You know, I had this same idea a few times in the past but since I'm not a pro I thought it would just sound like some random newbie asking for help. lol
2 obvious suggestions: RPG Maker DS and Uncharted Waters Rota Nova DS. I don't even need to elaborate on why they're good picks. That said...
I personally think that if we are to invest our blood into it, it must be a game that 1-) has enough depth/replay value to make it worth but at the same time 2-) it shouldn't be extremely complex (like 3D stuff) so even newbies can "walk the walk" and get involved in the proccess if they wish to (or at least can learn from it after all is said and done). And finally, 3-) it would be ideal that this game uses the main DS functionality - the touch screen - in a decent way as I feel that this justifies choosing a DS game as the target of this project.
Considering this, my personal suggestion is RAGNAROK DS.
This game is a stylus-focused ARPG based on the popular MMORPG with millions of fans. Every other attempt to create any derivative work on this franchise was a let down and not what people wanted: Tactics on PSP, Odyssey on Vita, Valkyrie on Android, etc. Surpsiringly, the DS exclusive is VERY reminiscent of the original game which is impressive. With a few doable hacking tweaks, it does indeed have an absurd potential.
As far as the GFX are concerned, I know it's newbie friendly enough because I have hacked them myself (I managed to swap all Class sprites for their female counterparts, female hairs and other colours work too). The in-game GFX are the exact same sprites from the MMO except with lower quality. Since all of those sprites are easily available online that resource is already there to all modders who may wish to insert famous enemies, Class sprites or cool looking hats from the MMO.
The coding was apparently done in Lua. If we can reverse engineer some scripts we can essentially recode the game events. But even if we fail at this, the improvements we can accomplish via hex editing enemy stats and spawns is worth pursuing considering the game already has enough depth to make it fun, all it lacks is enough content (read: more challenging enemies) or a better distribution of the contents already available.
On the musical aspect, Ragnarok is well known to have some awesome music...and some trashy ones. Unfortunately, the highly replayable post-game dungeon in this game has a bad one. The story mode final boss has the awesome 1 and it's only played ONCE. This is a good place to start.
For those not familiar with the MMO (or the DS version), here is a quick look at how this ARPG works and why it has actual depth despite simple and functional controls:
After lvling up as the Novice Class a little, you choose between Swordsman, Thief, Mage, Taekwon, Archer, Acolyte, Merchant. Each of these classes has 1 advanced form, ie: Merchant becomes Blacksmith (and 2 secret, advanced unlockable classes). You can spend your stat points and skill points earned via lvl ups freely, but it's not possible to have all skills or max all stats by the time you reach max lvl. That's where part of the depth comes from: every class has multiple stat/skill/equipment builds to experiment and have fun with.
After reaching a certain lvl, you can Rebirth. It basically resets your character in exchange for bonus stat points and eventually bonus skill points. So after a very long cycle of Rebirths it's possible to have a max stat/skills character of any Class. Since you can recruit and train NPCs of any Class, this process becomes easier and faster each time as you can have a high lvl support healer helping you, etc.
The Mirage Tower is a semi-random dungeon with 50 floors and a chance to get the best equipment in the game. It's also the only multiplayer part of the game but it be soloed. Unfortunately, after finishing the main story this is the only place in the game that still offers some challenge and good rewards.
But the game has all of the structure to fix this already in place if we decide to mod it:
- There are repeatable quests, except their challenge/reward stops being meaningful (easy to fix by replacing the enemy/reward);
- There is an open world but unlike the MMO, boss monsters don't respawn and eventually it's not worth exploring anymore (fixed by changing monster spawns/drops/their stats);
- The NPC partner system is awesome but NPCs can't go into Mirage Tower in adition to the issue mentioned above (if open world exploration and repeatable quests become valuable again this point is fixed);
- Although equipment variation is already nice, redundancies exist so they can be modded into useful stuff without any negative impact;
- The point above also applies to some monster "Cards" that grant special effects by being allotaced into slotted equipment. ie: The anti-freeze card is awesome on the MMO but useless here. We can fix this by replacing the card effect for something useful OR by changing enemy spawn points+their stats in such a way that anti-freeze becomes a serious consideration (example: by putting "Marin" monsters together with Sonic Blow users and buffing their stats considerably);
- There is a cool game mechanic of customer>shops relationship that unlocks better equipment and goods as you buy more from each shop. But once maxed, it becomes irrelevant. Since this is ultimately accomplished via the final scripted quest unlocked after spending X money, we can edit whatever SRAM value is used to upgrade the shop and give players some rare reward from the already existing Mirage Tower floor clear drops (then reseting player relationship with shp owner), effectively circunventing the dead end on this cool sub-system (and giving player rewards for more money loops from buying shop potions and stuff);
That last point actually applies to everything in the game, really. The Mirage Tower can be considered a small, replayable game on its own. Floor challenges are designed to test specific things like player ASPD (atk speed), monster killing capacity (monsters of different races and sizes that can be affected by DMG mods) VS boss killing capacity (if you equip a weapon strong against the boss you're sacrificing general monster killing speed), etc. So the "invisible" objective in the Tower is to be as efficient and quick as possible in order to have more chances at rare equipment drops.
Basically what I'm saying is that all game design concepts used in the Tower can be applied elsewhere and they're solid enough to fesh out the rest of the game world via hex edit swaps on enemy spawn points/base stats/drops even without reversing the Lua scripts. If reversing them is possible then we could try more advanced stuff like reusing the Tower semi-random floor generation on other maps too, theoretically expanding the game world without adding extra GFX. Or better yet, hacking the touch-screen commands to create new spells/skills (probably too hard but at the very least we could mess with charge times/delays/cooldowns for a similar effect).
And if someone is wondering why people would play RO DS (via emulators) instead of playing the actual MMO instead, well, that's the thing. The majority of the players who quitted RO did so for basically only 2 reasons: it takes too much time to get anywhere in the game and/or they didn't like the Renewal mechanics. So even disregarding the actual DS portability (and being offline), RO DS has the potential to deliver on both ends even with simple mods if the community manages to make this possible. The faster lvling loop is already integrated anyways, it's possible to reach end-game levels in 1 afternoon with a few good NPC partners so Rebirthing doesn't feel like a chore or unrewarding grinding that takes months to accomplish in the MMO. It's not a big deal to swap classes in RO DS so being creative is fun.