Okay, so what follows is the list for several OSP projects that I have ideas for. However my personal skill set is not optimally suited for them. So while I potentially can do everything listed below, it will take longer and look uglier than it otherwise could. Hence for all the these ideas, either I get help and publish them in the nearest future, or I "keep them in mind" for an indeterminate amount of time and eventually get around to implement them solo - but with no guarantee as to when. 
So, without further ado:
So, without further ado:
Redesigning Heraldry
OSP needs: 2D artist.
To be frank, M&B tableau system is an incredibly versatile mechanism, however it's actual implementation for heraldry leaves a lot to be desired. My primary gripe is with the fact that the entire heraldic system is based on scene prop banners. When applying heraldry on an item, modder is essentially projecting a banner on the item's texture. This has the following major drawbacks:
1. The banner has limited size, which necessitates the use of background color, which must be hand-picked for each and every banner, and mismatched color will make your heraldic items look like ****.
2. Many banners have non-trivial graphics at the edges, especially at the top and the bottom. This imposes additional limits to how these banners (and by extension, absolutely all banners) can be placed on item's mesh. When determining the banner's position on the tableau, it is necessary to avoid any artifacts while still making sure the main part of the image is not too large or too small and positioned roughly at the chest. This is not always possible, sometimes making for quite awkward heraldry positions.
3. Banners have their own texture and shadows which look fine on banner itself. However when combined with item's tableau, banner's texture can only degrade the image quality. If the item has a very detailed texture, the heraldry may already be quite distorted, so adding extra noise is not something we should be happy for. Or the item might be polished enough so that banner's texture is actually discernible - and you definitely do not want a cloth texture to be visible on your plate armor.
4. The entire heraldic system depends on entity ranges too much. For each new banner, you must have a new mesh, a new scene prop, a new map icon flag and a new record in banner color array, and all these entities must have the same relative position to the first game banner. As the result, adding a new banner is a finicky work and in 99% cases will break savegame-compatibility with previous versions of your module.
What would be an ideal (from modder's point of view) heraldic system? It would be the one where to add a new banner, modder must add a new square mesh with his new crest, and then add his new mesh to the mod's list of banners (probably with several parameters determining how that crest should be painted). Map icon banners and scene props must use tableaus to display heraldry, thus eliminating any need to create new versions of them. All game entries (map icons, scene props, items etc) use original clean version of the crest - therefore the only texture and noise imposed on it is by the item itself.
There may be different visions of what a game crest should be composed of. At the moment my personal idea is as follows:
1. Crest itself, as a simple square mesh. Crest image must fit into a square, and all important details should fit in the circle inscribed into said square. Background should be transparent.
2. Associated background color or background texture mesh. As the background should ideally be the same for all heraldic items, it should be a simple repeated pattern.
3. Optionally, additional graphical meshes, for as long as some standard is imposed on all of them. So some crest may have an associated "bottom" decoration, which would be ignored for round shields, but displayed on the lower hem of a tabard or pavise.
Essentially, when generating the heraldic tableau, background is drawn first, then the crest itself and any optional elements are positioned and scaled in a way determined by item's texture, and finally item's tableau texture is applied to provide the necessary "look&feel" and impose item's material and defects.
Additional benefit is the increased control over heraldry positioning - right now horizontal map icon flags look pretty poorly, with the banner image itself turned 90 degrees from the position it's best viewed at. This is the most glaring example, but not the only one.
To accomplish this, the following needs to be done:
1. All game banners need to be remade into clean fit-in-circle-inscribed-in-a-square versions with transparent backgrounds. This is the major part of the job, and not something I'm really suited to.
2. Map icon flags must be "heraldified" (I've already done this in FISH&CHIP).
3. Scene prop banners must be "heraldified" - i.e. replaced with a single heraldry-supporting banner. This is something I've already done and tested.
4. Tableau scripts must be adapted to the new heraldic system. This is a lot of boring fiddling with coordinates to ensure precise positioning of the new heraldry (which will be thrown astray as new crest meshes will obviously have different dimensions than vanilla banner meshes). Luckily, this only needs to be done once per item, and it's nothing I haven't done before.
5. Mission template triggers replacing banner scene props on all scenes need to be removed. Instead, scene prop trigger should determine what heraldry to display. Should be simple enough.
OSP needs: 2D artist.
To be frank, M&B tableau system is an incredibly versatile mechanism, however it's actual implementation for heraldry leaves a lot to be desired. My primary gripe is with the fact that the entire heraldic system is based on scene prop banners. When applying heraldry on an item, modder is essentially projecting a banner on the item's texture. This has the following major drawbacks:
1. The banner has limited size, which necessitates the use of background color, which must be hand-picked for each and every banner, and mismatched color will make your heraldic items look like ****.
2. Many banners have non-trivial graphics at the edges, especially at the top and the bottom. This imposes additional limits to how these banners (and by extension, absolutely all banners) can be placed on item's mesh. When determining the banner's position on the tableau, it is necessary to avoid any artifacts while still making sure the main part of the image is not too large or too small and positioned roughly at the chest. This is not always possible, sometimes making for quite awkward heraldry positions.
3. Banners have their own texture and shadows which look fine on banner itself. However when combined with item's tableau, banner's texture can only degrade the image quality. If the item has a very detailed texture, the heraldry may already be quite distorted, so adding extra noise is not something we should be happy for. Or the item might be polished enough so that banner's texture is actually discernible - and you definitely do not want a cloth texture to be visible on your plate armor.
4. The entire heraldic system depends on entity ranges too much. For each new banner, you must have a new mesh, a new scene prop, a new map icon flag and a new record in banner color array, and all these entities must have the same relative position to the first game banner. As the result, adding a new banner is a finicky work and in 99% cases will break savegame-compatibility with previous versions of your module.
What would be an ideal (from modder's point of view) heraldic system? It would be the one where to add a new banner, modder must add a new square mesh with his new crest, and then add his new mesh to the mod's list of banners (probably with several parameters determining how that crest should be painted). Map icon banners and scene props must use tableaus to display heraldry, thus eliminating any need to create new versions of them. All game entries (map icons, scene props, items etc) use original clean version of the crest - therefore the only texture and noise imposed on it is by the item itself.
There may be different visions of what a game crest should be composed of. At the moment my personal idea is as follows:
1. Crest itself, as a simple square mesh. Crest image must fit into a square, and all important details should fit in the circle inscribed into said square. Background should be transparent.
2. Associated background color or background texture mesh. As the background should ideally be the same for all heraldic items, it should be a simple repeated pattern.
3. Optionally, additional graphical meshes, for as long as some standard is imposed on all of them. So some crest may have an associated "bottom" decoration, which would be ignored for round shields, but displayed on the lower hem of a tabard or pavise.
Essentially, when generating the heraldic tableau, background is drawn first, then the crest itself and any optional elements are positioned and scaled in a way determined by item's texture, and finally item's tableau texture is applied to provide the necessary "look&feel" and impose item's material and defects.
Additional benefit is the increased control over heraldry positioning - right now horizontal map icon flags look pretty poorly, with the banner image itself turned 90 degrees from the position it's best viewed at. This is the most glaring example, but not the only one.
To accomplish this, the following needs to be done:
1. All game banners need to be remade into clean fit-in-circle-inscribed-in-a-square versions with transparent backgrounds. This is the major part of the job, and not something I'm really suited to.
2. Map icon flags must be "heraldified" (I've already done this in FISH&CHIP).
3. Scene prop banners must be "heraldified" - i.e. replaced with a single heraldry-supporting banner. This is something I've already done and tested.
4. Tableau scripts must be adapted to the new heraldic system. This is a lot of boring fiddling with coordinates to ensure precise positioning of the new heraldry (which will be thrown astray as new crest meshes will obviously have different dimensions than vanilla banner meshes). Luckily, this only needs to be done once per item, and it's nothing I haven't done before.
5. Mission template triggers replacing banner scene props on all scenes need to be removed. Instead, scene prop trigger should determine what heraldry to display. Should be simple enough.
Tableau Portraits
Needed: 3D artist, 2D artist.
While hanging heraldic banners around your castle is fun, hanging your own portraits (or the portraits of your spouse) would be way cooler. Obviously, it is possible with creative use of tableaus, and there are plenty of vanilla examples generating images of characters in various equipment and positions.
What needs to be done:
1. Portraits themselves need to be designed. This includes 3D design work for portait mesh, and 2D artistry (or screenshottery) to create portrait backgrounds (over which the character image will be drawn). To properly draw a tableau-generated portrait, two textures are necessary: one is a standard heraldic tableau (with solid texture for frame and semi-transparent for image part), and another with a fitting background image texture (some interior or scenery shot).
2. Once again, some fiddling with numbers to make character properly posed and positioned on the portrait.
3. Portrait trigger script, which will identify the owner of current location (or just pick some global variables) and use this information with combination of portait's scene prop instance variation numbers to determine whose portrait to draw (and possibly on what background if several background images are available for this portrait mesh).
Needed: 3D artist, 2D artist.
While hanging heraldic banners around your castle is fun, hanging your own portraits (or the portraits of your spouse) would be way cooler. Obviously, it is possible with creative use of tableaus, and there are plenty of vanilla examples generating images of characters in various equipment and positions.
What needs to be done:
1. Portraits themselves need to be designed. This includes 3D design work for portait mesh, and 2D artistry (or screenshottery) to create portrait backgrounds (over which the character image will be drawn). To properly draw a tableau-generated portrait, two textures are necessary: one is a standard heraldic tableau (with solid texture for frame and semi-transparent for image part), and another with a fitting background image texture (some interior or scenery shot).
2. Once again, some fiddling with numbers to make character properly posed and positioned on the portrait.
3. Portrait trigger script, which will identify the owner of current location (or just pick some global variables) and use this information with combination of portait's scene prop instance variation numbers to determine whose portrait to draw (and possibly on what background if several background images are available for this portrait mesh).
City Interiors.
Needed: scene editors (tons of them).
There are plenty of doors on city and castle scenes, and only a few of those doors are actually usable. This, I believe, is an oversight, and fixing it could potentially improve the "look&feel" of the game, which is frankly a bit of a bland sandbox.
Okay, so the game focus is not on actual location exploring, but that wouldn't hurt either, right?
So the idea is to go through city and castle scenes and add passages to various dummy doors, leading to some utterly non-consequential locations. There are balconies and other inaccessible areas in castles which look like they could be accessible through some doors - why not make them actually accessible? There are many businesses in towns, not just those started by player - why not let the player actually visit those businesses, even if he won't be able to do much there? Some doors could lead to private houses, which could be designed to have randomized furniture and decorations. Or to local shops selling stuff that is of no consequence to an adventurer/lord. Strip clubs... er, forget I said that.
Anyway, the sky's the limit (and the number of available entry points). Some framework would have to be established to provide support for the extra passages and entry points, some scripting for randomized houses, and some new dialog for various new locations and NPCs, but nothing unmanageable. Best thing is, once the framework is done, actual improvement of cities and castles (and maybe villages as well?) can be done location by location, as time and sceners' imagination permit.
Note that unlike previous two ideas, I can easily do all the preliminary job for this one. However I'm not sure there's going to be enough volunteer support to implement this project, and furthermore, sceners may have some specific demands to the framework and scripts which I cannot predict. So the first stage for this project would be to discuss and design the requirements for the framework, which I would then be able to provide and allow the community work to start.
Needed: scene editors (tons of them).
There are plenty of doors on city and castle scenes, and only a few of those doors are actually usable. This, I believe, is an oversight, and fixing it could potentially improve the "look&feel" of the game, which is frankly a bit of a bland sandbox.
Okay, so the game focus is not on actual location exploring, but that wouldn't hurt either, right?
So the idea is to go through city and castle scenes and add passages to various dummy doors, leading to some utterly non-consequential locations. There are balconies and other inaccessible areas in castles which look like they could be accessible through some doors - why not make them actually accessible? There are many businesses in towns, not just those started by player - why not let the player actually visit those businesses, even if he won't be able to do much there? Some doors could lead to private houses, which could be designed to have randomized furniture and decorations. Or to local shops selling stuff that is of no consequence to an adventurer/lord. Strip clubs... er, forget I said that.
Anyway, the sky's the limit (and the number of available entry points). Some framework would have to be established to provide support for the extra passages and entry points, some scripting for randomized houses, and some new dialog for various new locations and NPCs, but nothing unmanageable. Best thing is, once the framework is done, actual improvement of cities and castles (and maybe villages as well?) can be done location by location, as time and sceners' imagination permit.
Note that unlike previous two ideas, I can easily do all the preliminary job for this one. However I'm not sure there's going to be enough volunteer support to implement this project, and furthermore, sceners may have some specific demands to the framework and scripts which I cannot predict. So the first stage for this project would be to discuss and design the requirements for the framework, which I would then be able to provide and allow the community work to start.
