MAMEWorld Forums - EmuChat - Mess with the RetroArch + MAME 0.187 last night.

> > I'm not really sure what you expect from RA, it's just butchered up junk that
> should
> > be avoided.
>
> From an end-user perspective, RetroArch (and by association, libretro) is also - how
> can I say this kindly - complete and utter dogshit.
>
> The concept of a unifying frontend for disparate emulators is great, and I'll
> absolutely give it credit for that. It even looks nice and presents a largely-usable
> UI. But the implementation is a fresh, fat, steamy dog egg lying in wait to be trod
> upon. More on that below.
>
> > Unfortunately I guess on mobile you have no choice right now.
>
> Or on just about any low-end platform (RasPi, I'm looking at you). It's as though RA
> is being seen as the ultimate answer to how to wrapper multiple emulators in such a
> way that anyone barely registering a pulse can use them without having to think about
> it, but with total disregard for the fact that adapting the emulators to work with
> the frontend platform generally ends up hacking them into a craptacular mess that
> contributes to sub-par user experiences with said emulators.
>
> What I see as the core of the problem is that RA (by way of libretro) promises
> bitchin' UI and UX. And in many ways, it does deliver that. But having to graft
> libretro into emulators that were never intended to use it in the first place leads
> to the aforementioned craptacular messes and sub-par experiences.
>
> This is essentially what I don't like about RA: the idea that it *is* 'the solution'.
> Thing is, when it comes to frontends, there are 32767 other solutions out there,
> virtually none of which require making major source changes and building against
> external libraries to integrate them with {insert emulator here}. It also obfuscates
> large chunks of the emulator being used from the end user, which I feel is a
> disservice to the people who worked on that emulator - you should know what it is
> that you're using so that you know who to give credit to for the hard work that went
> into it.
>
> If the libretro API was something that emulator authors sought to support natively,
> fine. I have absolutely no problem with that; it's an option that *should* be
> available. But forking source, grafting code into it that the architecture was never
> meant to be able to handle in the first place, then presenting a crappy end product
> to users and giving the excuse of, 'it's the other guy's fault, honest, my frontend
> library works fine, go talk to them' is just pathetic.
>
> My gripes with emulation on the RasPi have been stated here to some extent, and I
> still agree with them. But I don't see why - and particularly on low-end platforms -
> performance should be traded for eye-candy convenience.

and despite all this it's becoming massively popular, even on Windows, which is what I thought would end up happening as soon as they got a more slick marketing team on-board, which appears to have happened.

I'd still consider it one of the biggest threats we've seen to proper emulation in a long time.