Time for our third, albeit slightly belated, community update!
In October, we...
As of 10th November, Waterfall has 26,832 users, 45,732 blogs, and 246,403 posts.
Things have slowed down on the new user front - that's to be expected, given we've not done any extra marketing lately and we've been slower with updates. We'll be remedying that soon!
Coming in the next week or so, depending on testing results, will be the tagging and queue overhauls. Ideally these'd be launched seperately, but practically, because the queue overhaul plays off the new tagging system and the tagging system would need to update posts already in the queue, they'll be released together instead.
Queues pose an interesting challenge, since the way they're stored is being changed entirely to how it really should have been in the first place if I didn't decide to implement the system in two hours at 3am. The three options are to delete everything in the queue, forcibly post everything that's in the queue, or move everything into the new system but set every blog's queue frequency to 0. We ended up picking the latter since no content is lost that way, and - as a sign that we need better documentation - a third of all posts in the queue are on blogs that have a queue frequency set to 0 anyway.
Setting every blog's queue frequency to 0 is likely to be the best option based on that, and then everyone can re-adjust their settings based on the new features.
We're not entirely sure if any other new features will make it into the next update, but we're hoping to get it out in the next few days. After that, there'll be a few QoL patches - options for disabling anon asks and allowing logged out users to send messages are one - before we add the next big feature, which will likely be the CM, at long last.
For the site's UI, we're still drafting the revised version and our designer is on vacation right now, so stay tuned for that!
We're also playing with some algorithms to improve blog recommendations. We're not sure which one we'll pick yet, but we've narrowed it down to four.
- Text posts can now be made without text content if there's a title present.
It would be really cool if in the future you could make it possible to view photosets as you're getting ready to post them. Sometimes you don't know what order you want pics in until you see them and can swtich it all around
Planned for the rewrite! We're overhauling the whole user interface, and actually previewing and being able to adjust photosets is one of the features I have marked down as "will literally delay launch if it doesn't work properly".
How will import from Tumblr work for art since Tumblr doesn't have an art post? Also art posts only allow for one image and some of the art I've posted on Tumblr is more than one image. Will I have to upload manually to have my art posted properly?
Rewrite will support more than one image per art post, dont' worry there.
For importing, that's just something we'll have to deal with - you can set it to exclude certain image types or tags, so if you have a consistent tag you use that's one solution, but you'll have to re-post them manually. We MIGHT be able to set it so anything with that tag is converted into an art post though, I'll add that to the to-do list.
Time for our second monthly community update!
In September, we...
As of 3rd October 2019, Waterfall has 26,672 users, 45,466 blogs, and 229,258 posts.
There's not much to say this month for this given everything has been mostly behind the scenes, but we can give some app news! We've designed a new API to use in the rewrite - this will enable both the app as well as any fancy applications you guys want to write. The iOS app has successfully loaded and presented test data in the format the API will use, meaning we've made significant progress there. However, app development is still slow since we can't launch the apps until the rewrite is done and bug free - no point building an extension if the foundations of the original aren't solid.
For the release of the rewrite, we're aiming for no later than November 22nd, with December 17th as a fallback date - in the first instance, because that's when Doom Eternal releases and I wanna play it, and in the latter case, if we miss the first date then launching on the first anniversary of the public announcement would be quite nice.
Can I transfer my original posts from Tumblr to my blog here?
And is there a timeline for getting an app?
Tumblr importer will roll out with the rewrite - some backend changes mean we'd have to rewrite the importer from scratch for the new system anyway, so we decided to just wait. It shouldn't be long for that.
No timeline on the app since we're learning a lot and fiddling with basics still, but we are making progress.
One of the main things people wanted in the survey was more communication. So, we're going to try and do monthly updates on the site! These will contain a summary of what we've done, what's coming, and, on occasion, some insights into how things are being developed. We hope you enjoy!
In August, we...
As of 6th September 2019, Waterfall has 28,498 Users, 45,199 blogs, and 214,037 posts.
The Kickstarter ended, and money is now being collected! We're anticipating getting the money on the 16th, at which point backers will be sent a survey so we can collect some initial data so we can figure out what devices we need to buy for testing.
We also got 457 responses on the survey - that's more than enough to determine where it is we've been going wrong, so we just closed it for responses and will start going through to get more detailed info.
Right now, Analytics is the biggest stumbling block - if you're interested, hit the read more! Recommendations is something marked as major, because the current system sucks - it's broken, and we know it. Pools and communities are coming soon too - right now, the only thing to solve is how to go about letting regular users create them without staff intervention, while also not allowing duplicate pools to be formed - after all, if that were the case, it defeats the purpose of a central pool of content for a specific fandom or interest.
In the meantime, if you're not interested in the deep dive on the Analytics roadblock - let us know in the comments what you'd like to see in the next update!
With Analytics, we're inherently having to store data. Since the full version is a premium feature, we need to make sure the information it provides is useful - but, we also need to make sure it's private. While developing it, I have three things in mind - not necesarily true, but principles that are dictating how I go about designing it.
For the first point, we're doing everything we can do avoid that, but programming defensively is key with projects like this - you try not to have one, but you assume you will. After all, if Tumblr and Google and Facebook can be hacked, then so can we. How do you minimise the impact when it happens?
For the second point, we assume folks want as little stored on them as possible. There's some stuff that will be because the site needs it to function, but how can we minimise it?
For the third point, it's in the name - the data needs to be useful, or what's the point of having the feature?
So, with these in mind, we need to store data in a way that anonymises it, but is also able to provide drill-downs for statistical purposes. Regular users won't get access to all the data, just an aggregated version. Even then, staff shouldn't be able to go "ah, this is that user".
After spending weeks researching, the solution seems to be a neat thing called Differential Privacy. These will allow analytics to function correctly, as well as giving a good statistical base for more interesting stuff, while letting the actual users in said dataset be pretty private. The simplest test I'm using with this is, assuming someone had the full database, could they figure out what kind of porn I've been searching for? If they can, it's not secure or private enough.
There seems to be an issue with the "update email" section. I've been trying for a couple of weeks now to change my email address but it keeps telling me that the email has been updated and I should check my email account. I'm not getting any emails though, neither in my inbox nor in the spam folder. When I log off and back on here it's reset to the previous email address. I thought I drop note here maybe there's a bug or something?
There's some known issues with emails right now - notable, we know that Yahoo rejects all our emails outright and will never deliver them, and a few other providers send them to spam. We're still looking into the problem and possible solutions already.
- Reblog notifications will now show content from the reblog in the notification. This only affects reblogs after this patch is deployed, with older reblogs being regenerated in a future update.
- Turned commenting on DNR posts back on.
- Added additional special case tags - "dni" and "do not interact" which will disable both reblogs and comments on a post.
- Re-extended the dashboard slightly based on additional feedback and to resolve issues with post buttons.
... I know you just did all this work making the post width narrower and all but from someone who liked the wider width bc less cramped feel... is there much chance of there ever being an option switch or something in the future so people could choose to have wider or narrower posts? (if not, we can deal, but the wider posts did look so much nicer...)
Almost certainly not built in, but someone is free to make an extension to restore it. The problem is we want art to look good - so if we have an option for the wider dash, that means stuff is either going to get stretched, or it's going to get squashed, depending which resolution the artist designs for. There needs to be a standard resolution the site uses, and the majority of feedback was overwhelmingly that we were too wide.
To contrast quickly - Tumblr's width is 540px, Waterfall's old width was 825px, and the new width is 635px. Of the options presented in the Discord (between 825px, 730px, and 635px), this one was the winner.
However, if feedback is negative enough, I'll try out 730px and see if that's more tolerable for everyone.