I wrote a big wall of text that nobody wants to read but deleted it because it boils down to this: we have limited manpower and keeping any given libretro core in lockstep with a (sometimes hostile) upstream is tertiary–at best–to our actual goals, which are focused on growing the libretro ecosystem and making a good gameplay experience for users.
What’s interesting to me is how you define those goals, in practice. I mean, what is currently lacking? This is not a critic on the mindset of the project, but I’m grasping for something… tangible. Once again, I’m not saying some parts of RetroArch can’t be improved, that the project has reached complete maturity so to speak, but I cannot think of anything that I would need more than say, a stable pcsx2 core. 
Well, the biggest thing to keep in mind in that respect is that we’re a small team of volunteers with varying skillsets, so time spent on one thing doesn’t imply time taken away from something else.
I could spend all day banging my head against trying to update (or debug) pcsx2-libretro and get exactly nowhere (because the issues with that core are not within my skillset to investigate, much less fix), or I can spend my day writing shader code and actually get somewhere. That is, if I spent 2 hrs updating the slang repo to play nicer with the new append/prepend feature, that doesn’t mean 2 hrs were taken away from pcsx2-libretro because I was never going to do shit there anyway.
5 Likes
Personally I always felt Retroarch is universally accepted as the go-to reference and easier setup, specially for frontends, EmulationStation and Launchbox just to name two encourages RA over standalone not only because it’s easy to setup, provided the user has a basic understanding on how to operate a menu, since there are those who get lost even on kodi. But we all know the internet is a place where people has a voice whereas it wasn’t possible before, behind a screen and a keyboard, even those who just want to rock the boat to their own amusement. Some people need drama to fill the void. I honestly find Hunterk’s participation here astounding, what was that, over 10 years still promptly replying with expertise and patience, probably more? As a side note, I rewatched the Analog Frontiers mini series from My Life in Gaming and I marked the part where Retroarch is mentioned as the template for what a service such as Netflix or Spotify could be in the near future, have you guys watched this series? I’d say this is the best this channel has produced so far, along with M2 Complete Works.
As far as PSP, GameCube, Wii, etc, yes, it’s probably too much work to maintain them as updated as frequently as they get and in certain cases RA cores will do a better job, specially for 4:3 locked games, but that could change. From Dreamcast and below, it’s a no-brainer that setting RA once is the easiest and best way, specially for those who wants to enhance their look, because shaders. But there are specific technicalities I’m not equipped to explain which makes it the best way to play these games, at least in my experience, I hardly even think of trying a standalone for the above systems (DC and below), if ever, when I try, I usually feel the difference.
2 Likes
Fair. I guess this brings to mine the implication of those standalone emulators. Of course I’m not gonna throw shade at them, but I have to wonder why a developer would want to provide a standalone when libretro exists. As a developer myself I’d rather leverage an existing ecosystem that already handles file system, inputs etc. That’s a huge chunk of work I wouldn’t need to care about. I suppose they see limitations in it that are not obvious without looking at the code.
I wish I could contribute to all of that but like you said, emulator architecture is a very specific skill. Even if cores were written in a language I have more experience with, I probably wouldn’t have the slightest idea what to do.
2 Likes
There are indeed some libretro-first/only emulators, but there are also plenty of valid reasons not to want to mess with it.
For one, the API is focused on playing games rather than, for example, debugging them, which is something emudevs often want/need to do. Many devs–emu or otherwise–also like to be in control of as much of the software stack as possible, and/or they may not want to mess with the complexity of RetroArch (though there are other frontends for libretro around, of course) just to play/test their own emu.
It’s also only fair to mention that libretro as an API is a Procrustean bed: frontends can bring a lot of benefits, but devs need to run through the API to work with them, and the API is simple/basic on purpose but that also means it has significant limitations. For example, the retropad input abstract is just 18 digital buttons and 2 analog axes, so if a core needs more than that (e.g., Jaguar and Intellivision’s many keypad-like buttons or n3DS’ dual analog sticks plus dual-axis motion control) they have to drop that functionality altogether, work around it (using a dedicated ‘shift’ button to get more buttons, for example) or deal with it on their own outside of the API (e.g., dolphin-libretro includes libusb functionality to sync wiimotes to a dolphinbar; this adds an external dependency, which we don’t love, but it allows the motion control functionality despite of the limitations of the API).
Interacting with hardware rendering APIs and dynamic recompilers via libretro adds additional hassles vs standalone development, as well. E.g., all dynarecs have to be PIC-compliant, which many older dynarecs are not.
So, I won’t pretend it’s all sunshine and rainbows working with libretro/RetroArch (and these are just technical issues; there are also plenty of personal clashes and disagreements that can and do occur, as is common with open source development), but many developers/teams have decided the benefits outweigh the hassles, which is great.
4 Likes
Thanks for taking the time to paint a clear picture on the situation. From what you’re saying it’s probably unlikely that this will ever change too, unless a large amount of cash can be fueled in libretro or in either of the emulators. I know some open-source projects end up being funded in large parts by private companies (OBS for instance) but since we’re talking about emulation, I hardly see that happening.
1 Like
You don’t know anything, if you look for any emulator in the prehistoric era, Hunterk appears. 
If anything impresses RetroArch is that it has a range of professionals doing magic, sometimes with sweets, sometimes with garlic, but they are always here, persevering, not giving up. Is a FOSS project at the level of the largest, Blender, Ubuntu…
I can only feel respect and admiration for them. Thank you for your hard work. 
PS: I get lost using Kodi ¬¬
We are living in an era where people don’t dig deep, and in the collective the idea persists that RetroArch is a frontend or a multi-emulator and they assume that it works or should work the same and that cores are a direct port.
In reality the LibRetro-RetroArch eco-system is much more, (as the snap says) “multi engine”. RA is currently a reference in the emulation scenario, it would be timely for the project to be presented with a clearer concept.
…
What if you create a bounty to upgrade those cores and every time someone asks for the newest, you send them there? Maybe someone will get motivated and upgrade them.
2 Likes
I think the biggest problem is that, for whatever reason, many emu devs seem to oppose RetroArch. MAME devs in particular absolutely hate it and never miss the chance to remind it to everyone whenever the name “RetroArch” appears, more specifically in the Emulation Reddit forum.
So i feel like there’s a climate of antagonism between RA and the emulation community. How can more cores be created or maintained if that’s the case?
1 Like
Well, it’s Reddit, lol. MAME had always drama around itself, same as the romhacking “community” , cracking scene etc. Reasons slightly differ, but people need to remember that not everyone of the dedicated folks who bring the good stuff are necessarly well-adjusted people living happy family lives for starters. This is not to say that plenty of the involved people are nuts (although they well may be). It is simply in the nature of things that once you invest a lot of your energy into a private thing that you do, chances are you have strong opinions about it.
2 Likes
Despite all of the visible scene drama, we get along well with many/most developers and teams. You just won’t ever see/hear about that because drama makes a spicier meatball for scenesters to chew over.
The majority of drama comes from people who want to exert control over retrogaming/emulation, sometimes for clout but usually related to commercialization efforts (or, frequently, both surreptitiously). Companies regularly reach out to emu devs and contract them to port emus to their devices/products, but having RetroArch/libretro around on every platform/architecture for free takes a lot of wind out of those sails (and money off the table; I know this for a fact because companies reach out to me/us about it constantly).
The same thing happened with for-pay Android emus: people still pay/paid for premium emus that provide specific benefits (like Drastic’s mind-blowing performance, or some of the handheld emus that do link cable over the internet for trading pokemans, or even just emus that provide a premium, native experience, which we admittedly lack), but the market for low-effort for-pay ports pretty much dried up once we released there for free, and a number of folks got very upset by that.
When people complain about RetroArch and money/Patreon, this is what they’re really complaining about. Our Patreon brings in <$15k/yr, which would be below the poverty line for a single individual (let alone an entire team) even if it all went into that person’s pocket instead of almost entirely going to fixed costs (server infrastructure, hardware, hosting, bandwidth) each month. We’re not getting rich (I’ve never made a dime off RA/LR in all this time), but just by existing we’ve made it harder for others to get rich, too, and that breeds a lot of simmering animosity.
5 Likes
One thing that comes to mind when we talk about classic games such as DOOM and basically any console pre PS3 era is that now we have availability of everything almost all of us dreamed as kids, since we were limited to what we could try and play. Today these games are more appreciated than any other time, even in their heyday. With that, comes Retroarch that is possibly the main choice to play these games. Sometimes we come across those hate posts you mentioned, but I wonder who really cares or buy into this, I mean, I have my own opinions and experiences, that’s more valid to me. With all that, when I use MAME to play classic arcade games, it’s the RA core that is the best option for me and FBNEO if I want to reduce input lag in some games and in a handheld.
2 Likes
No doubt every element influences, a particular personality, disagreements, professional jealousy, ego and even the economic issue that has always been treated as a taboo. But from my point of view, I think the community has a great responsibility in all this, more specifically the ignorant user, and I include myself because it happened to me, when I heard of RetroArch I thought “another frontend and steals the emulators” until I met him.
The average user does not investigate, he gets carried away by the first thing he hears or reads. As in the recent case of the PSX emulator, a “let’s burn labruja” crowd is created, without knowing anything. Most of us are very ignorant that they don’t earn anything, they do it because they are passionate about it, you are working for free. It’s practically a gift what they are giving us.
All this accumulates like a snowball, it hurts, that’s why the presentation, the documents, even any material that serves as publicity are important. Although, it is a bit difficult to ask a person who invests so much time to also do the documents (here I start to feel guilty).
1 Like
Are you talking about DuckStation? I’ve read some things, but I don’t know what to make of it all.
Yes, I precisely omitted the name so as not to go into details. That’s behind us, we have to focus on what lies ahead.
I wasn’t my intention to wake up old things, but I never knew about that until you mentioned it. Trying to sweep things under the rug never helps anyone I believe. You cannot move forward if you can’t acknowledge the past.
In general I thin the libretro team is doing a fantastic job. Finding the balance between front end development and core development is hard.
I agree that dolphin, ppsspp are pretty outdated. But maintaining all cores is hard work. In my job I have to maintain hardware revisions and different software brachnes for each revision depending on the region where the hardware is used. It is a nightmare. The key was to reduce the hardware revisions to a bare minimum and keep the software branches small as possible. So I can imagine what kind of work that is with regard to the cores.
Maybe the solution is to focus on main cores for each gaming hardware. Means for nes its mesen, for snes its snes9x for n64 its mupen-nx etc So the ressources can be used more efficient. But as said I think the team is already doing a great job.
With regard to standalone va libretro. I would also like to switch completely to libretro but there are some features which are currently not possible in libretro. For example the movement assignment of the wiimote in dolphin, or the mapping of touchscreen buttons in citra.
PPSSPP is up to date, although you are right that the Dolphin core hasn’t had an update in a while. I don’t use the Dolphin core, though not because it hadn’t been updated but because I have a fairly obscure requirement from the controllers that Retroarch is unlikely to support.
The PCSX2 core probably isn’t a fair comparison to the other cores since it’s still in Beta.
1 Like
What other unofficial cores exist?
Maybe the solution is to focus on main cores for each gaming hardware. Means for nes its mesen, for snes its snes9x for n64 its mupen-nx etc So the ressources can be used more efficient.
It means agreeing on which core should be the main focus. You name snes9x, but I believe it’s not as accurate as bsnes for instance. Reaching a consensus on this topic is complicated.