It really has been a while since my last log, hasn't it?
Much of what I did recently, you already know about: Symbiotes. The release was a huge deal, and it's incredibly cool (and slightly frightening) to see people digging deep into the API and documentation details! The release was a huge success with only a few bugs in the API and positive reception, but it's also shown us several places we need to work on - overall I'm super excited to see where things will go from here.
But I want to talk about something other than Symbiotes today! Over the last weeks, we've brainstormed on asset search within TaleSpire. Currently, the asset search in the library is serviceable if you already know the names and tags of the asset you're looking for. We do have some problems with consistency of what is tagged and how exactly, but that's fixable, even if it's time-consuming. The bigger problem we're starting to encounter is actually because of a good thing:
Throughout the project, the art team has been releasing tons of amazing asset packs (and they're not about to stop!) leading to an ever-increasing amount of content available in TaleSpire - the problem is the more assets there are, the harder it is to find specific things and the harder it is for new players to get an idea of what is available.
I won't get too far into the details of our plans, mainly because it's all still very early on, and there are few details to go into. So far, they're just ideas that sound good in theory that we'll have to test and see if they work well in practice.
What I do want to tell you is what we want to achieve with an overhauled tagging/search system:
- Discoverability: We want new players to be immediately able to get into building without first needing to memorize what assets they have at their disposal. A big part of this is to make tags more predictable - if we tag one asset that contains wood with "Wood", all should be.
- Ease of authoring: A big part of why our current tag system is falling behind in usefulness is not an inherent problem with the system itself but with the tag authoring tools within TaleWeaver, which are a bit too cumbersome. Our poor artists already work hard enough on the assets themselves; categorizing them should be as quick and straightforward as possible. Having good systems there will also help with modded content once that's available because the easier we make tagging (for our art team and modders), the likelier it is modders take the time to tag their creations properly!
- Translations: While actual language support for TaleSpire is still a while out, we don't want to create systems we need to replace again as soon as we start translating. Because of that, we want to account for translated tags in the core design of the new system.
- Scalability: The system needs to be scalable, not just with new assets in existing categories, but also other "things". Ideally, whatever system we end up with can be used under the hood for everything we may want to add that is searchable - from assets in the library, the music player, maybe even the board list, etc. We are still far too early in the design phase to be able to guarantee anything in this regard, though.
Additionally to all this, it's important to us not to break the simplicity of the UI we currently have - optionally showing a more in-depth search or filtering UI is excellent, but we want the "default case" to stay as simple as possible - we want TaleSpire to be usable without needing a PhD in Data Science (which, incidentally, none of us have).
I'll end this with a disclaimer: This is a longer-term endeavor, so don't expect any results soon. We can likely implement some parts of the design standalone from the rest, making it more likely that we'll have partial overhauls earlier - Still, it will take some time even just to go through and reevaluate the tags on our current assets.
Disclaimer: This DevLog is from the perspective of one developer. It doesn't reflect everything going on with the team
: This will obviously also benefit all other players, not just new ones.
: Of course, in the end, it doesn't matter to any user whether or not we manage to use one system for everything - as long as the final result is "I can easily find things that I am looking for" everything else is just "boring" implementation details.