I'd like to ask a few questions:
1. Can we write equivalents to new simple_behaviors? Will there be example code provided? One of the things that really concerned me, watching the videos from the last public showing at Gamescon, strongly indicated that the horseman AI is pretty much a direct port (and therefore, incredibly slow compared to the well-optimized code of the infantry AI).
2. How soon in advance of the release can we see the new model / mesh formats for characters, armor, and weapons, and examples of how they're implemented? Personally, I think this is vital; I don't think I would've made Blood and Steel if OpenBRF didn't exist already; the Taleworlds-provided Blender tutorial was weak, arrived too late, and wasn't really all that useful by comparison.
3. Will there be an automated tool for developing pathfinding in Scenes?
4. Will there (finally) be ways to manipulate the pathfinding mesh properly via code?
5. Will the pathfinding mesh include flags indicating whether an area's passable by infantry, but not by horses?
6. When will we see examples of new skeletons, and a workflow for implementing them, tying them to animations, etc.?
7. Will the modular-weapons idea be mandatory, or can a mod simply say that a given weapon doesn't have any variations?
8. How will we write new Party behaviors, interactions, etc.? What does the structure of Conversations look like?
9. We've gotten vague answers about access to the shaders. Can we at least write new ones and assign them to Materials? Or are we stuck with whatever setup is built in? If the latter, I'm probably not even bothering to mod Bannerlord, frankly; custom shaders allow for whole new looks and behaviors (and given that this is a DirectX 11+ game, I'm expecting to be able to use geometry shaders, etc., finally).
10. How will Scene terrains get generated, and how will code be able to interact with it dynamically? Will we be able to alter the heightmaps, random placement of objects, etc., via code?
11. Will we able to implement new core sets of animations that can be targeted on a character depending on the weapon / armor / shield they're carrying?
12. Will we finally be able to define all-new Weapon Types from scratch? This limitation in Warband is seriously annoying; I hated having to sacrifice one important weapon type to get another one working at all.
13. How will weapon animations be designed? I presume that the new engine will finally dispense with vertex animations, but what does that mean, exactly?
Will we rig weapons to a skeleton and animations? What if we have to make a new animation and new rig for it? Will we have the ability to tie parts of weapons to an IK bone and have physics performed by the engine, or coded in by modders? For example, what about all of the important weapons of the Middle Ages that aren't in Warband- flails, stave-slings, slings in general, spear-throwers, whips, nets, etc.? There are a ton of weapons that just aren't very practical to even add right now, and I see no signs they're in Bannerlord, which is really disappointing.
15. When will we be able to see the tools used to implement morphs on body armor? It's very important that we're able to understand how the morphs get used; the old male / female morphs were complex enough and required a fair bit of tuning to replicate in OpenBRF.
16. Has TW finally hired Mtarini to think about these issues with modeling and asset creation by modders for them?
Seriously; he's about the only person I'd trust with understanding how modders work, what we actually want (from simple n00b "I wanna take Kewl Sword from Some Other Mod and stick in My Mod" to "I want to implement objects with entirely custom shaders, vertex colors as control systems and passthrough data manipulation via middleware", and presenting all of it in a reasonably-coherent workflow. I really hope TW has at least made him an offer.
17. How will the new equivalent to flora_kinds work? How much low-level control will we be able to have via code?
18. Will the Bannerlord engine finally allow access to all object types in a Scene? It's a crying shame that we cannot iterate through a bunch of the major object types, let alone do anything with them. I'm really hoping all object types that aren't fundamental (i.e., terrain, sky system) are all Objects (in the coding sense) that are sub-classes of a fundamental Scene Object type. Then, say, writing systems to deal with object elimination in a Scene to keep performance high, etc., won't be fundamentally limited. I get that it's attractive to treat, say, grass objects as a special type (for performance reasons) but it's actually a terrible idea at this point in computing history.
19. How much explicit support will the engine have for multithreaded operations? I presume the main engine's multithreaded and takes advantage of today's multi-core CPUs, with well-conceived safe access to game variables, etc., but what about mod code?
Anyhow, those are my questions at this point. I really wish TW would talk a LOT more about how the workflow for content works, how they expect coding a mod to work (from Hello World to Custom DLL, compile-at-runtime, or another middle-level bytecode format), and how the engine works at a deeper level.
If they're hoping for the same mod scene, they really should have things set up so that newbies can put in a simple weapon into the game within an hour (using a provided example mesh, textures, etc.) and coders can implement new features rapidly and easily. Not like the early days of M&B, where only a select elite even knew how to get a mesh turned into a BRF
If the staff at Taleworlds want to see what an excellent, modder-friendly C# codebase should look like, in terms of quality of documentation, clarity of purpose and design simplicity, I'd like to recommend that they read through the API for Starsector (which was written by, and maintained by, one person!!!). That's what a really good codebase should look like at release, if it's aiming to be modder-friendly and attract a lot of interest at release.
I really hope we're not going to see a repeat of Warband, where a great deal of the core Module System code's not documented at all, is documented in Turkish, or in a mishmash of English and Turkish, with misleading comments, code segments that aren't Turing-complete, etc. that only a few elite coders can even wade through, let alone make new things with
1. Can we write equivalents to new simple_behaviors? Will there be example code provided? One of the things that really concerned me, watching the videos from the last public showing at Gamescon, strongly indicated that the horseman AI is pretty much a direct port (and therefore, incredibly slow compared to the well-optimized code of the infantry AI).
2. How soon in advance of the release can we see the new model / mesh formats for characters, armor, and weapons, and examples of how they're implemented? Personally, I think this is vital; I don't think I would've made Blood and Steel if OpenBRF didn't exist already; the Taleworlds-provided Blender tutorial was weak, arrived too late, and wasn't really all that useful by comparison.
3. Will there be an automated tool for developing pathfinding in Scenes?
4. Will there (finally) be ways to manipulate the pathfinding mesh properly via code?
5. Will the pathfinding mesh include flags indicating whether an area's passable by infantry, but not by horses?
6. When will we see examples of new skeletons, and a workflow for implementing them, tying them to animations, etc.?
7. Will the modular-weapons idea be mandatory, or can a mod simply say that a given weapon doesn't have any variations?
8. How will we write new Party behaviors, interactions, etc.? What does the structure of Conversations look like?
9. We've gotten vague answers about access to the shaders. Can we at least write new ones and assign them to Materials? Or are we stuck with whatever setup is built in? If the latter, I'm probably not even bothering to mod Bannerlord, frankly; custom shaders allow for whole new looks and behaviors (and given that this is a DirectX 11+ game, I'm expecting to be able to use geometry shaders, etc., finally).
10. How will Scene terrains get generated, and how will code be able to interact with it dynamically? Will we be able to alter the heightmaps, random placement of objects, etc., via code?
11. Will we able to implement new core sets of animations that can be targeted on a character depending on the weapon / armor / shield they're carrying?
12. Will we finally be able to define all-new Weapon Types from scratch? This limitation in Warband is seriously annoying; I hated having to sacrifice one important weapon type to get another one working at all.
13. How will weapon animations be designed? I presume that the new engine will finally dispense with vertex animations, but what does that mean, exactly?
Will we rig weapons to a skeleton and animations? What if we have to make a new animation and new rig for it? Will we have the ability to tie parts of weapons to an IK bone and have physics performed by the engine, or coded in by modders? For example, what about all of the important weapons of the Middle Ages that aren't in Warband- flails, stave-slings, slings in general, spear-throwers, whips, nets, etc.? There are a ton of weapons that just aren't very practical to even add right now, and I see no signs they're in Bannerlord, which is really disappointing.
15. When will we be able to see the tools used to implement morphs on body armor? It's very important that we're able to understand how the morphs get used; the old male / female morphs were complex enough and required a fair bit of tuning to replicate in OpenBRF.
16. Has TW finally hired Mtarini to think about these issues with modeling and asset creation by modders for them?
Seriously; he's about the only person I'd trust with understanding how modders work, what we actually want (from simple n00b "I wanna take Kewl Sword from Some Other Mod and stick in My Mod" to "I want to implement objects with entirely custom shaders, vertex colors as control systems and passthrough data manipulation via middleware", and presenting all of it in a reasonably-coherent workflow. I really hope TW has at least made him an offer.
17. How will the new equivalent to flora_kinds work? How much low-level control will we be able to have via code?
18. Will the Bannerlord engine finally allow access to all object types in a Scene? It's a crying shame that we cannot iterate through a bunch of the major object types, let alone do anything with them. I'm really hoping all object types that aren't fundamental (i.e., terrain, sky system) are all Objects (in the coding sense) that are sub-classes of a fundamental Scene Object type. Then, say, writing systems to deal with object elimination in a Scene to keep performance high, etc., won't be fundamentally limited. I get that it's attractive to treat, say, grass objects as a special type (for performance reasons) but it's actually a terrible idea at this point in computing history.
19. How much explicit support will the engine have for multithreaded operations? I presume the main engine's multithreaded and takes advantage of today's multi-core CPUs, with well-conceived safe access to game variables, etc., but what about mod code?
Anyhow, those are my questions at this point. I really wish TW would talk a LOT more about how the workflow for content works, how they expect coding a mod to work (from Hello World to Custom DLL, compile-at-runtime, or another middle-level bytecode format), and how the engine works at a deeper level.
If they're hoping for the same mod scene, they really should have things set up so that newbies can put in a simple weapon into the game within an hour (using a provided example mesh, textures, etc.) and coders can implement new features rapidly and easily. Not like the early days of M&B, where only a select elite even knew how to get a mesh turned into a BRF
If the staff at Taleworlds want to see what an excellent, modder-friendly C# codebase should look like, in terms of quality of documentation, clarity of purpose and design simplicity, I'd like to recommend that they read through the API for Starsector (which was written by, and maintained by, one person!!!). That's what a really good codebase should look like at release, if it's aiming to be modder-friendly and attract a lot of interest at release.
I really hope we're not going to see a repeat of Warband, where a great deal of the core Module System code's not documented at all, is documented in Turkish, or in a mishmash of English and Turkish, with misleading comments, code segments that aren't Turing-complete, etc. that only a few elite coders can even wade through, let alone make new things with