Zetheros
Developer
Hey Exanimates,
The main goals for this update have been reached, we've had some time to polish things too, add some smaller features and QoL, deal with some old issues and all sorts, but the new content is still being worked on, and so development must continue.
With this update we introduced features to support the Potential system, which is how your thaumaturgy powers scale and how you choose which powers and domains you specialise into. For this system to be fully functional however we also need various ways to acquire potential that are quite involved, and we also need to introduce scaling for every single power, which often requires whole new mechanics for them. The former of these requirements should be partly built on top of new role, event and scripting with persistent state functionality, which we're not adding until the following update. It's a lot to do at once, so we've been gradually working towards it, if possible we'd like to get it mostly finished in a smaller follow-up update along with a few other things.
Right now we're working on various elements of this, and going into some depth with how it will be supported by the role system etc. features. There's a lot to analyse in terms of requirements and how we can transition into this, and we can already take some steps. This stuff will be the main development effort for the next update, and the sooner we can climb that hill the more we can do. This is the core feature set that will support the more structured interactions and gameplay that we discussed in the last diary.
Madoc's also finally had some time to do some work on improving loading times, this was on our radar for a while, and the new content in this update is particularly large or dense so it's more of an issue. Things are generally well optimised, we do asynchronous IO, dynamically prioritised streaming of GPU resources and split workloads, but still some levels are very large and filled with diverse objects, characters and items are unique and procedurally generated, everything is dynamic and computed on the fly, so loading can take several seconds. The first optimisation we wanted to attempt is kind of a raw performance one, where assets are being immediately processed for use in parallel to more assets being loaded, and can even function in a lower performance mode until fully processed, which can continue in the background even after loading. This already shaved a huge 40% off our loading times, a good first step, but eventually we'd like to get the times much lower still, so that transitioning between areas feels almost instant. This will need some aggressive streaming strategies, which are different and not as easy to do here as they are for outdoor environments, but we know we can make it work.
Have a great week!
-the BM team
The main goals for this update have been reached, we've had some time to polish things too, add some smaller features and QoL, deal with some old issues and all sorts, but the new content is still being worked on, and so development must continue.
With this update we introduced features to support the Potential system, which is how your thaumaturgy powers scale and how you choose which powers and domains you specialise into. For this system to be fully functional however we also need various ways to acquire potential that are quite involved, and we also need to introduce scaling for every single power, which often requires whole new mechanics for them. The former of these requirements should be partly built on top of new role, event and scripting with persistent state functionality, which we're not adding until the following update. It's a lot to do at once, so we've been gradually working towards it, if possible we'd like to get it mostly finished in a smaller follow-up update along with a few other things.
Right now we're working on various elements of this, and going into some depth with how it will be supported by the role system etc. features. There's a lot to analyse in terms of requirements and how we can transition into this, and we can already take some steps. This stuff will be the main development effort for the next update, and the sooner we can climb that hill the more we can do. This is the core feature set that will support the more structured interactions and gameplay that we discussed in the last diary.
Madoc's also finally had some time to do some work on improving loading times, this was on our radar for a while, and the new content in this update is particularly large or dense so it's more of an issue. Things are generally well optimised, we do asynchronous IO, dynamically prioritised streaming of GPU resources and split workloads, but still some levels are very large and filled with diverse objects, characters and items are unique and procedurally generated, everything is dynamic and computed on the fly, so loading can take several seconds. The first optimisation we wanted to attempt is kind of a raw performance one, where assets are being immediately processed for use in parallel to more assets being loaded, and can even function in a lower performance mode until fully processed, which can continue in the background even after loading. This already shaved a huge 40% off our loading times, a good first step, but eventually we'd like to get the times much lower still, so that transitioning between areas feels almost instant. This will need some aggressive streaming strategies, which are different and not as easy to do here as they are for outdoor environments, but we know we can make it work.
Have a great week!
-the BM team
