Some stuff

[Amunak]

because it keeps the style consistent. the same reason why I like the low resolution graphics. right now, because of the limited palette, you are so limited in what you are able to do that almost anything you do will end up looking similar to the original, whether you want it or not. that's not a bad thing.

It's neither a good thing. If you want, you can make a mess of the game even with a 4-bit palette. Or you can improve it with smoother gradients with a full (32-bit) or at least 24-bit palette. Since openxcom will probably always (due to the copyright it has to I think?) come without graphics (actually using the "bought original" graphics), this won't be an issue. And then... let the modders do whatever they can.
Well... I see that I'm actually repeating others. NVM.

And speaking of LIMBO.. It might use only shades of white, but I'm pretty sure that they use both alpha and also whole scale (probably 255 colors) as a palette
Daiky, nice work on the palette. Maybe we could choose some "basic colors" xcom uses and advise the use of those only? I mean making something like this.

[luke83]

I honestly think this discussion would be better after TFTD has been fully supported ( Are we going to call that V2.0 ? ). 

[SupSuper]

Boy these modding discussions sure keep coming and going. Let me clarify the technical hurdles we have to overcome to implement this stuff:

- Support new image formats: Technically this isn't too hard. We just need to add image loaders for different formats and have them properly converted to 8bpp, which SDL can mostly handle on its own. It's just there is currently no easy way to specify those different files to load, the resource manager is pretty hardcoded (just loads some set files at the start) and would need some rewriting.

- Support 24bit: Ok now you have to get rid of palettes too. This sounds simple, just convert old graphics to 24bit right? Except... there's a lot of stuff set up around palettes. Screen background colorization, geoscape/battlescape shading, lots of drawing tricks based around 8bpp and lots of stuff drawn on-the-fly like windows, buttons, etc. Lots of stuff that would need to be either pre-generated or need new drawing solutions in 24bpp while preserving the original look.

- Support higher resolutions: This is probably the hardest. First of all, the engine gets away with a lot because it's all actually drawn at 320x200 and scaled up to whatever resolution, but once you actually need native high resolution, it all starts to chug with the increased workload, due to a messy mix of SDL software rendering, dirty rendering tricks, unoptimized resource management, etc. Second, you then have to figure out how everything is gonna fit on the new resolution, given it's all designed for 320x200. Does the UI just stay at its size centered on the screen? Make it flexible with the resolution and increase the viewport? What about elements that aren't flexible like original sprites (background images), do they just get resized? How does this all mix with new graphics? What about preserving the original look? Etc etc etc. This would pretty much require an overhaul everywhere and is very v1.0+.

So none of this stuff is impossible, it'll just take a long while, but it is planned. Be patient.

P.S.

Also... Someone already mentioned it - there's no reason for converting the original data. I think that it would be best to have folder "original" in the data directory where would be all the original resources. Idea is that any other folders would be displayed in a menu ingame and you could chose if you want to have them active (and presumably the order of loading) so that you could have like more "texture packs" or "sound packs". And if some file isn't in these custom-made folders, it would fall back to the original. What do you think? It might be pesky to implement it, but it would add great and simple extensibility to the game.

"Data packs" are the plan for moddability in the future.

[Volutar]

My comment on "Support 24bit":
Screen background colorization is a palette trick, taken from original xcom. Issue is there are number of screens which have same background, but with slightly different palette. So this will require either number of differently colorized screens, or simple colorizing them while loading (those loading functions anyways will get palette variables, just to get them converted into 24bit). Second is much easier and more "natural".
Geoscape/battlescape shading can be redone, though shading might become different (if there will be simple 24bit lightness control over depalettized 24bit sprites). While geoscape shading probably become more "smooth" and nice, battlescape lighting may become too altered, and probably will require additional tweaking (to keep shaded parts "in-palette") or just depalettizing-on-the-fly.
Interception window probably will require for little tweaking.
Windows/buttons and other "lots of stuff" is not a problem of any kind at all.

[SupSuper]

Windows/buttons and other "lots of stuff" is not a problem of any kind at all.

Well UI also uses lots of tricks like text colorization, inverted buttons, etc. And drawing on-the-fly is a bad approach in the long term since people will wanna replace those elements with their own graphics too, plus pixel manipulation isn't hardware-accelerated.

[Volutar]

SupSuper, what "long-term" are you talking about, when everything made with SDL and "pixel" approach is not anyhow "long" at all! If we getting everything working like in original, but with some kind of modding, and paying attention mainly on gameplay replication, planning using pixel-SDL in "long term" is quite a silly idea. As much as hardware-accelerated SDL

[luke83]

For what is it worth, i like keeping all the same map formats, graphic types as the original, i dont really want anything Fancy until both xcom games are supported. Just my 2 cents

[hsbckb]

For what is it worth, i like keeping all the same map formats, graphic types as the original, i dont really want anything Fancy until both xcom games are supported. Just my 2 cents

Yes,  running the two X-Com games without trouble is the first priority. Others major improvement is welcome but after 1.0

[Volutar]

Yeah running xcom1 game without any modification. Without new maps or new units or new weapon. The only things allowed - is UI improvements, which make management easier.

[luke83]

Volutar , you can already add new maps without any trouble and have RANDOM ufo of a given type, this is why i am mostly focusing my efforts on these points right now. It would be good to sneak in new units OR weapons to increase game variety but  no its not a must have before version 1.0 , Mostly we want to know What is planned ( and achievible without coding skills) and is there anyway we can test it?

  However it is important to find some sort of balance with the Modders, I can think of nothing worse than writing a program like this and have no-one interested in using it as they can simply play the original in DOSBOX. You still need to provide people with something here that they cant easily do using UFO EXTENDER or any off the alternatives and so far i think Suspuper and Daiky have done a great job giving us things to keep us modders going whilst still keeping the project on the moving towards the main objective.

[Daiky]

"You can simply play the original in DOSBOX" ... not :p Well, I can't anymore. If you are used to unlimited savegames, better pathfinding, possibility to abort a unit walking, etc... It's hard to go back Those for me are the main strengths of openxcom actually. Those things can't be done with UFO Extender.

Modding is nice, but it totally depends on the content that is created. It was possible to do modding for the original game too, but it's never been that popular... People have been able to create xcom maps and sprites already for more than 10 years, you would expect a giant repository of new aliens, maps, objects, but I don't think I can't find more than a handful.

[luke83]

Modding for the original game is a Pain because of the MCD limits per map. Also it not easy to have Large sets of maps because it ha set numbers of mapblock sets per type, in Openxcom i can have 10 version of the same "Road" all with slight variations and the result are  interesting map variety. Other than those 2 points i am not sure why there is not giant repository of people work and besides the stuff from HOBBES ( which is usually very good quality) , most of the work from others is rather second rate 
  I can tell you so far i have HELPED answer 5 peoples questions directly for how to mod openxcom but so far only one has shown any work, maybe other people cant commit the amount of time required to do these mods, i can easily spend 3 hours on 1 UFO so you can imagine how long a Set takes. My Hope is my websites How To section will create a few more modders as there is very little help files on most of the software available. https://openxcommods.weebly.com/how-to.html

  I want to create a completely New version of Xcom one day for the OpenXcom community, new story, new aliens, new UFOs,new weapons, set after Apocalypse...First things first, i need to create a reasonable story ontop of the Xcom universe, learn all the software required to make the art assets ( mostly done), workout what each alien race should look like, work out what human soilders i can recruit ( Human, Sectoid hybrids & Mutants) the biggest problem to my planned mod will be the changes i want to make in Geoscape which will require programs skills
  Because this "New game" idea is such a big task, i will just work towards it  piece by piece, so if i never reach the ultimate goal i still have new items for anyone wanting to add to there normal game

[Amunak]

luke83: We can still hope for some very nice modding API or something. If LUA/Python are at least considered, it could mean very easy but complex modding opportunities.

[michal]

gchevallereau already was experimentiong with python scripting:

https://openxcom.org/forum/index.php/topic,360.0.html