MAMEWorld Forums - Welcome to the War Room - Is MAME Killing it´s frontends?

> > > I prefer program agnostic frontends. I don't use Kodi because it's not truly
> > generic.
> > > It's tailored towards media files and streams, and it isn't really intended for
> > video
> > > games.
> > >
> > > What's wrong with a frontend where you create your own lists, and each item in
> the
> > > list is something like:
> > >
> > > 1) binary location
> > >
> > > 2) args
> > >
> > > 3) label
> > >
> > > Is that so difficult? That's all you need. Anything you can do with MAME you can
> do
> > > with command line args anyway, so you immediately support anything MAME's ever
> > going
> > > to do once you support arbitrary args in a generic launcher.
> >
> > because it needs to be more than a frontend.
> >
> > the idea that MAME is a 'launch once' thing is an outdated concept.
> >
> > if you're running a computer / console inside in for example you need a good
> > interface for managing the tapes / disks, inserting new ones, creating blank ones,
> > rewinding, fast forwarding etc.
> >
> > designing a fancy frontend 'launcher' or anything similar to what you describe has
> > limited use as we actually need users to be able to pull things up *while the
> > emulation is running* so creating reusable menus etc. and a good interface is
> > important.
> >
> > I'm not saying what you describe is useless, there are situations where it would be
> > useful, but it doesn't negate the need for the internal frontend / interface etc.
> > having decent functionality, because regardless of how MAME is launched there are
> use
> > cases where you'll need it, even if it's just something 'simple' like inserting the
> > correct game disks when requested; you can't really exit MAME, and reload MAME with
> a
> > different launch parameter every time you want to do that ;-)
>
> You're basically talking about a sophisticated UI to simulate physical interaction
> with devices possibly leveraging things like the layout system, which simulates the
> physical interface objects of a machine, but I think you're over-estimating how
> useful that actually is. Commandline args could provision all the individual disks
> you're ever going to switch to in that instance of the program anyway, and a simple
> input binding could activate switching between those disks (the "H" key switches to
> disk 1, the "J" key switches to disk 2, etc.). And if you're doing something really
> complicated such as basically emulating a Windows environment and wanting to do real
> work in it (changing an arbitrary number of devices with an arbitrary number of data
> sources), at that point, why can't you just use a generic Windows or Linux UI to
> select different disks and so on? Standalone X86000 emulators already do this; it's
> already a solved problem. For this kind of a use case, a frontend wouldn't make sense
> anyway; frontends are meant to be simplified entry points to running software. This
> is the exact scenario where you WOULDN'T be swapping an arbitrary number of devices.
>
> It's no different than the "Reset machine" input in MAME. Yes, when you run a Genesis
> game in MAME, you don't have a physical Genesis reset button to press, so you
> abstract it away into a hotkey. Why does switching disks or doing anything else need
> to be different? Why is "fast forwarding" a tape different than any other input I
> could just map to a keyboard or mouse control? I don't see a compelling argument for
> not being able to do everything from the command line and simply controlling physical
> actions with input mappings. I think you're over-complicating things, and I think
> you're overestimating what the majority of MAME users are actually going to do with
> the software. Really, nobody cares about archaic computer emulation in MAME, which is
> why every non-developer instantly falls asleep whenever someone starts talking about
> Apple IIe emulation, etc.. Most people just use MAME to play games. The dirty secret
> about MAME is that the development team and user base have almost completely
> different and arguably incompatible interests; it just so happens that the interests
> overlapped enough to get some popular video games running in it, and that made it
> popular, but there's a harsh line between what developers are interested in and what
> the majority of its users are interested in. The reason garbage like Hyperspin is
> popular is because it caters exactly to what a lot of the users want: a visually
> flashy, functionally simplified entry point to running games. You're a very
> intelligent person who knows a lot about computers, and I think you need to step back
> and realize that most of us are really dumb and just want a simple way to play games
> without having to understand devices and software lists and so on.

I'm sure that's understood.

However, just because something is what the majority like doesn't mean it's the right thing to do.

As you say, that might sound harsh, but integrity etc. is important, the project needs to make sure it can handle the more extreme use cases well, not cater entirely for the most simple ones; the true value comes from the ability to handle unusual / difficult situations and the people who do put in the effort to learn and understand, and likewise those who have a strong need for them appreciate that greatly.

But yes, the harsh truth is that it's more important to do things right for the 10 people who really need something than do what 10000 people *think* is right. Developing a piece of software like MAME based on 'popularity' is a road to ruin.

It's a very intentional decision, much as things like wanting to create reusable components is rather than have every machine using it's own implementation of something, it's own timing framework etc. etc. We care greatly about that, even if again, users don't and are quite happy to choose an 'emulator' that just glues other ones together without making any attempt to properly share emulation code between systems (and therefore lacks definitive and easily reusable versions of cores etc.) General users don't see the importance of such things at all, real emulation developers do, so yes, again you get a big disconnect between the two groups.

Likewise, stuff like the Fruit Machines are unpopular, should we listen to the crowds of people demanding they go away, or should we stick to our guns and try emulating them. The people who want to see progress really appreciate it, even if the majority start moaning every time we spend time on them as if we owe it to them to spend time only emulating the things they care about. Popular opinion is worthless here too.

I might have had my own disagreements with the team in the past, and often still do, but things like this are what keeps the project going, what differentiates MAME from others.

I don't think your genesis 'Reset' thing really makes much sense.

There's a 'reset' button on the Console, it's a button, we treat it like any other input. That might become more obvious if the input configuration menus end up divided into sections for each device (which they really should be)

Also disks was a simplified example, there are going to be times later when we're going to be wanting to insert / remove many different devices, add additional machines to a running emulation session, hook up specific serial ports on those machines to other machines; to do that you need an intuitive gui, a generic launcher will never handle it.