LibRetro hard forks: branding brainstorming

What forks specifically are we talking about? Maybe if we list them all we can find/think of a common pattern?

The problem with “lr-(system name)” is the redundancy as has been said. I like the idea of codifying the speed of emulation (fastest/fast) in the core name.

From what I understand one of the main drives behind offering so many cores is portability and make RA/LibRetro available on a wide range of devices with very different hardware (and software) configurations. Could we thus try to hint at the type of device/hardware requirement the core was developed for in the name?

This might sounds over-complicated but we could have 3 categories (3 tiers! just like in the brawlers of yore!). Each category would have an easily identifiable core suffix associated to it. We could list a few platforms under each category somewhere in the Wiki/Doc (or expand the info file for easy identification from the menu) to guide the user.

For example:

Low power (low memory?) devices Example: Raspberry Pi and clone SoCs Intel Atom SoCs Pre-201x smartphones and tablets (Exynos xxXx + Mali 400) Wii

Suffix: lr-(system)-nimble

lr-snes-nimble (snes9x2002) lr-nes-nimble (quicknes) lr-arcade-nimble (mame2000) lr-gba-nimble (gpsp) etc…

Mid-Range devices ???

Suffix: lr-(system)-balanced lr-n64-balanced (Mupen64+) lr-snes-balanced (bsnes-mercury)

Highend devices nVidia Shield OnePlus 3 / Nexus … x86 PCs with Intel HD4000 and above

lr-(system)-mighty lr-psx-mighty (beetle HW) lr-n64-mighty (paraLLel) lr-wiigc-mighty (Dolphin)

Obviously the big issue with this approach is that some emulators/sytems (nes, gameboy etc…) run full speed even on modest hardware… I also understand the team has been reluctant to suggest some cores above others and the moment we start categorizing some may take offense?

1 Like

In the case of fast cores omitting the suffix would be fine I think.

The problem with this is that there are multiple cores for the same system (Fceumm, Nestopia, etc).

Why not simply select the most compatible core or the core with the most active development for each system, propagate this to “libretro-$system” and then add the original names to the other cores (i.e. the cores that have not gained “libretro-” status) of that system? Or maybe even drop them from the buildbot? It may be rather confusing for non-experienced users to see multiple cores for a system in the Core Updater…

I think just add a sequential numbering scheme to it to make it easy for everyone. For example. Lib-nes001, lib-nes002, lib-nes003. Each respectively represents nestopia, fceum, quicknes. This permanent numbering scheme would be applied for the all future cores of that system.

Obviously, the same can be done for other consoles with several cores.

1 Like