending on window resolutions and devices.

It may actually be quite an unpleasant experience to regularly see half a screen or a full screen of one or two people's thoughts, while so many others who are more economical wind up being squeezed out.

If we're to have the ability to post in 5,000 characters at a time, I think it would be best to have a feature so that users could filter messages by a user-defined size (say, let them have the ability to hide messages longer than 1,000 or 3,000 characters or any other number), but I think it's very likely that many users would want to filter out such large messages, and those who were interested in writing so much, might find an inverse proportional interest from readers: the more walls of text one writes, the more messages of theirs that will be hidden.

Given this, perhaps we can negotiate a size in between the current one and the proposed one that is an increase by tenfold?

--

The text above from the at sign to the question mark inclusive is 1,357 characters. Should we make this space one in which people can post more than 3 times that in one toot?

Beyond that though, where social.coop is on other servers' federated timeline, how nice would it be for us to send walls of text there? Might it result in server admins block listing us? I don't know the answers to these questions. So, perhaps some of us should first experiment by spending more time on these servers with 5,000 character maximums and reflecting on how we feel about the experience.

But thank you for bringing this request to us, brother. I'm sure we can figure something out.

mastodon will hide text beyond ~500 chars, with a "read more" link or something like that. the timeline won't get filled with walls of text.... in theory at least, I can't find any long toots to check it...

https://github.com/mastodon/mastodon/pull/8205#issuecomment-416071326 is where that feature got added...

I think it would be best to have a feature so that users could filter messages by a user-defined size

we do not have the capacity to add custom features to mastodon, you could see if people have requested it before, and if not open the topic yourself, in the github issues for the mastodon project --> https://github.com/mastodon/mastodon/issues - I suspect the "read more" folding is sufficient for most people though...

I wonder if a 1000 char limit would be a good trial.... and see how that goes down?

1,000 characters together with the ‘Read more’ foldover feels to me like a reasonable place to be. Although, having the character count go red when 500 words is reached would also be helpful?

At the same time, it could be good to have a Writeas account too. For blogging that is not ‘micro’ aka 1,000 words.

I feel like our scarce WG resources would be better spent elsewhere, such as smoothing out the sign up flow.

Mastodon isn't a blog, it is not intended for long-form content. Folks are welcome to toot links to their longer-form content elsewhere, and many of us already do.

For the occasional longer disquisition, folks figured out threads long ago.

As others have mentioned the micro-blogging format keeps things concise.

I would lean into an existing solution such as the Hometown fork of mastodon, it seems well maintained (up to date), and comes with the discussed custom features, as well as a few others:

• Custom max toot characters to change the default 500 characters limits imposed by Mastodon

• In-line reading of full articles behind a cut for length

• Local only posting (aka posts that never federate)

• Exclusive lists which let you follow accounts without them clogging up your home timeline

• Themes of a pastel variety!

Personally, I would favour switching to Hometown, but leaving the max toot length as-is or a small increase, while the in-inline reading of Articles (from Writefreely, Plume, WordPress) would be a good compromise for most users.

+1

My opinion, for what it's worth, is that even if the 500 character limit is increased it's very unlikely there will be many posts using more than 500 characters. To my mind, arguments from positions such as "I've grown used to the limit" don't hold much water. It is unlikely to affect you, as you can carry on as you were.

Personally, I'm ambivalent. I have a blog and am happy making threads for things longer than 500 characters, but I see this proposal as "good enough for now, and safe enough to try". So it's a vote in favour from me 🙂

I wanted to see what it looks like when mastodon has long texts, but struggled to find an example (I found posts of 1000 chars that were not folding... so I'm not sure the exact mechanism), ... but someone finally shared one with me that did.

If you put @[email protected] in the searchbox in mastodon, then scroll down their public feed, you should see some "Read more" links...

And, in case it got lost, it's not clear we actually have technical capacity to make this change (see my comment here --> https://www.loomio.org/d/0bCHvrTj/expand-the-character-limit-for-posts-toots-in-our-mastodon-instance/15 )

If there was tech team capacity, switching to hometown fork would seem a good idea to me, as suggested by @Django - as maintaining our own patches for increasing post length is not pretty imho.

@Nick Sellen thanks for finding an example. In my own experience I find the Read More links work well to keep the flow of the timeline going, without burdening it with the full text of the post. And we get this already on our own instance, which is nice. The question is do others want to be able compose posts like this, and it seems like the answer is (for some) yes.

But I totally agree that given our technical capacity we should avoid maintaining our own patches. I wonder if maybe we should set up a meeting to talk about Hometown sometime? Or has that already happened before and I missed it?

In theory I would vote yes for this because:

• I find the character limit too short.

• I don't like that it is a technically-defined limit to communication, and a pretty arbitrary one at that. If there's some expectation of optimal message length, I would prefer that to be socially determined, not technologically enforced...

But given it doesn't seem to just be a simple config option, it doesn't bother me enough in practice for it to warrant the technical changes needed.

I don't have a strong opinion about character limits, but the killer feature of the Hometown fork in my humble opinion is user-configurable expiring posts (e.g, you can set all of your own posts to delete after one year). So yes, I also think we should consider the switch so that we can take advantage of all of Hometown's cool features. But I'm new to social.coop, and don't have a deep understanding of the culture around undertaking big projects like this, so I want to be clear that I'm also totally fine with the status quo if it makes life easier for folks.

Dear all.

I'm sorry I had a couple offline days but after seeing what was flowing and being the OP, I preferred to let others express themselves without intervening. Now that it seems that the incoming opinions and views are slowing, I made a thematic analysis on the results for all folxs to interpret.

Each response was tagged with 1 or more synthesis; a new tag if needed and reusing existing tags when possible. The tags were counted per reply and in total and with this information a co-ocurrence matrix was created to visualize which tag appeared with which other one/s (if any).

Both axes of the matrix are the same, except that the vertical were grouped by a big category created from the results (yes, condition, maybe, proposal, no) and the horizontal had the original order. You may observe that the total on the last right column is not the sum of the row because if a tag A appears in 1 answer with other 3 B, C and D, that will sum 4 (A, AB, AC, AD).

Of course since I may be biased even if I'm trying not to be, I am not interpreting anything, just sharing here:

• the tag co-ocurrence matrix as it is for you all to interpret it (both png and pdf versions) and

• the whole dataset (the tagged responses up to now, as an .org plain text file).

Of course, I'll be happy to explain anything that may be needed.

I hope this helps us to move on forward. :)

Warmest regards...

Also, I'd like to thank you all for the warm welcome and great participation. :)

Very nice! Thanks for the tabulating work, @Eduardo Mercovich !

Thanks Bob, it's my pleasure.

I use free tools for this (and almost all) kind of work that I'll happily share if anyone is on these things. :)

Didn't make the time to read through all this. Notice that here on Loomio, it's easy to be verbose and long-winded. I can SO easily get extra verbose myself. (I could have just left out that point and this parenthetical and my message would be okay still!)

If we remove the char limit, people are going to likely get excessively-long posts from me.

My wish would be to have features Mastodon probably doesn't offer yet.

• Individuals setting our own char-limits!

• Short tldr summaries with expand options for longer beyond-limit content

• The ability to opt (be prompted) when passing a limit

• Formal recognition of multi-toot posts as a set (not just casually marked as replies with manually labeled numbers)

And so on (perhaps some of these are already requested or can be requested for the software upstream). I think if people want to post long messages, that might be best done in a different place and then link to the longer post in a toot.

Hi Aaron.

A few of your wishes are present in Hometown (https://github.com/hometown-fork/hometown) along with some more features:

• Custom max toot characters

• Local only posting

• Reading more content types

• In-line reading of full articles behind a cut for length

• Better list (feeds) management

• Better accessibility defaults

• And as it's maintainer says, "Hometown is still 99.999% Mastodon".

Migrating from Mastodon to Hometown is a possibility that appeared in various responses, understanding that it depends on having the time and knowledge in the tech team to make the change. The migration seems very similar to updating Mastodon itself (seehttps://github.com/hometown-fork/hometown/wiki/Initial-migration).

Maybe this information could be useful for us all to decide what to do next. :)

Best...

I don't understand. The main relevant point of "custom max chars" seems to be the topic of this whole discussion. So that must be in regular Mastodon. Hometown doesn't mention giving users the capacity to each set max-chars for themselves on a shared instance, it still looks like just instance-wide fixed max-char setting (per the Hometown wiki)

> I don't understand. The main relevant point of "custom max chars" seems to be the topic of this whole discussion. So that must be in regular Mastodon.

No, Gargron (mastodon lead) opposes to this even if it's quite asked for. And this creates patches and forks on mastodon.

> Hometown doesn't mention giving users the capacity to each set max-chars for themselves on a shared instance, it still looks like just instance-wide fixed max-char setting (per the Hometown wiki)

Yes, it is an instance wide setting. Now it's I that don't understand, you were talking about a per-user setting? If it is so, sorry, I didn't thought of that possibility (I was talking about a instance wide setting). :)

Best...

Since this whole topic was about changing the max chars, the idea of even doing that itself always required switching to that fork at all (regardless of my suggestions)?

No, you can patch/build the code, which I described here --> https://www.loomio.com/d/0bCHvrTj/expand-the-character-limit-for-posts-toots-in-our-mastodon-instance/15

... but I don't think that would be a good idea, and I'm not up for that personally.

It wasn't clear to me/others at the start what it would involve, which is what we've explored here.

Hi Nick.

It wasn't clear to me/others at the start what it would involve, which is what we've explored here.

Neither to me, I never administered a Mastodon instance and so I didn't knew it required a patch. In that case, I support your opinion and wouldn't do that neither. :)

But if Hometown comes with that and other interesting features (personally I like a lot the publish to instance only), my question to the tech team (*1) is: how complicated could be to migrate to Hometown as it's documented in http://github.com/hometown-fork/hometown/wiki/Initial-migration ?

Best...

(*1) Since I'm new I still don't know who you all are, but thanks a lot for your work. 🙌

Yes, I agree, hometown would be a nice option in my view. It needs tech team to have willingness and capacity for that.

I think tech team is more active on matrix than on here, so if you're motivated, I'd suggest hopping into https://matrix.to/#/#socialcoop-tech:matrix.org and raising it there.

Thanks Nick!

I will ask there, great to know the tech channel. :)

Just to keep everyone in sync, @Flancian replied in Matrix:

Thank you for the FR and the suggestion to move to Hometown

I think we could spin up a hometown container to test it?

I've never updated Mastodon myself but it's definitely something we need to get good at anyway :)

I already offered help and will try to keep up with the matrix chat, and let's see if that is doable with our resources.

If it is doable, does anyone has an objection? An objection is not a problem or something negative, but a gift or a present of seeing something that could imply danger or going against our objectives as a Cooperative, so please do go ahead if you see any. :)

If not, we have consent to go on and try it. As @Doug Belshaw said and I totally share it is "good enough for now, and safe enough to try".

Please remember that with this replacement we get:

• custom max toot characters that we can increase the limit up to ~1500 only (3X instead of 10X) as a test as it was suggested by @M. Page-Lieberman - @[email protected]

• a fold as @Nick Sellen and @mike_hales sayed so the reading is not interrupted by default

• more content types

• better list (feeds) management

• user-configurable expiring posts as per @Eamon Caddigan discovery,

• better accessibility defaults that are great for everyone but specially for people with some disabilities

• Local only posting (great for internal discussions in our own Fedi server)

• all the rest exactly as it is because as it's maintainer says, "Hometown is still 99.999% Mastodon".

So, WDYT?

Again, thanks a lot for the enriching discussion.

Best...

Since this is a significant change -- moving to Hometown -- it should probably be a new proposal to be voted on by social.coop members. No harm in setting up a test instance (hosted where?), but this kind of change should be made with more than just no objections, I think. Also, we haven't heard from @David Mynors or @Noah from the tech team...

I've been interested in switching to Hometown for quite some time now; I just never had the spoons to gather consensus on it. I'd love to see it happen and would be glad to help.

Dear @Matt Noyes , thanks again for showing me the way here. :)

In that case, let's wait to hear from the rest of the Tech Team and if the Hometown test is supported, then we will propose it formally again here in Loomio.

Is that OK? :)

Dear all.

The Tech Team meeting was today. The full meeting notes are in https://www.loomio.com/d/UwAeiBgE/tech-meeting-minutes, but in brief:

1. The migration to Hometown seems a good move and it is aligned with our current need to upgrade our Mastodon instance.

2. At this moment, in order to do that, we need to make a brief plan, but most importantly, a backup.

3. Once we have that backup we can try Hometown in a separate, parallel space/url (maybe federating only with ourselves so as not to pollute the Fediverse with our tests).

4. For this trial we may ask not only previous Tech Team members (some documentation remains, but better safe than sorry), but also Darius Kazemi (hometown creator and maintainer).

5. If this trial goes OK after some time we can make a formal proposal for the migration here in Loomio in a new thread that references this one but starts clean, for everyone to make their voices heard.

Please @Flancian , @Akshay and @Ed Summers check if this summary is ok, and thanks a lot for the very comfortable welcome. 😃

As always, thank you all for your time and attention... 😺

Really like the way this is unfolding. Thanks to all who are making it happen.

Hi all!

We've started a draft of a plan to 'upgrade' our Mastodon instance in several possible ways. This can be found currently at anagora.org/mastodon-upgrade (to edit: https://doc.anagora.org/mastodon-upgrade, this should be moved to https://anagora.org/go/twg/wiki once stable.)

Our recommendation is to lease a new dedicated server (see https://www.loomio.com/d/jSvw79qE/we-have-room-in-our-budget-to-do-more-what-should-we-do-, currently running, which seems relevant) and use it to ramp up new TWG members and get backup restores tested (crucial for keeping the instance running safely); then use this newfound independence to try out new Mastodon versions, plugins and forks including likely Hometown.

Thanks a lot @Flancian ! Yes, the fact that we have a bif of money to invest in another server allows us to move forward in a more secure way. How can we help those of us that don't do installation or administration in the tech area?

Again, thanks a lot... 😃

I realize that I'm commenting on this long after this conversation seems to have died down but thought I would add some observations: in addition to here, I have an account on hcommons.social (Humanities Commons) which runs hometown and a 1,000 character limit. My impresison so far is that only minority of posts go beyond the usual 500 characters, but they do make it a seamless environment for more thoughtful pieces, or posts that have lists of bullet point or instruction like materials that easily take someone over 500 and get broken into threads. Something in the 1,000 to 5,000 range running on hometown sounds wonderful and I look forward to seeing how the upgrade plan proceeds.

+1 to trying out hometown

I agree that any decision to use a fork like Hometown should be put up for a vote among all members, with a generous timeline and announcement, given the significant differences in behavior, and the implications for future updates.

Personally I would lean towards staying with the mainline release - both because it has more sustainable funding and development traction behind it, and because I don't think any of Hometown's capabilities really are must-have.

I am skeptical of the utility of high character counts in a micro-blogging context -- it degrades the value of the feed for me. There are other federated blogging platforms like write.as that work well in combination with Mastodon.

All of this is written with much gratitude for the folks working on this stuff, and I'd definitely be game to play with a demo site before any such formal vote!

I'm intrigued by some of Hometown's features, especially local-only posting. But Hometown's philosophy, as laid out in the https://runyourown.social/#keep-it-small document, suggests a cap of 50-100 daily active users. The front page of social.coop says we have 1.54K users, 461 active (in terms of monthly active users). About how many daily active users do we have? That would help me think about how Hometown-compatible we'd be likely to be.

I'm on another instance that has a 2222 character limit. It's nice sometimes to go beyond 500 characters, though I think 5000 is a bit too much, and I'm glad I'm not the only one that's brought up a limit in between those two because for some time the discussion seemed to be either stay at 500 or go up to 5000 when there are other options available.

I'd be for something between 1000 and 1500 myself. In any case I'd love to see us either upgrade our version of Mastodon or switch to a fork with features that will benefit us.

I'm on another instance with a longer character limit, and I honestly don't like it nearly as much. People get rambly. So my preference is to stay shorter.

I feel that ideally we'd default to a good-enough, well-known value (it might be 500, it might be 512 ;)) but let users redefine the default according to their preferences.

Mastodon doesn't support this overriding yet I think, but we could explore a patch or workaround -- that might come with maintenance or implementation cost, though. Other Fediverse software might support this now, or in the future; we could check.

If we can't provide overrides soon, I'd rather we upped the limit by default and let people self-limit. It is the safest default in the sense that it maximizes freedom at no clear (but I'd say not huge) cost.

Note in particular that if the UI limited the number of characters shown by default to 500, there would be no extra cost to the reader if they want to scroll past.

What happened with the idea of trying Hometown that was talked before? Is that still on track once we have the new server? :)

Fedibird https://github.com/fedibird/mastodon offers more features and is used in a pretty big instance, it might be interesting to check it out or create a sister instance. this would also be in line with the "budding" idea of spinning out more cooperative instances.

I would be much more interested in checking my feed if there were a higher character limit.