Split the Amiga games database into OCS/ECS/AGA?

[Kodoichi]

The Games [ADF] database is so huge that I had to split it into three folders on my hard drive. When I use RomCenter or want to copy files it's faster to load smaller folders than a huge one. Recently I found a download archive that had AGA games in a separate folder.

Would that be something for Tosec? Split the games database into the chipsets OCS, ECS and AGA. Keep any flags like (AGA), [KS1.2] or [KS3.1] in the filename. If a game works with multiple chipsets (like ECS-AGA), put it in the lowest common denominator (ECS database). This would also be useful to load games in emulators, without having to guess the correct configuration.

[NLS]

No I don't agree.
TOSEC is not made with these things in mind.
A good front-end that also keeps the metadata that interest you (OS supported, chipset supported) to filter, is what you need.

I agree that the folder is huge, but this again is an issue of your storage speed etc.
I would prefer splitting by letter maybe.

[Kodoichi]

Explain why there are separate databases for public domain or educational games then, which is basically just metadata too and might as well be added to the "Games - ADF" database with a corresponding tag. Tosec used to split all commercial games by alphabet, then one day all games were put together into the current big database. 26 (+ numbers) databases seem too much when I think of it now, but three for each chipset would be manageable.

[NLS]

Everything could be metadata, it is an arbitrary decision how to split.
I don't see the way you mention very viable because it would mean re-scanning every single to find out what KS and chipset it needs.
I don't see it happening.

[TKaos]

Have to agree with NLS, it's also just more work for the guys maintaining the DATs to split them and then even more work to create all the new DATs for releases.

Also just splitting the games folder wouldn't be an option, either all or nothing imo.

[Kodoichi]

Another argument for splitting:

Tosec has separate databases for the Commodore 64 and 128 (and some more Cxx), as well as MSX and MSX2. The newer systems can play most games released for the older ones too. But in the Amiga database, everything is mixed together, even though A1000 (OCS), A500 (ECS) and A1200 (AGA) are different systems.

Rescanning shouldn't be much of a problem as most AGA games already have that tag in the filename.

[NLS]

Another argument for splitting:

Tosec has separate databases for the Commodore 64 and 128 (and some more Cxx), as well as MSX and MSX2. The newer systems can play most games released for the older ones too. But in the Amiga database, everything is mixed together, even though A1000 (OCS), A500 (ECS) and A1200 (AGA) are different systems.

Rescanning shouldn't be much of a problem as most AGA games already have that tag in the filename.

Actually, Amiga is way more complex, so I am not sure splitting would help.
You see:
- 99% of OCS works on ECS (and the separation is not A1000 vs A500 as there are A500 with OCS too).
- Most ECS actually works on OCS too.
- 80%+ of ECS works on AGA.
- AGA doesn't work on older chipsets BUT many games claiming "AGA" are not really AGA, just declaring their compatibility to it, or (few times) enabling extra stuff because of AGA... and in reality because of more chip RAM not really because of AGA.
- And the secret sauce. Kickstart version. Many times it matters.

So I would say the classification of games is a matter of external databases (or if we ever implement TOSEC-META ...meaning metadata, not the new trend word), not splitting.

Actually MOST things nowadays that need to be classified, because it is a complex task of interconnecting parameters, it is a matter of proper tags, not splitting some way.