posted this
Time ago

Hello everyone - it’s been a long time, hasn’t it?

You might have noticed that Waterfall has kind of fallen by the wayside for the last year or so, and it’s time to discuss why. There’s a few main reasons. The first is just a severe bout of depression on my end, but that’s boring so we won’t discuss that (sidenote: It turns out therapy works. 10/10, highly recommended).

The much more pressing issue is the matter of the British Government. For those who’ve been around for a while, you might remember a post a while back about something called the Online Safety Bill. For those who don’t -

Each word here is a separate link about it. In truth I started writing this post months ago, but the evolution of the Bill meant that whenever I had something written, it was already out of date - just like those links probably are at this point. I’ll let you do your own reading, but it’ll be kind of obvious why development stalled - when half of what people do online is about to be criminalised, is there a point to continuing? Luckily (imagine that in air quotes) things are settling with it now, so we can talk about it. Let’s start with something fairly upsetting to say, and I apologise for my French in advance - The Online Safety Bill is absolutely fucking horrifying. It is SO much worse than the posts a year ago about it expected. It’s evolved to mandate jail time for site owners for the actions of their users, have mandatory ID verification for users to prove their age (as well as segregating unverified users from the verified ones), and a bunch of other horrible things. Take a look through CyberLeagle and Ms. Burns’ blogs and you’ll see the worst of it, including the cool and fun “hostage taking” part of things. So, policy-wise, things aren’t great.I’ve been thinking long and hard about things. I firmly believe that smaller, alternative sites like Waterfall need to exist. With that in mind, I don’t want to shut up shop and leave a sign hanging on the door. To try and deal with the problem, I’ve taken the very time consuming, very expensive, option of incorporating in the United States. This means… a lot of different things, which I’ll go into in another post, though I will rattle the tip jar (Patreon and KoFi are in the footer) if you have a spare five bucks, this has cost me a grand and a bit so far.So - that means the project isn’t abandoned. However, there’s some inescapable pitfalls and I need to cast aside the “I made this” defensiveness for a moment to look at the site with a critical eye. Waterfall as it is doesn’t quite work. It started as a Tumblr alternative when they were bringing in bad policies, but… Tumblr didn’t die. This leaves Waterfall filling a niche that doesn’t need to exist. There’s been a handful of discussions taking place in a small part of the community and the general agreement is that it needs to change, at least a bit.

To that end, I’ve put together a quick poll. It goes a little in-depth and has some free-form answers, so please take your time and think about what you want - your answers will be used to shape Waterfall, or, if more people think a totally fresh start is in order, its replacement. We’ll have the opportunity to shape and re-mould the site into something far better than it is now, both as a platform, and on a technology level, so if you have twenty minutes to spare, please do consider filling it in.

There’ll be more coming soon about what’s going on, but a heads up - I’ll be doing some minor upgrades soon, and moving the server. This’ll result in a short downtime in a few days.

But for now - we’re back.

posted this
Time ago

Community Update - Waterfall 1.3, and The Future™

Hello friends, Waterfall 1.3 has now been deployed. In addition, after a month of putting it off and throwing around ideas in the Discord, this post will finally talk about the potential plans for WF 2.0.

Patch Notes


  • Fixed post editing issues.
  • Fixed some error log spam.
  • Fixed video posting giving the wrong message on success.
  • Added a language switcher to user settings.
  • Added Spanish localisation.
  • Added German localisation.
  • General improvements to localisation code.
  • Re-added OpenGraph support, bringing Discord embeds back.
  • Fixed being able to view the blogs of users you've blocked.
  • Fixed features showing you posts of people you've blocked.


  • Removed redundant files, condensed existing ones.
  • Raven doesn't require Sentry now.

WF 2.0 Planning and Discussion

Right now, Waterfall has a bit of an identity crisis on its hands. The site is positioned as a Tumblr alternative, with its only really unique feature being the art theft prevention system, which itself has been adopted by DeviantArt now, and exists in a kneecapped form to begin with due to artists not being able to agree on how it should work.

But it's in the name - Tumblr Alternative. As this implies, if Tumblr fixes their issues (which, given the quality of life improvements made to it since their last buyout, basically boils down to "officially allowing NSFW again" at this point), there's no reason to use Waterfall, or any of the other platforms that positioned themselves as alternatives, with the exceptions of those holdouts who use them specifically because they're not one of the Blue Monopoly Sites.

Because of this, I've advertised Waterfall even less than I used to before - people who've been around know that I don't like active advertising and prefer word of mouth, but I've not even been happy enough with the state of the site to encourage that lately.

So, I think Waterfall needs to become it's own thing, not just a Tumblr alternative. This'll be a long post, so I'll be putting this under a readmore. Everyone is encouraged to read it and share their thoughts, and Patrons will get polls on specific areas of feedback too. But anyway, readmore time.

First it's probably worth going over some of the notes I've taken over the last few years. Some are observations, and some are things I think I've just handled wrong with Waterfall, and should have done differently.


  • Despite being modelled after Tumblr, the majority of the site's users came from Twitter. Probably because sharing stuff there is a horrific experience, on Tumblr it's still moderately decent.
  • The smallest visible change is appreciated by users a lot more than the largest backend changes - making corners slightly rounder was noticed and appreciated more than any other update, including the one where I cut loading times by 80%.
  • Frequent art uploaders interact significantly less with other people than those who don't draw often - as a rule, they only produce their own content, they don't help spread what others have produced unless there's a pre-existing relationship.
  • Many artists use the site as an archive, with no intention of joining the community itself.
  • There's a lot of tribalism based entirely on whether somewhere lets you draw naked children or not. It's weird.


  • Marketing Waterfall as "for artists". See above for one reason, but also, it stopped regular users from signing up due to a miscommunication of what the site was for.
  • Artist-centric philosophy led to artists having too much sway over design and functionality considerations, which caused deadlock on various aspects (how the features system should work, the arguments that led to v3 of the Art Theft system being cancelled), reduced community cohesion, and reduced developer motivation
  • The site was open from the beginning. This is a surprising one, but when you look at the facts, it makes more sense. Having the site be closed drums up hype, because people want in. While it's under development, only a subset of loyal users who are invested in the site see the WIP and broken bits, and they won't mind as much since they actively chose to be there. Having the site be public access from Day One means people saw the broken bits, and never came back. This limited WF's growth in a way that may not have happened if it was restricted access.

With this in mind, it gives a few options, as well as considerations for how to proceed going forward. The first one is obvious, albeit controversial - whatever happens, there needs to be less focus on artists. Even ignoring the CM debacle, they've had too much sway in design while, in terms of what's actually useful to growing the site and community, contributing very little. This isn't to say things that are useful to them are going away - they still play a part in the ecosystem, and I intend for whatever form WF2 takes to have the art theft system (it'll be v3 this time, despite people arguing for the less effective one) and uncompressed uploads. It just means focus will shift to regular users, since they're the ones who actually spread content, and there's not a point to allowing art to be uploaded if there's no audience. There'll still likely be some artist specific features regardless, just because pretty drawings are nice to look at.

For the rest though, we need to figure out how to proceed. There's three major options I can think of.

First, continue as we are, with WF2.0 being the next major update. Given this is changing the entire way the site works, this may be a problem - people joined for a Tumblr-like, to make the site suddenly less of a Tumblr-like might alienate those users. It also doesn't deal with any reputation WF built up in the first few weeks for being a buggy mess.

Second, as above, but rebrand the site. Content stays and site gets a boost from a new name.

Third, however, is to make WF2 be a new site entirely, with a new name. This way, Waterfall co-exists with it so those who want to keep using it can do so, and it can still be worked on alongside the new site since it's open source, and I don't intend to stop hosting it any time soon. It also allows for solving the problem of a reputation, since all the bugs can be solved with a closed audience before gradually opening it up. This'll likely be patrons and prominent members of the community who are invited over. It also means that several technical concerns can be ignored - there's no need to worry about existing data and losing information, or figuring out how to generate any extra information that's needed on posts/blogs/etc for whatever the new format ends up being, outside of writing an importer for existing WF, Tumblr, and Twitter content.

In truth, I'm leaning towards the third option - the rewrite was pretty painful when I had to take into account existing data. To reiterate, going with the third option does not mean Waterfall dies - it will remain up as long as people use it, and I'll still be developing for it, as can members of the community. However, I think we all agree Waterfall needs to become its own thing to survive, so one of these options is the way forward.

As the community, I think it's important that you guys get looped in on this discussion and get to air your preferences, opinions, concerns etc, and argue for your preferred of the three options. Feel free to reblog or comment, or if you're in the Discord, discuss it there - I likely won't respond to everything, but I will read it all while I make my decision. For now though, I'm retraining to be an ESL teacher so I have a better shot of getting off plague island (the UK), and working on some paid projects. This'll take maybe another month or so (a couple weeks at the earliest), at which point I'll be making my final decision on the way forward.

posted this
Time ago

Community Update - Waterfall 1.2

Waterfall 1.2 is out! Here's the patch notes, straight from the repo.


  • Fixed searches being case sensitive
  • Fixed error log spam relating to non-existent blogs
  • Fixed blocking (probably)


  • Improved dynamic message box rendering and added a utility function to do it instead of the very stupid, very convoluted way it was being done before
  • Post buttons now change when pressed to let you know that stuff is happening
  • Pages can now be deleted
  • Re-added mutual notes mode, but renamed it to "notes from people I follow"
  • Re-added OC Only mode


  • All text on the site is ready to be localised. See the GitHub for details on how to get involved.


  • Language checks have been separated out into their own Action workflow
  • Test strategy updated. PHP version support has been expanded
  • Updated Pillow requirement for Raven to mitigate a few CVEs
  • Added PHPStan to force me to write better code and make things less terrible for everyone
  • Code now passes PHPStan Level 0 checks

posted this
Time ago

Patch Notes - Waterfall 1.1.1

Just a small update this time.

- Removed restriction on pronoun selection, they can now be entered freeform rather than being restricted to only the most common three.

posted this
Time ago

Community Update - Waterfall 1.1

Waterfall 1.1

The site's first update since being open sourced! Much of this update is from community contributors, so a special shoutout to @blobcat for doing a lot of the legwork! Patch notes are as follows:


  • Fixed an issue preventing account deletion
  • Logout option has been re-added
  • Two-factor authentication is available again
  • Fixed a security issue with tag searching on blogs


  • Sentry is now optional in development environments
  • Set-up documentation improved
  • Improved Raven Comments

Waterfall 1.2 will follow soon with additional fixes and improvements, and then it's on to designing 2.0.

For any features or requests, please use the GitHub, and feel free to try your hand at implementing something too!

posted this
Time ago

Community Update - Waterfall is now Open Source!

It's been a long time coming, and I finally stopped procrastinating.

Waterfall is now Open Source.

What is Open Source?

Open Source means that the full source code of the site is available. The idea behind Open Source is that anyone can inspect, modify, and commit their changes to improve the project. Think of the saying "two heads are better than one" - while I'm certainly not a bad coder, I'm not an expert either. Someone will, undoubtedly, look at the code and find something I did wrong. With Open Source, they can just fix it, and let me know about it - and then bam! Waterfall just got improved. Except instead of two heads, it's the entire planet able to help if they feel like it.

Why make Waterfall open source?

I've made a few posts on my personal blog about how I hate the current trend of the Internet - that trend being "there's like seven sites". I miss the old days where people could set up their own forums and stuff. phpBB exists for those, and Mastodon exists for Twitters, but there's not really been a good Tumblr thing you could spin up yourself. Then I realised I made Waterfall and oh shit, I can just publish my code so people can do that.

It also means people can verify statements on things like "the site doesn't track you", etc etc. It also opens up the floor to others to contribute and improve the site in ways I couldn't do on my own. Additionally, more people working on it means I have more time to work on other things, so it benefits me too.

Waterfall is still Waterfall - but this way it survives forever, and, if we're lucky, might get a little of the old Internet back, where niche sites can carve out their own little corner instead of there being just Reddit and Twitter.

This also means anyone can run their own little version of the site. I'm nicknaming them cascades. In theory, anyone can run their own little private version of Waterfall. In reality, you'll probably need some technical chops about you - it's not an easy thing to run, and needs some beefy stuff and knowledge behind the scenes. The preferred option is you help improve Waterfall proper to make it better for everyone, but if - for example - you want to run a private little thing for you and your friends to roleplay as your D&D characters, that's a perfect use for it. Larger scale stuff is not recommended though - Waterfall got expensive pretty fast as people started joining. This site is still the "canonical" instance, and as stuff gets pushed to the repo, the site will be updated with all the new features.

The code is licensed under the Affero General Public License, which means if you run a private instance, any modifications you make to the code must be made publicly available. You're also not allowed to pretend you made it, change or remove the license, and you have to list any changes you made too.

How can I contribute?

Check the repo! In particular, contains some stuff to get you started. There's now also a developer channel in the Discord. I look forward to working with you all! Developers who consistently contribute good code might get a little cash every now and then as thanks, given the Patreon is still a thing. Don't worry if you're not a coder - feature suggestions, bug hunting, and helping write documentation are all good ways to contribute too. If you do want to take a crack at it though, check the issues - some are specifically marked as good options for beginners.

What can I contribute?

Check the issues and see what needs doing. At the moment, the code is in a transition period of "this is just Thell and has caffeine dependency", moving towards "this is a serious open source project now". The code is a mess and very much reflects that only one person was maintaining it, so commenting code, cleaning it up a little, and writing more tests are all things that need doing and are things I'll be working on over the next few days.

If you have good ideas for features - go ahead and implement them! Throw in a PR, and it'll be merged in and added to the site after being reviewed.

After that - Waterfall 2.0 is going to be a thing, designed from the ground up (while maintaining backwards compatibility, of course), designed from the beginning with this open source, self-hosting stuff in mind. Feel free to join the Discord to get in on the design stage of that.

As always, join the Discord to discuss!

Time ago

posted this
Time ago

Community Update

Some major news and then some minor this update.

First though, a big welcome to everyone who's joined over the last few days - I hope you're enjoying the site so far!


First, the major news, and unfortunately, it's not good. As you know I've been working on the app alongside the rewrite, and I've been trying to get it into testing programmes. Unfortunately, the apps have been rejected.

After following up with contacts, it turns out this is primarily because of the type of app it is in relation to the team behind it - in short, to get a UGC (user generated content) app published, you need a large moderation team, and the balance sheet to prove you can pay them. Since Waterfall is one guy and a Patreon, that's clearly not happening any time soon. Unofficial figures I've been quoted say at least 15 full time, paid moderation staff, which means about a $500k a year income.

It's unlikely the Patreon will suddenly see a few zeroes on the end of the current income, so this, unfortunately, means the app is effectively dead at this point. There are ways around it - sideloading on Android for example - but I don't think putting effort into something that'll benefit 1% of the site's existing users who are okay with disabling their phone's security and technically competent enough to sideload in the first place is the solution here, and I think that effort is better spent elsewhere.

Following Bug Fixed

You may have noticed a bug with following last night - this has now been fixed. Unfortunately this probably means blocking is weird in some way, so expect that to be compromised right now. I'll be looking into this.

UI Update and Future

The staff post poll has ended, and updating the UI has won out. This'll be the next major feature update and, given the app isn't a thing anymore, I'll try to make it even more mobile friendly since right now mobiles miss out on a few features because of space constraints.

I'll also be adapting my schedule a bit - I spent a year flat out working on the rewrite, and a year before that working on the original Waterfall, which has left me tired. Given this wasn't meant to be a full time job, I'm going to adjust the hours I spend working on the site to reflect that, and finally take some time to work on the other projects I've shelved over the last couple of years. All of these can be supported at the site's Patreon, so feel free to donate a few bucks to keep things running. There'll be a post-mortem of the whole rewrite process on there soon, for those curious about the takeaway and technical details.

This doesn't mean the site won't be updated - indeed, security bugs will be fixed as soon as humanly possible when they're discovered/reported - but minor bugs will be a little slower. This reflects the fact that, hey, this is one guy who isn't really doing this to make money, so I think that going forward the "best effort" approach is the safest and sanest. I'll be targeting at least one bug fix a day.

As usual, suggestions and bug reports can be filed on the GitHub, including letting me know of areas where the help articles are lacking. Check the footer for a link!

posted this
Time ago

Hey all - image re-conversion is almost done, with about 4 hours left on the clock. Given 90%+ of the images on the site work now I went ahead and removed the banner. For the techies wondering, the storage burden of images has increased from 60.7GB with just WebP, to 580GB with legacy formats (and still going up thanks to the last 10%).

While that's been going on, there's been plenty of hotfixing going on behind the scenes. You can see more or less what's been fixed on the GitHub, but first, two major things that need going over.

First, audio posts - they've been changed a bit! A couple users started to upload podcasts to the site, which is fantastic. To help support this kind of content, audio posts now behave similarly to video posts - the audio is queued for conversion, and then the post will go live once it's finished. This should help alleviate the timeout errors people have been getting on longer audio files.

Second, there's been a significant uptick in artists neither tagging NSFW on their pieces, nor marking their blogs NSFW if they're an adult artist. I want to take this opportunity to reiterate that tagging NSFW is not optional. The site cannot prevent minors from seeing NSFW if you don't tag it, and, ignoring the legal and ethical ramifications for the site of letting that continue, having NSFW out in the open all but guarantees Waterfall will not be able to get an app released if it's easily accessible to minors. It's the sort of thing that makes sites eventually end up being hostile to NSFW creators, and I really don't want that to happen here. Repeated failure to tag NSFW at a minimum bars you from being searchable and at worst is a bannable offence, so please make sure your adult works are tagged, and your blog flagged if it's your main thing. Tagging it is as simple as adding "nsfw" as a tag.

Thank you, and see you soon for the next slightly more major update.

posted this
Time ago

Waterfall Bug Bounty Programme

While I work on fixing the minor bugs that have been reported, my mind started to wander. While there's been no showstopping issues reported yet, there's almost certainly a few lurking in the code, waiting to be found. So I thought - why not run a bounty?

What's a bug bounty programme?

Wikipedia can do a better job of explaining, but in short, a bug bounty is an official programme where people who find serious bugs (in this case, security ones) can be rewarded.

Why is Waterfall running one?

There hasn't been one reported yet, but somewhere in the code, there's going to be a serious bug that could have serious consequences. Waterfall is a solo project that's basically for fun and, while I'm not terrible at security by myself, it also means there's nobody to really review my code and find anything glaringly obvious. This means that occasionally, something serious might sneak by, and having that in public with a "here's basically the instructions to ruin everyone's day" attached in the form of a normal bug report is a bad idea. This is where the programme comes in.

How does it work?

Here's the page going into more detail.

In short, if you find a serious bug with security implications, you email it to the address on that page, giving me as much info as you can. It's STRONGLY advised you use a sideblog or side account for this, but specific considerations are outlined on the page, including how to get dev environment access for anything that might be particularly destructive.

It should be noted this bounty is for serious bugs only, mainly affecting security or privacy. Other bugs should still be reported through the GitHub linked in the site footer.

So what are the rewards?

Trivial bugs already get badges rewarded, and the bounty programme is no exception - you'll get the standard ones plus a special VRP badge. In addition, for stuff that's impactful, there are cash rewards on a sliding scale of severity. It's worth noting here - while most companies offer a bit more than this in their BBPs, Waterfall is a solo project, so it's limited by what one guy could realistically afford. Better than nothing though, right?

You can take the cash yourself, or choose to donate it to charity.

With any luck, nothing major will ever be reported - but the chances of that are slim, and it'd be utterly hubristic to think there's no flaws in my code that could be dangerous, and people who help defend against that definitely deserve to be rewarded.

posted this
Time ago

The re-conversion process for images is now underway, and is about 5% done - thoes of you on older devices should start seeing stuff again soon. I'll keep the warning there until it's at 95% done or so just so any new users know what's happening. This'll also fix the low quality on older images, I cranked that setting way up. You don't need to do anything to have those fixed, it'll happen automatically when it gets to those images.

One small side effect - a very small number of blogs have their avatars bugged right now. This is fixable simply by re-uploading the avatar. Not many blogs should have been affected before I caught it, but it's worth mentioning just in case.

One further note - I'm aware that editing image and art posts currently doesn't show the image stuff, meaning you can't re-order or change the images in there right now - I know about this, and re-adding it is my current priority. Sorry if you need that in the meantime, I'll get it back in as fast as I can.

Time ago

Regarding Images


As some of you have noticed, older browsers can't display WebP images, and you get a warning about it. This decision was made based on two factors:

1) WebP has been a standard for a while now

2) The cost of storage

Storing two image formats gets expensive, real fast - especially when older image formats, like PNG, are up to 15x the size of the WebP image.

Technical data from CanIUse said that the number of people who wouldn't be able to see images would be negligible - however, it turns out this isn't the case! Firefox and Chrome are listed as supported for a long time, but apparently on iOS, they rely on Apple to provide the libraries, which they apparently didn't until iOS 14. This means the math was WAY off.

So, this is a heads up - without a downtime I'll be reconverting all the images, this time including a PNG version. This does mean the site gets a bit costlier to run, but a LOT more people can't use it than I anticipated, so I'm hoping it's worth it. This'll take a couple of days, so the warning will stay in the header in the meantime, but I wanted to let you know.

In addition, this will also solve the compression issue with stuff that was converted - a library I used had some internal changes that I didn't catch so the quality setting didn't work anymore. This is now fixed, and has been pushed to live, so images uploaded going forward should work fine too.


Update! The transcode server has been updated to create both modern and legacy formats of any new images uploaded. I'll be working on the stuff to get existing stuff re-converted as soon as possible, but people should start seeing stuff again soon.

posted this
Time ago

Regarding Images

As some of you have noticed, older browsers can't display WebP images, and you get a warning about it. This decision was made based on two factors:

1) WebP has been a standard for a while now

2) The cost of storage

Storing two image formats gets expensive, real fast - especially when older image formats, like PNG, are up to 15x the size of the WebP image.

Technical data from CanIUse said that the number of people who wouldn't be able to see images would be negligible - however, it turns out this isn't the case! Firefox and Chrome are listed as supported for a long time, but apparently on iOS, they rely on Apple to provide the libraries, which they apparently didn't until iOS 14. This means the math was WAY off.

So, this is a heads up - without a downtime I'll be reconverting all the images, this time including a PNG version. This does mean the site gets a bit costlier to run, but a LOT more people can't use it than I anticipated, so I'm hoping it's worth it. This'll take a couple of days, so the warning will stay in the header in the meantime, but I wanted to let you know.

In addition, this will also solve the compression issue with stuff that was converted - a library I used had some internal changes that I didn't catch so the quality setting didn't work anymore. This is now fixed, and has been pushed to live, so images uploaded going forward should work fine too.

posted this
Time ago

Welcome Back

If you're seeing this, it means that the update is out, and I made it through the frantic few hours of emergency bug fixes that were inevitable from people who aren't me starting to use the site.

What's Changed?

For the most part, the site should look more or less the same - this wasn't a UI update (see my previous posts for details on how aesthetically challenged I am), save for a few QoL updates. On the backend though, everything has changed. This means we're out of beta.

Does that mean no more changes?

The opposite actually! The beta marker is honestly arbitrary, and now we're in the "perpetual development forever" phase. In the footer below, you'll see a link to report bugs and request features. You can use that to tell me what you want, and, over time, it'll be added if enough people also want it.

What's Next?

For the next day or so, I'll be in triage mode - there's bound to be a few more bugs I've missed (especially minor ones - a few have already been reported). I'll be steadily working through those until we're bug free, then I'll start introducing the holdout features I mentioned - things like giveaway posts, comic mode, etc. Once THOSE are in, I'll work on whatever you guys vote on below first, alongside the app - after all, the whole point of this rewrite was to make the app easier.

This has been a shorter post than usual, but understandable - the whole process left me a little exhausted, so there's not many words to say right now. In the meantime, if you encounter any major bugs or missing features - the only one intentionally missing is chat posts, since those were only ever glorified text posts anyway - let me know in the Discord. Otherwise, use the GitHub form to report them and I'll start working through them after a little rest.

In the meantime - welcome back!

What major feature should be added first, after bug triage is finished?

18.994413407821% of votes
Commission Market 2.0 18%
26.815642458101% of votes
UI overhaul 26%
11.731843575419% of votes
Advanced Blog Activity 11%
24.022346368715% of votes
Themes 24%
18.435754189944% of votes
Disregard new features; get the app done first 18%

posted this
Time ago

Community Update - Waterfall 1.0

As promised - the next Community Update would be the release announcement.


After more than a year in the works, it's finally time to deploy the Rewrite.

What's the Rewrite?

For those who haven't been here since the beginning and so are unaware, the Rewrite is a project I started late 2019. The site was getting bigger than I anticipated and a successful Kickstarter had just happened to make an app. Unfortunately, the code showed it - what had been slapped together as a proof of concept was now the foundations for what you're using right now, and it was getting increasingly difficult to add new features and adjust the existing ones. So, I pretty much decided to start from scratch, and rewrote the entire site. That's what's being deployed tomorrow.

Those who were here in the early days will remember how fast new features came - sometimes, I was able to deploy two or more major features a day. The rewrite means those days return, and I can make things happen much, much faster than over the last 9 months or so.


See the post directly underneath this one for the formal downtime announcement!

There's no formal estimation given for how long it'll take, since that'll largely be determined by how many images are on the site, since that's where the bottleneck is in the dry runs. I've optimised it as much as I can, but it'll probably be at least 24 hours.

What'll happen during the downtime?

During the downtime, the site will be completely unavailable, save for a maintenance page. It'll have some basic status updates on there, but you're encouraged to come hang out in the Discord while it's going on! We'll play a few games together or something.

Almost all data on the site is being converted into a new format. The only things I can think of that "break" per se are blocklists, blog descriptions, and badges. Blocklists have been talked about enough so I won't reiterate on that. Blog descriptions are being changed - given the app will start development immediately, it felt reasonable to prep for those now. They're being changed to be shorter rather than full bios, so that when they show in the app, someone doesn't have to swipe two full screens down. This means they'll be cleared in the update, alongside your badge settings - this is because the internal structure of badges has changed, so you can pick them all again (you can change the order of them now!) Your blog and dashboard theme may also be reset for the same reasons, so all in all, nothing too destructive is happening.

What'll happen after the update?

I'll be on standby to fix any major bugs that crop up. Given every line of code has been rewritten, there'll likely be a few that I missed.

Minor bugs will be put on a list to be triaged, and they'll be fixed in patches over the coming days and weeks, along with adding the hold-backs.

Hold-backs in this case refers to new features that are mostly complete, but are being held back a bit to make bugfixing the major stuff a little easier. It also includes one or two things, such as the advanced page editor, that are finished but I've found major bugs (i.e. security issues) in that need fixing, but aren't major enough features that it's worth holding the update itself back for.

After that, the GitHub repo and Discord will be open again to accept new feature requests, and a staff post will be made with a poll on what you want to see first. There'll also be a Patron update that gives a post-mortem on the whole rewrite saga, so keep an eye out for that if you're supporting the site.

It's been a long time coming, but we're finally there, and here's the 24 hour warning I promised.

See you on the other side.

posted this
Time ago

Downtime Announcement

At 8.00pm UK time (3:00pm EST) on the 22nd of January, Waterfall will go down for Extended Maintenance.

The purpose of this update is to deploy Waterfall 1.0, and move the site out of beta. No ETA is currently available for completion.

posted this
Time ago

Community Update

We're almost there - you can probably expect the next update to be the downtime announcement. But first, a few things to go over.

Blocking Returns

You might remember me saying a few CUs ago that blocking would be temporarily removed in the update - this is no longer the case! It's been redesigned and reimplemented to be simpler and more managable. Rather than trying to make it state of the art, I decided a simpler system that does what you need it to most - stop people contacting you - and divert the effort in making it foolproof towards making it harder for you to see content you don't want to see. I feel like this fits better with what blocking is - a nuclear option - and that tag blocking etc can be improved further to deal with just not wanting to see certain content, so that's what my effort will be going towards going forward.

Existing blocks still need to be cleared out due to the nature of the changes, but once they're re-added you won't need to add them again. There's no need to do this until the update hits.

Art Discovery and UI

If you've looked at the site, you'll have figured out by now that I'm crap at UI design. Art Discovery is implemented in the update but looks very boring. I'm aware of this and a UI update is planned to coincide with - or be released before - the app, since having them be at least vaguely thematically similar is probably a good idea. Also, nobody could agree how the Discovery pages should look, so that played a part too.

To help with that, I'm planning a workshop on the UI and Art systems soon - feel free to drop by and give me your thoughts on both of them. This'll get a dedicated post after the update hits with more info and times.

New Help Site

I've been working on setting up a dedicated, proper help site for the rewrite.

I won't link it just get, but if you guys could comment on this with aspects of the site you find confusing - or found confusing as a new user - that'd be extremely helpful! I'll make sure to write an article on the topic ready for the release.

Release Schedule

There's 29 issues left on the "essentials" list, and I closed another 14 today. That means we're talking a matter of days now. I have my Red Team doing their thing already (for those who aren't computer people - this basically means hackers for hire doing a quick security sweep and letting me know of any security issues I missed) while I finish up.

The plan is a staggered release - most of the update will release in a few days, with extra features rolling out piecemeal over the following days/weeks. This is mostly due to the scale of the rewrite and new features - I actually counted earlier, and only 17 lines of code remain from the current build of Waterfall, out of a few hundred thousand (more in the rewrite!). That's as close to a total rewrite as you can get. It also means that, almost certainly, some bugs are going to be introduced because of how much has changed. So, the bigger features get held back a bit so that while I try and fix those bugs, there's fewer moving parts - less new weird stuff that could be affecting things so I can squash them better. Not terribly ideal, but that's the trade-off for being a team of one.

Anyway, we're very close now. I'll finish up and run a few test converts on the dev server to make sure things run smoothly, and then give you all at least 12 hours warning about the downtime.

Assuming nothing terrible happens, next CU will be the downtime announcement.

posted this
Time ago

Community Update

This will be a more text heavy update than the others because honestly, there's not a whole bunch interesting to show about most of it. Buttons are only interesting for so long.

Group Blog Overhaul

I got halfway through rebuilding group blogs when I realised there was a much easier way of doing things than emails with convoluted codes. Instead, group blogs are now managed entirely from within user settings, including accepting invites and leaving them.

Also, I added a few extra permissions since I figured if I'm going for granularity I might as well go all the way.

Badges Overhaul

There's been some new badges added, and I've completely redesigned the way badges work internally. What's staying the same is that blogs can have up to three.

The differences are that now, they're broken into categories. The initial idea was that badges can be used to display achievements of some kind, and, when BiNet pulled the shit they did, was extended to have the Bi flag as a fuck you to them. Then the rainbow pride flag was added, and it sort of snowballed into the 12 we have right now for various reasons. Anyone who looks at the option list and how it displays can probably guess that it wasn't intended to have so many badges as options. At the same time, I can't really get rid of them now they're there without upsetting people, so the obvious solution is going to be categories of badges.

This is split into two bigger categories - User Badges, and Blog Badges. User badges are stuff available at the user level - stuff like the pride badges, any patron badges, Doctor/Scientist, Exterminator etc etc. Blog badges will be locekd to that blog, and will be mostly sort of achievement related, as well as the Artist badge when the CM is re-added. There'll be a lot more to choose from, and also earn - and this time the page is actually designed to take it into account. Oh yeah, you can reorder them now too.

There's a lot of badges added for actually participating in the site, that sort of serve as achievements. They'll be awarded automatically without need for intervention from me.


I realised I use emoji a lot on Discord, and that bleeds over into some of my shitposts on here as well. So I went ahead and added them, meaning that when the update hits, all your :sparkling_heart: stuff will show as the actual emoji. In fact, if you're reading this after the update, that should actually show as the emoji rather than text.

There's also a second component to this system. I've designed it to support custom emoji sets, meaning that, as you can see form the second one there, custom stuff can be added. This means that in addition to the inevitable Official Waterfall Emoji Set, individual artists can upload their own sets for the site to use - or, if I ever get off my hands at make it/figure out what licenses I need, the Patreon alternative I talked about building a while back can allow artist's patrons to get access to special subscriber only stuff for artists.

For now though, it'll be manually curated by me. I have a couple artists in mind who I'll approach soon about it (or you can offer in the Discord, if you feel brave), but the exact criteria I'll be using for now on which artists have custom sets uploaded is still in the air and will likely evolve as the site grows. Uploaded emoji (other than the patron stuff) will be usable by all users on the site.


Been a long time coming, but I have a working prototype of a chat system. However, it's a little buggy right now, so if this makes it in, it'll be marked as experimental and opt-in. It currently supports one-on-one, and group chats with no defined maximum limit. More on this next time, but I wanted to mention it.

It's especially broken when I try it in the app prototype, so this might be someting pushed back to a future update.

posted this
Time ago

Welcome to all our new users today!

The site is still a work in progress (with major upgrades to basically all aspects of the site just around the corner), but I hope you'll enjoy it here!

Some useful info:

The Discord is where you can hang out with us and chat. There's also a channel for shilling your blog a little, alongside somewhere to get help if you're confused about something. Announcements tend to be posted here first! There's also a feature request channel for if you think there's something I should add to the site.
The #artists on waterfall tag is a good tag to start posting art in - a lot of users use this to check out new artists while I finish up the dedicated discovery system. Just remember to tag your NSFW and mark your blog appropriately if that's your jam (you can also use that setting to keep minors away if you're just not comfortable wth being followed by under 18s).

The featured posts system on the site works a little different than most other sites for the moment - simply put, all art can be featured, and automatically will be once it hits a certain notes threshold. Don't be shy, go ahead and post it.

Hop in the Discord if you have any questions, and I'll do my best to help!

posted this
Time ago

Community Update - More Teasers

The more progress I make, the more I have to show off. And since showing you guys what I'm working on means I can get feedback and hopefully not get too far down the wrong path that it's not easy to change, so you get to see more of what's done!

First, a quick note to NSFW artists - please make sure your work is tagged appropriately. The prescribed tag that must be present is "nsfw" - not a variation on it. You're welcome to add your own variations in addition to it, but that tag MUST be present. This is coming up because we've had people not tagging it and it showing in features. Using the tag correctly makes sure minors can't see it (or adults who just don't want to, for that matter), and is mandatory. The "this blog is for adults only" toggle isn't enough, since that's not explicitly an NSFW flag.

With that out of the way, let's see what we have for you. As always, WIPs and will likely change a little in the final version.

Accessible Fonts

One thing that a lot of sites don't do well is acessibility. Waterfall probably doesn't do it well either, but that's no reason not to try and fix it. So, first things first - font size, and dyslexia mode.

There's been plenty of research that specialised fonts can help dyslexic folks read better, so there's an option to enable that now. Doing so changes the font for the site to the one pictured, and also increases line spacings a little, since that apparently helps too.

In addition, there's a large font option too. The screenshot above has both dyslexia mode and large font enabled to give you an idea of what it looks like.

Image IDs and captions

This took a surprising amount of fiddling to get working, but it's finally working 100%. Image and Art posts now support Image IDs natively, if the uploader chooses to use them (which they absolutely should!), as well as captions. Image IDs are used by screenreaders too, while captions show on hover with the mouse. Head over to XKCD and hover over a comic to see what's possible here. Obviously, the image ID in the screenshot sucks and you should use actually good ones rather than jokes, but that's on a private testing build so I figured I'd have a little fun.

While we're on the subject of Art posts...

Art Theft 3.0

The art theft system has been completely overhauled, and we're skipping 2.0. That's how much it's improved. Unfortunately, I can't really show you for this since it's just a mess of code, so I'll describe it instead.

Art Theft 3.0 now uses five different algorithms to determine whether the site has seen an upload before. Instead of using the file itself as a comparator, it uses the actual image. As usual, I don't want to get too in depth to prevent workarounds being discovered before I can patch them, but it involves a bunch of stuff like wavelet decomposition, perceptual hashing, histogram analysis, yadda yadda.

You probably don't care what those words mean and just want to know what it changes, so here's a summary:

  • Lightened/darkened images are caught now.
  • Resized images are caught now.
  • The algorithms are crop resistant, so cropped off watermarks are caught.
  • JPEG compression/artifacting is caught and adjusted for.
  • Adding noise to an image is caught and adjusted for.
  • Changed aspect ratios are caught.

I borrowed 500 art posts from the site and ran them through a generator to create variations, then had the algorithms run on them. Only 12 weren't caught. Checking those 12 there was a common theme among them (sketchy style, no/little colour, or a lot of whitespace), but this is a staggering improvement on the existing system, so it'll be deployed in the next update. Best of all, it barely slows posting down at all.

This is the main reason the rewrite has taken about a year to do.

Group Blog Permissions

The page for this is ugly as sin right now so I want to see if I can improve it a bit before launch.

The motivation here was simple - Tumblr's way sucks. You can either do nothing, or everything. It's the same way here right now, so I wanted to fix that. Rewrite's blog permissions are fully granular.

URL Purge

I said a while ago that inactive URLs would be purged occasionally. That system is now finished, and the first purge will be on the day the rewrite is deployed. Here's your warning.

And that's it for this update - more to come soon!

posted this
Time ago

Community Update - Happy New Year!

If you're reading this, it's the New Year in the UK now and Big Ben didn't play O Fortuna instead of the bongs, so this year will hopefully be better.

Last year for New Year's I added a feature - since the scope of the next update is so big I can't do that this year, so instead I'll talk about a few I'm adding in the big one instead.

Note: Design work in progress, so these pics are WIP. Still gotta do theming, which'll be done as the last thing I do.

Mutual Checking

One of the most used features of XKit on Tumblr is the mutual checker, so I decided to go ahead and add it on Waterfall. This'll show on your Following and Followers pages, and the check will run against your currently active blog.

Pronouns in post headers

This was suggested by the wonderful @jellyenvy, and it seemed like a good idea, so I added it! You set your pronouns on your user page, and can enable it per blog. Then, it'll show up in post headers. It doesn't show up in post segments due to space constraints, but that might change in future.

For the first iteration, this'll be limited to the most common three - they/them, she/her, and he/him. Adding neopronouns will be on the to-do list, but I need to find a way of going about it that minimises the risk of abuse (such as the occasional user who might think the helicopter joke is still funny) while also not accidentally gatekeeping which pronouns are "valid", so that'll be an exercise for after the next update once any kinks are worked out. They also won't show on group blogs, due to the multiple user factor.

OP Tag Blocking

An update to tag blocking, it'll now check the OP's tags too, and let you know whether it's being blocked from the reblogger's tags or the OP. This should solve about 80% of the edge cases I've encountered so far.


Poll posts have been added in the dev build. You can set them to run for 24 hours, 3 days, or 1 week currently. Single vote or multiple options, you can have up to 10 options. These are limited to one vote per user rather than per blog. Options you voted for have a different colour than other options that have votes, to make reading it easier. I just didn't think to include that in the screenshot, oops.

Obviously, there's more behind the scenes and I'll post more about those when I have something a bit less WIP for them.

In the meantime, happy new year!

Time ago

Rewrite adds optional pronouns to post headers so folks don't need to check your bio for them

WIP, will likely look a little different in the final release. Credit to @UnidentifiedBlog for the idea!

posted this
Time ago

With the CASE Act now signed into law, expect most sites to start changing their rules regarding fanworks. Waterfall doesn't follow US law, but this is a paradigm shift for the Internet. This means I'll soon look at WF's rules and adjusting them to be more permissive for fan works. Everyone can weigh in on the new rules, and there'll be a proper workshop on it soon. Watch this space for more news.

posted this
Time ago

I woke up to the news that CASE had passed, and SISEA might be next, so let's talk about them and how they'll affect social media, artists, streamers, etc. This is mostly copy pasted from my Twitter thread on it, so sorry if the flow is a little off.

SISEA has been talked about a lot, so I'll start on CASE. It was passed as part of the funding bill in the US, because America is a functional democracy where you can apparently tack unrelated stuff onto important bills.

It introduces some pretty shitty copyright reform. Ironically, this is a neutered version, so at least you won't go to prison. That means that if you're a US citizen, you can have some fun stuff to look forward to now!

This means if you download or share copyrighted stuff, you can get turbofucked to the gentle rhythm of a $30,000 fine. Legally, they're bound to NOT consider intent, and if the claim is for less than $5k in damages, they don't even need to send you a letter first.
It can be applied retroactively. This means if you have streaming aspirations, do fanart (hi Nintendo!) - especially adult fanart that companies might not like - or stream yourself drawing, you're at risk from this new law.

This also affects memes, especially since a lot of them are based on stock images now. The Distracted Boyfriend meme? Hope you have a licence to use that stock image.
But hey, at least you get $600. SISEA is another bill, but I'm not too qualified to go into whether it's good, especially since the chamber is split on it depending who you talk to in the industry it effects.

It mostly means that it becomes almost impossible for any US based site to host NSFW. At least, without spending a ton of money. This affects Patreon, Twitter, Tumblr, Pillowfort, etc etc.

Any US based social media site is likely to just say "it's not worth it" because the extra budget that needs allocating to deal with it.

Something I've seen is "well those sites will just have to move out of the US" - that's not feasible, for many reasons. Traditional business models for social media mean the sites are profitable selling ads and your data on to other companies.

The US is the friendliest country in the world to that business model. Moving to Canada or Europe kneecaps that business model. So much so, that it stays more profitable for them to just formally ban NSFW altogether.

Will it work? Who knows, we saw Tumblr try it. Though that was mostly a token gesture, with it enshrined in law, they'd be obligated to actually try harder at it than leave it as an open secret.

Compliance with CASE is a different matter. It's not on the sites themselves to enforce that (though I'd expect a bunch more posts getting taken down), but a lot of people are going to be fined over it.

Anyway, fact of the matter is any US based site is now inherently hostile to content creators. Not through any fault of their own (for a change), but because they have to follow the law.

Waterfall is immune to these laws and will not be complying with them.

posted this
Time ago

Community Update - The Future of Features, Tracking, and SISEA

Time for another little infodump while I take a break from wrangling with the new search algorithm. This post goes into NSFW on the site (don't worry), ditching tracking, and how features will work going forward.

SISEA and sites banning NSFW

First, if you don't know what SISEA is, here's a thread. Essentially, stricter verifications and required actions from platforms mean a good chunk of sites will probably just outright ban NSFW in general if it passes (it hasn't yet).

I'm not personally qualified to comment on whether it's good or bad, but a few sex worker friends I've discussed it with consider it a mixed bag (in a sort of "maybe good intentions, but the execution sucks" way). Either way, it doesn't affect Waterfall - we're not a US site so we can safely ignore that. If push comes to shove and I DO have to implement something, I'll work with any sex workers or adult content producers who want to get involved (anonymously or otherwise) on making sure whatever system we're forced into is safe and secure, but NSFW as a whole will not be banned on Waterfall.


I got off my ass and removed Google Analytics. While it's been useful in planning when to do maintenance, I've started hating the tracking bullshit so I've removed it. TOS and Privacy Policy will be updated shortly, and the only analytics I have now is a user count for who's been active in the last week. I might go with a self-hosted analytics thing in future if I feel there's a need for it, but for now, no more tracking. That said, Spotify and YouTube still inject trackers into the page, as do some of the CDNs we use to load a few things - there's nothing I can do about that right now, but I'll be working on stopping that in the rewrite. Use Privacy Badger and uBlock Origin in the meantime.

Art and Features

A discussion that's lasted a few days in the Discord now is the state of the features system and art.

The system as it stands is everything that's an art post gets featured at a minimum notes threshold. This is good and I like it because smaller artists get attention too after a nominal period of time. However, as time goes on, it's apparent that there's a few problems with the system how it works right now.

First, the nomenclature - "features" - gives a certain impression, being that it's the best stuff on the site. This is an issue for new users who don't understand the system's workings - which, unfortunately, is leading to them seeing stuff on there, thinking it's the best on the site, then leaving because it doesn't meet their quality threshold. To be clear, there's nothing wrong with being a beginner, and beginner artists definitely deserve attention too.

Second, there's the abuse problem - if it's "everything that's an art post", that's open to a bit of abuse. Getting commission sheets featured becomes trivial and turns it into basically a free advertisement for anyone who posts a commission sheet as art, which is definitely not the intended purpose of the system. Likewise, WIPs - while ostensibly kind of art - are also put into the art system, which makes things dicey when it comes to Art Theft 2.0, since many WIPs are uploaded as screenshots of Sai and such.

Loathe as I am to start defining what art is, these are problems that need to be addressed, ultimately for the good of the site even if it pisses one or two people off. So, going forward, here's what I'm thinking pending more feedback. No more automatic featuring, everything is manual. Featured posts will be hand picked and presented as the best on the site, as users joining for the first time expect them to be. In place of the current system, there'll be a dedicated discovery area for all art, biased towards artists with less of a following.

All new art will show there, and give folks an opportunity to find new artists. Since I'll be going through everything submitted as an art post anyway to find features, I can keep stuff that shouldn't be art out of this area - stuff like WIPs, memes, OC ref sheets (which are mostly text by area), and anyone trying to get abusive/illegal content into the system can be downgraded to an image post or deleted as needed, while other stuff can be waved through.

This feels like the best compromise where new users aren't put off by things working different than they expect, while also promoting new and lesser known artists effectively.

As I said, this in particular is open to feedback, so do feel free to comment or join the Discord for discussion on it!