Regarding Controls.dat big picture'

[Jakobud]

Hey guys, I appreciate what you are doing with controls.dat, but I have a general question I was just thinking about.  Why are you guys making a dat file for Mame instead of just working with the Mame dev's on getting the controls listed correctly within the Mame source itself?  Is there some advantage to having a seperate dat file?

[Minwah]

Unfortunately, the MAME devs are not as interested in accurately documenting/emulating controls as we are.

I have made a few posts lately on the subject, but as u_rebel and HC will say, many input related fixes have been submitted and rejected time and time again by the MAME devs.  So the chances of them doing something as accurate as controls.dat are non-existent, imo

[SirPoonga]

Why are you guys making a dat file for Mame instead of just working with the Mame dev's on getting the controls listed correctly within the Mame source itself?

That is just not possible.  While through out this project we have found fixable problems, and have fixed them, there are some stuff that just isn't fixable.  One thing we (HowardC, Urebelm and I) have fixed is getting all the mame game to use new inputs that were in mame.  Like using both Pedal constants, using the Z axis for a pedal instead of robbing an axis from another player, etc...  Also like minwah said, the devs can be very stubborn.

Technically, for the most part, mame does document the right controls but it documents the wrong controls with it.  Example.  Discs of Tron, if oyu look at it in listxml it says it is a trackball.  While that is wrong int he output mame does not just use the tball for input.  The tball is a hack to simulate the up down spinner.  However int he driver dial, joy8way, all the buttons are in there correctly.  So it's actually a matter of listxml not showing all the controls defined for a game.  This has been mentioned many times in the last several years.  Hopefully now that mame is going to XML they can output that correctly.  The old listinfo format was limiting.

Then there are drivers that have one macro setup to define the controls.  Usually this is done because the only difference between games is a rom swap, the rest of the hardware is exactly the same.  So they just assign the control macro to all the games.  So if the macro defines 4 buttons but the games uses 3 buttons mame will output 4 buttons in listxml.  This is very widespread in mame.  This is easily fixable but takes ALOT of time.

Then there is the controls themselves.  same control, different name.  Example is spinner and 360 steering wheel.  They are the same thing, just called something different.  In mame they are listed as dial.  this is another task of the controls.dat project, to list a better description for the control used

And then, though I think this was fixed with .85, 2way joys showed up as 8way joys.  This is because int he code they didn't have any room for the bit pattern to define this without making major changes.

Then ther are lapses in thinking.  In .85 720 now uses an anlog joystick.  Well, dial was more accurate.  They should have added analog stick as a hack like they did tball for dotron.

So, to sum up.  The major reasons the controls.dat project exists is
1) To accurately document the controls a game used, using more descriptive terms.  Like 270 Wheel, 380 Wheel, Throttle, etc..
2) To accurately document the CORRECT controls, I.E. a game uses 3 buttons instead of 4, or uses and up/down spinner instead of a trackball.
3) To make dealing with inconsistancies easier.  720 is this example.  In the controls.dat this will always output 720 controller (or whatever we have defined for it) even if in mame it changes from dial to stick.
4) To relate the above controls with how they are defined/hacked into mame.  This allows FE devs to accurately display control information.  An example of this is the johnny5 viewer.  With this part HowardC can read your ctrlr files and accurately display what controls you have do what in mame for a particular game.

I know some of that will go over your head unless you understand the mame source code that deals with this.

[Jakobud]

No, I get it Sirp.  Thanks that summed it up very well.  I didn't know the mame dev's were that stubborn. Thanks!

[Howard_Casto]

I wouldn't say they are stubborn exactly, it's just their goals are very diferent than ours.  

Now this doesn't apply to all of them, so if you are the exception out there please don't gripe to me about it.  But in general, some mame devs don't care about the hardware aspect of a game, merely the game itself.  I year or two back I complained that one of the games (can't rmember which, it's been too long) didn't handle the input ports correctly so it couldn't be played on the original controls.  One of the devs said something very disappoointing.  They said something to the degree of:  "So?  The whole point of mame is so we can play them on the pc, so why would we want to go back a step and put them back on a cabinet?"

That is when I realized that some things would never get fixed.

My feeling is the game and the cabinet and the controls are all one thing and you can't seperate one from the other.  Apparently though, not everyone feels that way.  


Again, not all mame devs feel that way, but there are enough that do to hinder our efforts for accurate inputs.  

Also, I think after all is said and done and the dat is finished (by finished I mean around 90% completion as it'll never be finished)  Efforts could be made to transfer the labels to mame internally so that these would be displayed when you redefine controls in mame's menus.  The ability is already there, it would just be a global change that would effect every driver.  

[Tilzs]

very disappoointing.  They said something to the degree of:  "So?  The whole point of mame is so we can play them on the pc, so why would we want to go back a step and put them back on a cabinet?"

That is amazingly shortsighted for a mame dev. I'd beg to differ that the biggest supporters of MAME are the ones who would like to put things in original cabs with original controls. People who play things on a pc I feel are the ones who do so for a few weeks then move on to the newest FPS or whatever the flavor of the month game is. They may come back to MAME a year later for a few days until the next flavor of the month game is relased.

It is also amazing that if that is the attitude of some of the devs that there is as much documentation of controls as there is. Whats the use of even listing things as a dial if you can just list it as a trackball. Whats the difference in a 4 way joy as a 8 way joy if you are just useing a keyboard or 8way gamepad anyhow.

Controls are as big of a part of any emulation as any other part of the game. If you are going to ignore that fact why even bother emulateing games such as mortal kombat in Mame. The PC version is so close to the original anyhow. If you are going to ignore controls why bother with the improvements in sound as in some  games. Some of the improvements they do make I can't tell the difference anyhow.

[Tiger-Heli]

I year or two back I complained that one of the games (can't rmember which, it's been too long) didn't handle the input ports correctly so it couldn't be played on the original controls.  One of the devs said something very disappoointing.  They said something to the degree of:  "So?  The whole point of mame is so we can play them on the pc, so why would we want to go back a step and put them back on a cabinet?"

That is when I realized that some things would never get fixed.

In some cases I see the Dev's point, in others I don't.

Don't know if this game was one of the one's you complained about or not but here's two examples -

All SNK rotary games (Ikari, Bermuda Triangle, etc.) - The original joysticks used a 12-position switch on the bottom of the joystick to indicate direction.  Most people don't have 24 encoder inputs to give up just for joystick rotation, so you can't use these controls accurately, unless you use MAME Analog +, or use Druin's interface board.

Tin Star, Wild Western, Front Line - Here you have the opposite situation.  The game can be played perfectly with the original control, but that's not readily available.  It could also be played perfectly well with a DOT spinner, except that MAME does not support a dial input for it.

So the devs aren't even consistent.

[Howard_Casto]

In some cases I see the Dev's point, in others I don't.

Don't know if this game was one of the one's you complained about or not but here's two examples -

All SNK rotary games (Ikari, Bermuda Triangle, etc.) - The original joysticks used a 12-position switch on the bottom of the joystick to indicate direction.  Most people don't have 24 encoder inputs to give up just for joystick rotation, so you can't use these controls accurately, unless you use MAME Analog +, or use Druin's interface board.

Tin Star, Wild Western, Front Line - Here you have the opposite situation.  The game can be played perfectly with the original control, but that's not readily available.  It could also be played perfectly well with a DOT spinner, except that MAME does not support a dial input for it.

So the devs aren't even consistent.

SNK rotary games are the way they are because (and please correct me if i'm wrong)  those are the physical inputs the board actually recieved.  There was an interface board between the stick and the pcb. They would have to be, becase those games were supposed to be jamma, and you can't be jamma with 40 some odd inputs.  Regardless, if you buy a interface board for the pc you can use rotary sticks and even if you don't, you can use an optical rotary which is more accurate anyway.  


The reason wild western and other games like them aren't hacked is because the authors of the drivers choose not to.  And yet for years, games like defender, were hacked so that if cheat is turned on you can play with a joystick, even though everybody can manage to make a layout of just buttons.  The fact that the heads of the mame team allow such lee-way is what is hurting the project on the input end of things.  Most devs are good about keeping with the norm, but some aren't.  

I think poor old "frogs" is the best example.  That game's input has been changed about 15 times because every time someone works on the driver, they decide the input should be setup another way.  

And then their are inaccuraces in name only, which NEED to be in mame.  Games with directional buttons instead of a joystick, for example.   These games have each input mapped to a joystick direction.  This is done so that you won't have to remap every time you run across one of these games. And it's a digital input vs a digital input,  so interface wise, it doesn't hurt anything.  However, from a documentation standpoint, it does. See mame doesn't have any facility (other than overwriting with custom labels) to distinguish types of inputs other than the physical and certain joystick macros they've put in.  This is mostly done to avoid confusion.  

Mame doesn't have an up/down spinner type.  Instead, it "builds" one out of a dial and two buttons. This is why no matter how accurate we try to make mame, barring changing the core it is impossible to use it's inputs to document the original inputs. That's what sirp was getting at about it being impossible to make some corrections.  

[Minwah]

My feeling is the game and the cabinet and the controls are all one thing and you can't seperate one from the other.  Apparently though, not everyone feels that way.  

Aaron Giles summed his (opposite) view up in a recent post at MameTesters:

'Being able to hook a real 720 controller directly into MAME is not one of its goals, and is exactly what MAME Analog+ attempts to accomplish, so I think it is appropriate for that support to live on there.'

[Tiger-Heli]

My feeling is the game and the cabinet and the controls are all one thing and you can't seperate one from the other.  Apparently though, not everyone feels that way.  

Aaron Giles summed his (opposite) view up in a recent post at MameTesters:

'Being able to hook a real 720 controller directly into MAME is not one of its goals, and is exactly what MAME Analog+ attempts to accomplish, so I think it is appropriate for that support to live on there.'

Ouch!!! Kinda contradicts what they post at www.mame.net about:

'Even though MAME allows people to enjoy the long-lost arcade games and even some newer ones, the main purpose of the project is to document the hardware (and software) of the arcade games. There are already many dead arcade boards, whose function has been brought to life in MAME. Being able to play the games is just a nice side-effect. '

[u_rebelscum]

SNK rotary games are the way they are because (and please correct me if i'm wrong)  those are the physical inputs the board actually recieved.  There was an interface board between the stick and the pcb. They would have to be, becase those games were supposed to be jamma, and you can't be jamma with 40 some odd inputs.  Regardless, if you buy a interface board for the pc you can use rotary sticks and even if you don't, you can use an optical rotary which is more accurate anyway.  

 
These rotary games are Jamma+; JAMMA+ means jamma standard inputs plus extra non-Jamma stuff.  So the joystick directions & buttons are Jamma, the rotary is the "+".  I.E.: the rotaries are outside the Jamma standard.

To complicate things, different games had different rotary connections.  Most had straight connections for each of the 12 directions (and ground) directly to the PCB per rotary.  Others had a "counter" interface board that converted the pin number that was closed to binary (thus connecting to the PCB through 4 pins).  Some binary interfaces output 1-12, others 0-11 (in binary of course).

FWIW, no (zero) 12 way rotary game can process the output of druin's board.  Please don't model how the real arcades worked off of how druin's board works.

Nikpick: optical rotaries aren't more accurate; they're just more precise.


Besides that, HC is right.  It's not MameDev that's inconsistant.  It's that each driver writer gets to pick how to emulate/simulate the inputs.  

Maybe you can say MameDev doesn't have a written standard for the inputs.  But it's the hundreads of driver writers that are inconsistent.

[SirPoonga]

My feeling is the game and the cabinet and the controls are all one thing and you can't seperate one from the other.  Apparently though, not everyone feels that way.  

Aaron Giles summed his (opposite) view up in a recent post at MameTesters:

'Being able to hook a real 720 controller directly into MAME is not one of its goals, and is exactly what MAME Analog+ attempts to accomplish, so I think it is appropriate for that support to live on there.'

I think dial AND analog stick can both be defined.  tball is defines for dotron.....

[u_rebelscum]

My feeling is the game and the cabinet and the controls are all one thing and you can't seperate one from the other.  Apparently though, not everyone feels that way.  

Aaron Giles summed his (opposite) view up in a recent post at MameTesters:

'Being able to hook a real 720 controller directly into MAME is not one of its goals, and is exactly what MAME Analog+ attempts to accomplish, so I think it is appropriate for that support to live on there.'

I think dial AND analog stick can both be defined.  tball is defines for dotron.....

Sure they can.  Which is what analog+ will do.  Getting into official mame is a different story.  

We're talking about a driver originally by Aaron Giles, with changes made while Aaron is the master running the show.  We tried, and Aaron gave his option.  [shrug]

[Howard_Casto]

 To complicate things, different games had different rotary connections.  Most had straight connections for each of the 12 directions (and ground) directly to the PCB per rotary.  Others had a "counter" interface board that converted the pin number that was closed to binary (thus connecting to the PCB through 4 pins).  Some binary interfaces output 1-12, others 0-11 (in binary of course).

FWIW, no (zero) 12 way rotary game can process the output of druin's board.  Please don't model how the real arcades worked off of how druin's board works.

My bad, I have only seen the "counter" type and asumed they all worked that way.  No I'm not that dumb druin's board definately isn't the way it was done in the arcades as the arcades didn't have trouble skipping around.  

Regarding aaron's comments:

That doesn't suprise me.  He and I don't always get along.  The reason we don't get along is because of stuff like that.  His views don't bother me, rather the way his views change based on what mood he is in that day or what view he needs to put in some hack and justify it.  The 720 degree hack is a prime example of this.  

There was no rhyme nor reason to putting it in as the game is just as unplayable on standard hardware as it was before, if not more-so.  I just tried it on an analog stick and the new inputs make turning twitchy and akward. And as inaccurate as it was in the past, at least then I could actually steer fairly well with commonly available controls. Probably he's rigged something up and just put it in for his own benefit.  I hate to talk smack like that, but again, it wouldn't suprise me.  

[SirPoonga]

My feeling is the game and the cabinet and the controls are all one thing and you can't seperate one from the other.  Apparently though, not everyone feels that way.  

Aaron Giles summed his (opposite) view up in a recent post at MameTesters:

'Being able to hook a real 720 controller directly into MAME is not one of its goals, and is exactly what MAME Analog+ attempts to accomplish, so I think it is appropriate for that support to live on there.'

I think dial AND analog stick can both be defined.  tball is defines for dotron.....

Sure they can.  Which is what analog+ will do.  Getting into official mame is a different story.  

We're talking about a driver originally by Aaron Giles, with changes made while Aaron is the master running the show.  We tried, and Aaron gave his option.  [shrug]

Guess controls.dat needs updating  

[Howard_Casto]

Actually I would wait.... My guess is as soon as someone else takes the reigns it'll change back.  Like I said, the change didn't make any sense.

[Minwah]

Ouch!!! Kinda contradicts what they post at www.mame.net about:

'Even though MAME allows people to enjoy the long-lost arcade games and even some newer ones, the main purpose of the project is to document the hardware (and software) of the arcade games. There are already many dead arcade boards, whose function has been brought to life in MAME. Being able to play the games is just a nice side-effect. '

Exactly.  I have tryed to bring this point up within earshot of said dev(s) a couple of times lately (at mame.net and mametesters), and surprise surprise did not get a comment about that.

They should at least remove or revise the above piece of text at mame.net imo, as they clearly don't mean it.

[Howard_Casto]

Our real problem:

Nicola Salmoria isn't taking an active role anymore.  

He's the man and if he was still running the show the way he used to we wouldn't even be having this discusion.  

If we could somehow coax him out of "retirement" I think these problems would magically get solved without complaint #1.  

[Minwah]

Our real problem:

Nicola Salmoria isn't taking an active role anymore.  

He's the man and if he was still running the show the way he used to we wouldn't even be having this discusion.  

If we could somehow coax him out of "retirement" I think these problems would magically get solved without complaint #1.  

I bet contacting him is easier said than done, never mind coaxing!