Zetheros
Developer
Hey Exanimates,
We've kept at it during the holidays, but we've also been kind of taking it easy rather than crunching to get a release out. It was nice to just take a breath and carefully consider some important decisions rather than trying to rush to get what we had out.
For the most part we've been carefully testing and tweaking the new item randomisation systems. This gets surprisingly complicated, as we now have a lot of fine grained control over what's randomised and how. There's a lot of diversity, and it's really not at all "random". We want some fairly strict rules about what materials, what level of quality, condition and dirt etc. are appropriate in many different contexts, and yet still have a lot of variety. Basically things like a high end shop would never stock cheap, slightly worn or dirty clothes, a peasant would never wear a silk shirt, an undead would never wear a clean pristine shirt, or this type of clothing or part of an item would never use certain types of fabrics and so on. There's a lot of potential rules to navigate here.
The problem is complex because we need to coordinate many randomisation parameters that must interact in a logical way, and because we want both very detailed and simple control for content design. Most importantly we need to very easily get the results we want without having to tweak a bunch of obscure parameters, which leads to indecision and error. This means a high level layer of control that is simple, intuitive and clever, but also the ability to get very specific results when needed. Another part of this we've been trying to solve is keeping items in saved games up-to-date by automatically patching themselves from master copies as we tweak these systems in future.
This might all seem a bit over-engineered, but it's the nature of this game to have such details and consistency, and for these things to be quite dynamic, rather than a result of statically designed content. A lot of work needs to go into a system like this that can match what a human might make by hand.
The other thing we've been trying to get fully implemented this week is the procedural item descriptions. These are quite important now, they tell you things about an item that you might not understand by simply looking at it. Where previously you would find two "rusted helmets" that were identical in all but minor visual differences, they might now be in similar condition but have different traits, build and material quality, making one of them potentially quite valuable and useful if restored. For the story in particular, this means some of the best items in a playthrough might be ones that you find very early on. For the descriptions we've been solving some minor grammatical hiccups, but we've also refined the system from our original design to more accurately describe items.
Besides that we've finally managed to do a passable procedural velvet material, which was the last one we were missing. We've made some good progress and some important refinements, the itemisation features seemed like a small thing after getting the all new renderer and assets ready, but they are pretty big in their own way and very important. All of this also carries over exactly to the upcoming procedural weapons, as we've basically ended up adopting the biggest part of those systems for other items too.
Have a great week!
-the BM team
We've kept at it during the holidays, but we've also been kind of taking it easy rather than crunching to get a release out. It was nice to just take a breath and carefully consider some important decisions rather than trying to rush to get what we had out.
For the most part we've been carefully testing and tweaking the new item randomisation systems. This gets surprisingly complicated, as we now have a lot of fine grained control over what's randomised and how. There's a lot of diversity, and it's really not at all "random". We want some fairly strict rules about what materials, what level of quality, condition and dirt etc. are appropriate in many different contexts, and yet still have a lot of variety. Basically things like a high end shop would never stock cheap, slightly worn or dirty clothes, a peasant would never wear a silk shirt, an undead would never wear a clean pristine shirt, or this type of clothing or part of an item would never use certain types of fabrics and so on. There's a lot of potential rules to navigate here.
The problem is complex because we need to coordinate many randomisation parameters that must interact in a logical way, and because we want both very detailed and simple control for content design. Most importantly we need to very easily get the results we want without having to tweak a bunch of obscure parameters, which leads to indecision and error. This means a high level layer of control that is simple, intuitive and clever, but also the ability to get very specific results when needed. Another part of this we've been trying to solve is keeping items in saved games up-to-date by automatically patching themselves from master copies as we tweak these systems in future.
This might all seem a bit over-engineered, but it's the nature of this game to have such details and consistency, and for these things to be quite dynamic, rather than a result of statically designed content. A lot of work needs to go into a system like this that can match what a human might make by hand.
The other thing we've been trying to get fully implemented this week is the procedural item descriptions. These are quite important now, they tell you things about an item that you might not understand by simply looking at it. Where previously you would find two "rusted helmets" that were identical in all but minor visual differences, they might now be in similar condition but have different traits, build and material quality, making one of them potentially quite valuable and useful if restored. For the story in particular, this means some of the best items in a playthrough might be ones that you find very early on. For the descriptions we've been solving some minor grammatical hiccups, but we've also refined the system from our original design to more accurately describe items.
Besides that we've finally managed to do a passable procedural velvet material, which was the last one we were missing. We've made some good progress and some important refinements, the itemisation features seemed like a small thing after getting the all new renderer and assets ready, but they are pretty big in their own way and very important. All of this also carries over exactly to the upcoming procedural weapons, as we've basically ended up adopting the biggest part of those systems for other items too.
Have a great week!
-the BM team
