The Problem with Un-Pinterest

It’s been seven months since I started working on Saves?! Yikes, I should write an update.

The downside of creating things for yourself is that you immediately spot ways to make them better. So you tweak. Then you see other improvements you could make and you do those too. You tweak some more. You add some more. The cycle keeps repeating until you get bored or we all reach the heat death of the universe. Either/or.

The upside is… the same scenario, actually.

The first version of Saves was… okay. It worked, but it was old-school; actions were done through traditional HTTP requests with page reloads. It covered the basics, but it lacked polish. It sank into the swamp.

So I built a second version that used Livewire and modals! It a much more modern feel. It also sank into the swamp.

But the THIRD version of Saves didn’t sink! I built it on the sunken bones of the first two versions and it’s standing strong. It’s still using Livewire but it’s much more polished and feature complete.

Reviewing the original list of goals…

A ‘home’ feed that shows the firehose of what I’ve recently saved. Done.
Collections and subsections for organizing. Done.
Tagging for additional filtering. Done.
Title, description, and ratings for saves. Done, except ratings.
Automatic color palette extraction. Nope. Did find an interesting post talking about color extraction that I might use as a basis though.
Image recognition would be cool, but I’m not sure my current cheap hosting can handle that (maybe locally, though…) Nope. Did start testing out using PHPNative to see if that would work but I wasn’t happy with its image packages.
A browser extension for saving images directly (with semi-smart collection sorting) Working, but not permanently installable; it has to be reinstalled whenever I restart Firefox because making a permanent extension hasn’t been a top priority. Otherwise it does its job well. Would like to have an iOS extension going as well (curse Apple’s developer fees).
Manual uploads for edge cases or stuff that isn’t online anywhere. Done.
Automatic thumbnail generation (including flattening GIFs into stills) Done and done.
Masonry layout. Done. Of course, there’s a new API coming along that will make this ridiculously simple and remove the need for all the effort I put into making this work.
Human-friendly URLs. Done.
Light and dark modes. Done.
Accessibility built in from the start. So far, so good.
Duplicate detection to flag already-saved images. Done, insofar as hashes and other data points. I’m not running image recognition.
Search across tags, titles, descriptions, and colors. Done, except for colors.

… and I just added an extra feature to the search bar to save an image quickly by entering a URL instead of going through the usual process. It saves at least two clicks. Totally worth it.