The Multi-Pointer X server

worufu writes "Some weeks after releasing the MPX (Multi-Pointer X Server), the Linux world slowly seems to draw attention to the project which opens up the limits of simultaneous input devices of the current X server. The future possibities are unlimited and I cannot wait to see some nice applications supporting the advantages of multiple input devices.
From the project description: 'The Multi-Pointer X Server is an enhanced X server to support multiple mice. It provides users with one cursor per device. Each cursor can operate independently. A multicursor windowing system allows two-handed interaction with legacy applications, but also the creation of innovative applications and user interfaces.'"

Comments

Multi-Mouse

[Nation XII]

FINALLY! I can play multi-player xPong at work and never again have to argue who gets the mouse!

Settlers 2

[moderators_are_w*nke]

Didn't Settlers 2 multiplayer (Amiga I think) do this years back. Maybe we can get it to work.

Re:Settlers 2

[Anonymous Coward]

You're thinking of Amiga Settlers, Settlers 2 never came out on the Amiga sadly. But yes, it did have support for two mice, as did many other multiplayer games at the time on that platform.

Re:Settlers 2

[xaez]

I remember playing Settlers 2 on pc with 2 mice on the one screen years ago.

Re:Settlers 2

[gnud]

Yes, Settlers read the raw mouse data from the COM port. I never used it, but remember the dialog where you could specify a port.

Re:Settlers 2

[nutshell42]

Didn't Settlers 2 multiplayer (Amiga I think) do this years back. Maybe we can get it to work.

Didn't they just release a Settlers 2 remake? Coincidence?

Window stretching

[glowworm]

I was a little sceptical, but after I RTFA and watched the demo I would definately love to see some of those techniques make their way into window managers. Applying the snippet of the demo with the photo light box to the desktop would be quite nice, resizing windows by pulling the corners apart, flicking them into corners and so on.

Mind you, how do you keep the screen clean of fingerprints and pizza grease smears.

Re:Window stretching

[PrescriptionWarning]

screens will have to be replaced with holographic fields I'd imaged, a'la Minority report

Re:Window stretching

[Kancept]

I can already resize windows by pulling them apart. Man I wish Fingerworks woulda gotten somewhere instead of "bought out" Then maybe more would understand this is old tech that was hidden from the populace!

Re:Window stretching

[Ungrounded Lightning]

Mind you, how do you keep the screen clean of fingerprints and pizza grease smears.

Keep the cat out of the pizza.

Only one hand available

[Anonymous Coward]

All I use my home computer for is porn, so I couldn't use two mice anyway.

Re:Only one hand available

[Anonymous Coward]

Depends what you are pushing the mouse with.

Re:Only one hand available

[Gleng]

Maybe you could operate both the mice with your feet and free up both hands.

Reminds me of Anakin

[cerberusss]

Using two mice reminds me of Anakin Skywalker, with his two lightsabers. And I don't need to remind ANYONE here what happened to HIM.

Re:Reminds me of Anakin

[BlueBlade]

You mean he went on to be the second most powerful person in the universe for 30 years? Call me evil, but... where do I sign up?

Re:Reminds me of Anakin

[space_dude_27]

Yeah, George Lucas turned him from evil super-villain to whiny teenage brat in the space of just three films.

Re:Reminds me of Anakin

[hey!]

You mean he went on to be the second most powerful person in the universe for 30 years? Call me evil, but... where do I sign up?

Well, that's the thing about the Dark Side, isn't it? The #2 spot is actually worse than the #3, because #1 has to constantly crush you under his thumb, to keep you in your place and show you who's boss.

No, like Robinson Crusoe's father, I believe the middle station is best -- or rather, the middle station of the upper echelon. I'd go for, say #100 on the hit parade. High enough up to be to far too valuable to waste, not high enough to be considered a threat to the established order.

Re:Reminds me of Anakin

[hey!]

Yeah, George Lucas turned him from evil super-villain to whiny teenage brat in the space of just three films.

Which goes to show you that truth isn't so much stranger than fiction than it is less credible. Evil men aren't the titans they like to portray themselves as. They are typically petty men, puny vindictive characters whose insecurity worms away at their ego, driving them to seek external validation by extravagant displays of power and cruelty.

When social scientists and political thinkers looked at what had gone wrong with Germany after the Second World War, they were shocked, not by the party laders' excesses of cruelty, which they were prepared for, but by their dreary, commonplace character. Hannah Arendt coined a phrase for this: "The Banality of Evil".

Think Kim Jong Il. You wouldn't hire him to run a hot dog stand, much less a country. He is very obviously a clown. But he's a ruthless clown with a talent for spreading fear.

The problem with the episode one to three movies was that Lucas, freed from any artistic or financial limitations, produced a series of ponderous and excessive movies, like an over the hill prize fighter carrying fifty pounds of excess fat. But Anakin's annoying character was right on the money. Folks, he's supposed to be contemptible. In the original trilogy Darth Vader had a kind of evil glamour, like Satan in Paradise Lost:

So farewell, hope; and with hope farewell, fear;
Farewell, remorse! all good to me is lost;
Evil, be thou my good; by thee at least
Divided empire with Heaven's King I hold,
By thee, and more than half perhaps will reign;

Lucas wants to undermine this glamour, and show that Vader is really a pathetic character. Second baddest bad guy in the universe is not a happy job. Lucas obviously wants to make a philosophical point, and it's a good point. The problem is that he's not as entertaining doing it as when he's just hoping to be able to get the movie made.

Re:Reminds me of Anakin

[Svenne]

Yeah, he got to boink Natalie Portman. Thanks for the reminder, I think I'll be getting my second mouse just about now.

Re:Reminds me of Anakin

[cerberusss]

LOL better order a few extra...

Re:Reminds me of Anakin

[EnsilZah]

Are you suggesting this may lead people to the... windows side of the OS?

Re:Reminds me of Anakin

[Anonymous Coward]

How could it have been 30 years? Luke was only a teenager in the 4th film.

Re:Reminds me of Anakin

[ultranova]

Lucas wants to undermine this glamour, and show that Vader is really a pathetic character. Second baddest bad guy in the universe is not a happy job.

Actually, Vader is glamorous, while Anakin is a pathetic human being. Remember, int RotJ, when Luke removes Vader's helmet ? He's no longer an inhuman angel of death, but an old bald guy who can't even breath by himself.

It's the same as Nazi Germany: it was glamorous, but once the glamour was gone, those who had given themselves to it were no different than anyone else. After all, one meaning of "glamour" is "magic enchantment", something that's not quite real.

Vader was just a mask. Anakin was the real man inside. Masks can be seemingly invulnerable and glamorous in their inhumanity, while human beings can't. But human beings can wear masks, either metaphorically or literally. Doing so can make you a titan, but it also makes you less than human. And sooner or later the glamour is gone, and then what ? But the temptation of denying all weakness and appearing glorious, and maybe even believing yourself so, is so very hard to resist...

And of course we don't resist it, any of us, completely; we always keep some parts of ourselves hidden. To do otherwise is to make your weaknesses exposed, and to show weakness is to invite abuse. And so we keep on getting more and more inhuman monsters spreading death and destruction around the world.

So, sayeth the Horned Paperclip, what mask do you want to wear today >?-)

Re:Reminds me of Anakin

[SanityInAnarchy]

I'd go for, say #100 on the hit parade. High enough up to be to far too valuable to waste, not high enough to be considered a threat to the established order.

Won't work. Vader has a tendency to waste people who fail him, no matter where they are.

Hell, Vader has a tendency to kill anyone, with him or against him, including the person in the #1 spot. The safest ones would be old, familiar Jedi like Luke or Obi-Wan, because Vader will have the decency (stupidity?) to face them head-on, alone, in a lightsaber duel. Anyone else, he doesn't care, he'll just blow up their damned planet.

Anyway, back on topic, two mice does sound cool, but mostly it would be useless and confusing. For games, I'd really prefer a mouse and a joystick anyway -- joystick replaces WASD with actual, analog control, mouse is much quicker and accurate than having another joystick (like you would on a console game). Joystick feels much more natural than any mouse configuration I can think of for controlling walking, though -- if it's based on how far the mouse is from a central point, it'd be difficult to stop, but the alternative is to pedal the mouse to walk...

Re:Reminds me of Anakin

[SanityInAnarchy]

Yeah, then he killed her. You bastards!

Four hands..

[Anonymous Coward]

Two hands and two feet.. might be more useful than one thinks!

another way

[qwp]

another way i can be at the computer, and yet someone can come up and steal it away from --

Ah, memories

[Yeti.SSM]

This reminds me of "mice fight" we used to do sometimes with my friend.
We plugged PS/2 and a serial mouse into one Windows 95 or 98 box and then moved both the mice simultaneously. Great fun, you bet! Duh...

Re:Ah, memories

[gravij]

hehe, I remember plugging a second mouse into my older brother's computer, hiding the cable and then putting the mouse on my desk. He wouldn't notice anything was wrong until I started laughing. It was fun until he caught up with me. Another trick I remember was recalibrating the mouse so that moving it wouldn't move straight up and down. Funny though I can't find the setting for it here in win xp.

I wonder...

[durin]

... when Microsoft will "innovate" this.

Re:I wonder...

[mr_jrt]

Windows had this at a low level ages ago, and DirectX supports multiple mice too. In the switch to Win2K+ however, Microsoft deemed it a security issue and made it damn near impossible to query multiple mice easily. Can still be done now, but you have to go much lower level to do so (capturing raw mouse events and decoding them yourself). Annoyed the hell out of me when my multiple mice were only ever being detected by DX9 as a single device trying to play multiplayer air hockey.

Re:I wonder...

[Anonymous Coward]

Oh, you mean this?

http://research.microsoft.com/displayArticle.aspx? id=1471

I would be a lot more impressed

[Clockwurk]

if they made X natively support more than 3 mouse buttons. Or if X could automatically detect usable refresh rates/screen resolutions for my monitor.

Re:I would be a lot more impressed

[prefect42]

Detecting usable refresh rates and resolutions is a driver thing. The nvidia binary driver supports it just fine.

Re:I would be a lot more impressed

[Anonymous Coward]

ati the same, works best in conjunction with dvi connected displays anyway

Re:I would be a lot more impressed

[Anonymous Coward]

if they made X natively support more than 3 mouse buttons
As far as I remeber X was one of the first environments to use 3 buttons. So what exactly are you missing?

Re:I would be a lot more impressed

[ratboy666]

X "natively" does support more than three buttons. AFAIR (and, its been a while since I've worked on the X server input section), the limit is 128 buttons.

Ratboy.

Re:I would be a lot more impressed

[isj]

The X protocol directly supports 5 buttons. Additional buttons can be supported by X Input Extension (XIE). The scroll wheel is usually handled via XIE.

Re:I would be a lot more impressed

[Anonymous Coward]

As far as I have seen, so far, just about all scroll-wheel setups i've seen in XF86Config/xorg.conf's have been by mapping 'up' and 'down' to buttonpress 4 and 5 .

The original X11 docs (rev 5 at least) list that the max number of mouse-buttons supported (according to the mouse_button_press_event docs) was 5. I don't think that there's any reason why one couldnt expand that number by altering the definition of all related structures in the X-protocol.

However, doing that would most likely break a lot of legacy stuff that'd behave unpredictably on any 'button 6' presses.
To bypass the whole problem, the X Input Extension (XIE) was devised; allowing an application to 'ask' what a particular device supported and then choose how to handle them. In effect, it allows the device-driver within X to provide any number of possible 'controls' that can be read out and what names they should have. That could be anything from X/Y/Z axis, buttons, rotational sliders, etc. It doesnt really matter. I'm sure that, if one would wish, one could make a driver that would map all 101/103 keys on a typical keyboard to XIE events and read those out.

But, currently, scroll-wheel events are done using the line
Option "ZAxisMapping" "4 5"
in the x-config. That'll map the "z-axis" (meaning, the scrollwheel) to button 4 and 5; something any application can read out just using mouse-button events. Be aware that XIE isnt supported by any applications apart from those really needing it (Gimp, etc, since they'd want support for a Wacom tablet, etc). I'm sure there'd be a way to 'catch' XIE events and make sythetic 'normal pointer events' for any non-supporting apps.... but YMMV.

Re:I would be a lot more impressed

[simcop2387]

actually i've seen alot of applications (this might be due to gtk handling input) support my 6th and 7th buttons, works rather well, in firefox they are back and forward, and they do work in the gimp as part of the scroll bars allowing me to scroll vertically AND horizontally, that tends to work in just about every gtk app i've tried

Re:I would be a lot more impressed

[JohnnyBigodes]

Ditto to that. It's been SEVEN years since I first started dabbling in Linux and it *still* won't properly detect one of most basic things ever, the monitor's refresh rates. It's VESA DDC and their friends, folks, they're open standards, what's the big problem?

And never mind that at least for me, this puts a severe hamper in usability for those who can't bother to use fine-tuning tools and quickly kill the patience of those that do. Having used high resolutions for the past years and not having them properly detected even once causes to the picture to never be straight or correctly stretched in the monitor. And that's after over half an hour's fiddling for just *one* resolution.

Re:I would be a lot more impressed

[IBitOBear]

I don't knw what you've tried, but I find that for my hardware and software auto-selection is a simple matter of _not_ _setting_ any refresh rate at all in the monitor section and then enabling DPMS. Your milage will vary of course depending on the hardware and driver in use.

Notice what I commented out:

Section "Monitor"
                #DisplaySize 360 270 # mm
                Identifier "Monitor1"
                VendorName "VSC"
                ModelName "A91f+"
                #HorizSync 30.0 - 86.0
                #VertRefresh 808531264.0 - 808792640.0
                Option "DPMS"
EndSection

The problem is _actually_ that a lot of the configuration tools are not as well maintained as the server itself, so the data and patterns in the configuration programs are "a bit off".

As a matter of course I generally configure by hand. First I do "X -configure" to get a good starter file. Then I comment out the sync rates as above and add "DefaultDepth 24" to the screen section.

Also, if you use a KVM you have to be switched to the X machine when X starts or the DPMS can not do the configs.

Next I make sure that I boot linux with psmouse.proto=any (to maximize the /dev/input/mice functionality).

Finally I try to switch all the mice to evdev if there isn't a spesific driver with more functionality.

Yea, it isn't effort free, but it is easy and optimal.

Re:I would be a lot more impressed

[JohnnyBigodes]

Well I certainly appreciate all the pointers you've given, but it only goes as far to prove the point that some is *very* wrong with X for not doing something as simple as collecting the correct monitor data from DPMS :)

Thanks again for the tips though :)

Re:I would be a lot more impressed

[Schraegstrichpunkt]

Read the post. He wrote more than 3 buttons.

Re:I would be a lot more impressed

[Schraegstrichpunkt]

Do you know anything about how a KVM switch works? X can't detect your monitor's parameters if the monitor isn't connected to the video card when X starts.

Re:I would be a lot more impressed

[Schraegstrichpunkt]

Never mind. Apparently X supports that too.

Re:I would be a lot more impressed

[netfunk]

The scroll wheel is usually handled via XIE

No it's not. It generally maps to buttons 4 and 5, specified in your X server config file, and causes problems for hardware with more than 5 buttons. XInput is rarely used by most apps (and most X servers), and there're lots of programs out there now that would fail if you mapped the mousewheel to different buttons, since it's a de facto standard to have them on buttons 4 and 5.

Drives me nuts when people say fundamentally incorrect things with authority, since they always get modded "Insightful."

--ryan.

Re:I would be a lot more impressed

[JohnnyBigodes]

Wasn't talking in the KVM switch context.

Re:I would be a lot more impressed

[Schraegstrichpunkt]

Well, then you appear to be wrong on two counts: 1. X does support reading monitor specifications from the monitor itself (assuming the video driver supports it -- Mine does). 2. DPMS has nothing to do with it. You're probably referring to EDID or DDC.

Re:I would be a lot more impressed

[JohnnyBigodes]

Yes VESA DDC. And I know that X supports reading the monitors specs... it just never happens to get them right :) Might not be the case with the latest and greatest nVidia binary drivers, but every single time that I tried some distros (Ubuntu 5.something was the latest), the refresh rates were "almost right" but never there.

Re:I would be a lot more impressed

[Schraegstrichpunkt]

Hmm. Does it work on other OSes? I suspect that some monitors have screwed up EDID information. If not, you could file a bug report (though if you're using the proprietary nVidia binary drivers, you might not have much luck with that).

Race conditions

[Carewolf]

What an excellent way to make every toolkit vulnerable to thousands of race conditions ;)

You can probably crash 99% of all X11 applications using two pointers

Re:Race conditions

[orangesquid]

Hah!
What I'd like to know:
Does this work as a nested X server? In other words, can I run it in a window, using the native X's mouse as the primary mouse, and then have MPX have exclusive access to additional mice?

In the FAQ it talks about MPX having to 'guess' which pointer to use for QueryPointer. Looking at the xeyes screenshot, I wonder if it would be possible to extend it so that certain applications could be linked to a certain mouse in various ways (when multiple mice are possibly choices for function calls, define different rules for picking the cursor for different applications). Also, the FAQ talks about LeaveNotify events and xterm.. maybe add a special xterm handler that only generates LeaveNotify when no other cursors are still focused on the window?

which raises the question: exactly how is focus handled? Can I type in two xterms at once? Are there lots of different complex focus models to choose from now?

Re:Race conditions

[HTH NE1]

which raises the question: exactly how is focus handled? Can I type in two xterms at once? Are there lots of different complex focus models to choose from now?

Crisis or opportunity?

I've been wanting this for years. Unfortunately, not only have I had no time to work on it, but I still don't now that someone has done the basics, But I've thought of it as ways to have one user use two mice, not two users in one session each having his own mouse at the same display, and how to extend the UI to support it.

Having two mice allows you to eliminate the implied paperweight that says you can only resize windows by grabbing their edges or other special widgets. Instead, one mouse moves, two mice in different directions stretch. With dual focus-follows-mouse now you'll either want dual keyboard support or stop using focus-follows-mouse, or make it only follow when both mice are in the same object. Otherwise, conflicts for overlapping focus space can be handled by shifting windows out of conflict or require click-foward.

Re:Race conditions

[ultranova]

What an excellent way to make every toolkit vulnerable to thousands of race conditions ;)

Um, what race conditions ? Here's the basic pseudocode to an X application:

1. Initialize.
2. Get next event from X-server.
3. React to the event if neccessary.
4. Go to 2.

I don't really see how having 2, 20 or 200 pointing devices could cause any race conditions here. You just get 2, 20 or 200 times the amount of mouse events (assuming they are all actually used and won't just lie there) than before.

You can probably crash 99% of all X11 applications using two pointers

Can anyone out there test this ? Did your Gnome / KDE / Whatever desktop crash with two pointers ?

Re:Race conditions

[Carewolf]

Crash would be the ultimate reaction. To begin with most applications would just act weird, since every other event would put the pointer a completely different place. Stuff like tooltip, mouseover effects, window focus and even web-based JS would cease to work. Most likely many of the codepath would when experiencing these odd events enter broken code-paths and crash.

Re:Race conditions

[Anonymous Coward]

It will be like the Year 2000 all over again!

Re:Race conditions

[CastrTroy]

you could just designate 1 device as the primary device, and only events from that device are sent to the legacy applications, or at your discretion, applications from a list that don't behave properly. Then there wouldn't be any problems at all.

Good...

[tetabiate]

whenever I get tired of using the right hand (er, no, I'm not making allusion to that activity thought to be common between ./'ers, you sick bastard!) I can switch to a left-handed mouse immediately.

Re:Good...

[multipartmixed]

Did you really mean BETWEEN slashdotters, or *amongst* slashdotters?

Definately a different connotation.

Re:Good...

[tetabiate]

There you have me.

Re:Good...

[Anonymous Coward]

Just use a neutral mouse, and then move it around; I've been using my MS mouse with my left hand for quite a while now.

Re:Good...

[Anonymous Coward]

OK, so it's "between". At least for you and him :-)

About time!

[MMC Monster]

I always wondered when multiple mice would be supported.

It's great for when someone remotely logs on to help a user with a problem.

Gameplay would be very interesting - Use one mouse to point and shoot while using another to move around.

You can tutor someone else on the same computer. Maybe have the mice look different so there's no confusion.

Re:About time!

[sowth]

It's great for when someone remotely logs on to help a user with a problem.

Actually, that shouldn't make any difference. When you log into a remote computer and run an X client, it connects to your computer, and you use your local mouse and keboard because those are the ones your X server is connected.

Maybe if you were sitting right there, it may help, but remotely connected, it wouldn't make a difference...unless you are talking about something I've never heard of...

Re:About time!

[jafuser]

I bet it would make maniptulating things in 3D a lot easier, since now you have two 2D input axises.

Re:About time!

[jyx]

You want to fire up an Amiga then. Im pretty sure you could plug 2 mice/mouses/micee/dodads into and 'boink', a second cursor would appear on screen. Dueling pointers anyone?

It was great for playing a 2 player dungeon master style game called bloodwych (or something). Happy days indeed.

Re:About time!

[Ungrounded Lightning]

I always wondered when multiple mice would be supported.

Ditto.

I've wanted two hands "behind the glass" for years. (Like since I got one fingertip/finger-thumb-pincher back there with the current paradigm.)

It's not VR yet. But two hands is much cloaer than one.

Re:About time!

[Random832]

When you log into a remote computer and run an X client, it connects to your computer

And that's why X is completely inadequate for remotely helping a user with a problem. He's saying it would be nice to be able to log in in a vnc-like manner but have _your_ mouse pointer in one place, and the user's mouse pointer in another place.

I was just bitching about this

[bky1701]

The other day. Yay, I guess. Now how long does it take to get into disros.... I refuse to deal with X on a low-level, after what happened last time. SuSE update or bust.

Benefit vs. current two-handed input?

[syntaxglitch]

I'd be very doubtful of the value of this for conventional mice and single-user computers. Mice are best used for highly position sensitive input with limited input types... drawing lines, selecting groups of items, and other things where the location of the input is important and the limited range of input values (left or right click) isn't a hinderance. Adding multiple mice allows more input of this type, but at the expense input of other types.

Not many applications are likely to require a lot of mouse-type input, and ones that do (3d modelling, image manipulation...) are likely to either demand enough attention to detail that most people would only be able to work effectively with one pointer or the other at a time, or to require a substantial amount of input with a much richer range of values than a mouse provides, which will result in either sub-optimal mouse use (agonizingly slow hunt-and-click toolbar use) or switching hands from one mouse onto the keyboard or some other input device (which gets awkward).

Not to mention that the vast majority of applications being used regularly (email, IM, word processing...) require mostly keyboard-type input; people already often use the mouse for things it isn't best suited for, such as using mouse-driven toolbars or menus for common operations when keyboard shortcuts would be more appropriate.

On the other hand, this could potentially be very interesting for multi-user or otherwise non-traditional applications, with an interface designed around the concept, which I assume is where they're going with this and just using normal desktops as demonstrations right now.

I wish them luck, but definitely won't be plugging anymore mice into my desktop.

This should be default

[Ougarou]

This should be part of the default X.org server configuration, I can't see why it shouldn't, so I hope the next time I see this project, I already have it installed ;-)
I think this is the future, I would love to play a game where I could do first-person bare-fist fighting by moving two moses up and down over my table! (Sword and shield would also be cool!)
-- We create the future!

2 person PC

[muftak]

If you had 2 monitors, 2 mice and 2 keyboards, you could have 2 people using 1 PC, and they could drag windows across to each other.

Re:2 person PC

[hey!]

Well, I think you're on the right track here.

In user interfaces, we basically assume that the computer interacts with one person at a tim But real world scenarios, two or more people can work on the same thing at the same time, or on different parts of the same thing, or on different things in the same space then put them together. But it's not possible to do this naturally on a single computer.

The key I think will to be to find the right metaphor, without being excessively literal. In the Desktop metaphor, windows are really like sheets of paper that can be shuffled around. I remember seeing this for the first time, and there was a tremendous sense that this would liberate us from a lot of artificial limitations. But sheets of paper aren't resizable, and don't have scroll bars. Many times in the early years, attempts were made to make the metaphor literal, showing a representation of a desk with file drawers, papers on top, and desk accessories like clocks and calculators. These attempts at literalism turned out to be a waste of space and time, and the metaphor was pared down to its barest essential: a two dimensional surface on which things can be arranged.

The shared computer is an important evolutionary step, but there are going to be a lot of missteps created by excessive literalism, until the fundamental transformation takes place in peoples' heads.

Two trends are going to unite in the future: the continued reduction in size and cost of computing hardware, and the continued increase in ubiquity and affordability of wireless networking. I see the end result, some time in the next twenty-five to fifty years, as this: the computer as we know it will no longer be the primary interface of human beings to information processing technology. By this I mean your grand children won't routinely carry laptops and PDAs, any more than you carry a riding crop. Instead, we will have computer enhanced environments in which people work in a natural and shared way. Interfaces will be heterogenous, from large wall sized screens or interactive tabletops, goggles that computer enhance vision (e.g. overlaying an animation on top of the copy machine to show how to unjam it), and countless sensors.

Naturally these environments will be shared, otherwise they aren't really environments.

This Multi-Pointer technology doesn't get us there, but it enables some steps in the right direction.

Re:2 person PC

[zootm]

If you had 2 monitors, 2 mice and 2 keyboards, you could have 2 people using 1 PC, and they could drag windows across to each other.

Also interesting (I think): Could this have an effect (positive or negative) on the efficiency and/or effectiveness of pair programming?

Re:2 person PC

[npsimons]

In user interfaces, we basically assume that the computer interacts with one person at a tim But real world scenarios, two or more people can work on the same thing at the same time, or on different parts of the same thing, or on different things in the same space then put them together. But it's not possible to do this naturally on a single computer.

Really? I've heard of computers that can be used by more than one person at a time; they're called servers. Also, CVS and other change management systems allow two people to be working on different parts of the same thing, or different things and then put them together.

The shared computer is an important evolutionary step, but there are going to be a lot of missteps created by excessive literalism, until the fundamental transformation takes place in peoples' heads.

This is right on the money, and it's insightful comments like these that convince me to put you on my friends list ;) That being said, I think the problem is that we have been held back by the desktop metaphor that was developed perhaps too early. I mean, a server with SSH does shared computing perfectly; a desktop with a GUI, not so much, mainly because when the desktop metaphor was designed, you could only have one mouse and one keyboard.

My ultimate point is, I'm glad to see this happening, and I'm not surprised to see it come from the OSS/UNIX world. We've had multi-user text shells mastered for a while now, it's only natural to move onto GUIs/desktops. Windows and Mac have always focused on the "mememe" single egotistical user, and their interfaces (not to mention security) have suffered because of it. Now, here we have true innovation, yet again from the open source world. To all those naysayers who say that open source merely copies Windows and Mac, I say "how many users can your desktop support?".

Re:2 person PC

[Anonymous Coward]

I think that indeed this is one of the key use cases for MPX. Except you don't need to think in terms of two monitors if you don't want to. You could have just one really large monitor hanging on (or projected onto) the wall, with several sets of wireless mouse/keyboard pairs hooked up.

Think conference rooms... ...with off-site developers VPN'd and VNC'd in to interact with the virtual desktop space using their own keyboards and mice.

Think three kids doing their homework at the same time on one computer...
While another kid is playing "full screen" games in a VM window in the corner.

Which brings up an important question. Why don't the demo videos talk about multiple keyboards? Presumably we need to know how keyboard focus works in a multi-pointer system. Can two people edit a text file simultaneously (from one window)?

From the FAQ: Does MPX have support for multiple keyboards?

Answer: No. Not yet. It's somewhere down the pipe in my TODO list but not at the top yet.

Stink.

My personal favorite use-case involves joining this up with VPN, VNC, and VoIP chat to get a truly usable long-distance pair-programming setup.

I hope Steve Jobs and Redmond are watching and learning.

Re:2 person PC

[Anonymous Coward]

Modal dialogue boxes become sorely abusive in an environment like this.

Re:2 person PC

[tetabiate]

Why not run two copies of the X server on different monitors (:0 and :1) using either the same or different graphics cards and two sets of input devices?

Re:2 person PC

[lazarusdishwasher]

I have a link that will let you use the equipment you mentioned handle two seperate logins.http://blog.chris.tylers.info/index.php?/ar chives/14-Multiseat-X-Under-X11R6.97.0.html

I have never tried this solution but from what I have read in the past it seems to be getting easier to handle multiple input devices and displays. I would agree with others that if you wanted the same login to hanle both sets of keyboards mice and monitors you run the risk of confusing the operators or the programs.

Re:2 person PC

[LordMyren]

multi pointer is a huge first leap, and i think it will springboard us into something greater. but htere are some very profound changes in mpx alone. i think mpx will start to shift people into the idea of a distributed computing environment.

my own case was pretty simple. freshman year of college i built myself a video project. it was pretty damned bright, a not very efficient bulb but 600watts of non-efficiency still adds up to pretty bright. so i had a huge screen (aka, painted sheet). plus, i had two monitors. all of it rigged to one computer. so there'd be ten people in my basement, chilling out playing video games, watching tv / movies, listening to music, browsing the web or playing flash games sometimes all four at once. but rigging up the input for this was catastrophic. ZSNES or ePSXe had two usbplaystation adapters, and used those joysticks, the web browser got the keyboard and mouse, the movie & music had an ATI rf remote, and Dscaler ran the PS2 which had its own input. if anything needed adjusting (and something always did, the image quality was crap without extreme fine tuning and tweaking, the sound levels was always changing & needed mixing) i had to tell whomever was using the computer to pause for a second while I borrowed the mouse & keyboard. I had two logitech cordless sets, but of course they shared the same input "channel". there were some fd.o threads and X threads that were interesting on MPX, but i never seriously believed anything would ever happen. i had some loose code started to grab devices and try to construct virtual cursors, but it got put on back burner a couple years ago (and i never believed it would mature into an actual solution).

discovering mpx a couple weeks ago was one of the most life altering experiences i have ever had.
knowing there's someone else out there with a similar conception of how this whole game is playing out, that is just refreshing.
i really enjoyed your post, thank you.
lordmyren

Re:2 person PC

[dbIII]

If you had 2 monitors, 2 mice and 2 keyboards, you could have 2 people using 1 PC

That bit is easy with a two head nvidia video card (eg. VGA + TVout), ps/2 keyboard + mouse, usb keyboard + mouse and two copies of X running at once. You just need to add another layout with the name of your choice (eg. TVout) to your config file that tells it to use the correct monitor, keyboard and mouse. You then turn the second display on with a script that will contain a single line that looks like:

startx -- :1 -layout TVout

Gnome may freak out at having two instances running as the same user but most other display environments will be OK. I've used fluxbox and E17 on the TV just to make things easier to see, and with E17 I can ALT-TAB the mouse pointer between windows when the mouse batteries die. As memory - the big memory user for X is in the libraries that you already have in memory for the first version of X unless you have a lot of different large background pictures.

and they could drag windows across to each other.

That bit is harder - there was/is a thing called Xsend (or similar) used to send windows from one X display to another.

Re:2 person PC

[Anonymous Coward]

SGI already did this with their (a PDF, it's all that's left apparently) CADduo offering.

Legacy app problems?

[AnyoneEB]

Couldn't this cause problems with legacy apps that assume that there is only one mouse pointer? I don't know about you, but I usually that onMouseOut is called when there is no mouse pointer over an object. I usually do not keep track of the number of mouse pointers in that instance. Then again, mouse overs are tend to be used for minor things like highlighting buttons the mouse is over.

Re:Legacy app problems?

[Chris Pimlott]

From the demos, it looks like they just merge events from both mice into a single queue. If you only use one mouse, it acts as a normal single mouse. Once you move the other cursor, it's as though you instantly moved from the previous cursor's position to the other cursor's position. Since mouse events are only fired on movement or button presses, it doesn't matter than the other cursor is still "over" the element until you do something with it.

Re:Legacy app problems?

[DrMorris]

I didn't read the article, but would that be a pretty useless solution? It would be highly compatible of course, but else it would mean that you just recieve events from one mouse that is rapidly jumping around (in case you move both at once)...

Re:Legacy app problems?

[Chris Pimlott]

Well, this is just for legacy apps. It doesn't get you much, but it's probably preferably to having to designate one mouse the "master" and ignoring the other. There are potentionally some uses; in the Gimp example, one cursor is left over the tool palette and the other over the image. The user alternates between using the left mouse to change tools and the right mouse to apply them to the image, without having to constantly move the mouse back and forth. They also create some interesting effects with some of the paint tools by being able to instantly move the "effective" mouse position back and forth.

But, yeah, for the really interesting and innovate possibilities, you're going to have to program your app to a newer library or API to take advantage of the extra pointers.

Lemmings

[bhaak1]

Lemmings also had two mice support in the two player levels.

As did a lot of games back on the Amiga, even PD games.

I remember an Asteroid PD clone where you moved the character with one joystick and with the other joystick you controlled the fire beam.

Was quite an immersive feeling.

Re:Lemmings

[funkmasterbillis]

I remember an Asteroid PD clone where you moved the character with one joystick and with the other joystick you controlled the fire beam

Don't know if this is what you were referring to but gunroar does this, and it is quite entertaining even if I can't get past the third boss.

Can you do the same

[tmk]

...with a TV remote control?

Imagine: You and your girl friend get each an RC - you watch Jon Steward and she hears the sound of Gilmore Girls.

Re:Can you do the same

[hamfactorial]

For maximum enjoyment, I suggest listening to Stewart and looking at the girls.

Re:Oh yeah?

[eno2001]

Yeah but try actually clicking and dragging individual objects simultaneously with those multiple devices... That's this is all about. Just imagine having a virtual table covered with photos and different people sitting around the table dragging individual photos to them or duplicating and passing the image along to others around the table. THAT is what we're talking about here. Imagine a design meeting being held around a table like this and the presenter loads his documents into the table. He then, in real time, creates duplicates with his hands and pops them to each person around the table so they can read and possibly even edit in a collaborative mode. THAT is what makes this really cool. Of course, as usual, this was available on Macs first. HP is working on stuff like this too.

Robotron!!

[Gr8Apes]

with mice, instead of joysticks....

Why?

[Greyfox]

Hmm... second axis in a 3D UI? I hate the idea of taking another hand off the keyboard though... how about if each mouse had 52 buttons?

It's called the evdev interface...

[IBitOBear]

Um, I have a Logitech G5 mouse and X processesses all the buttions, the wheel works and the thing where you push the wheel left and right to do forward and back (e.g. a fourth axis) works just fine as well. [Note the "Name" option instead of spesifying the "Device" pathname, this lets me plug the mouse in, or not, without having to tweak any settings.]

See the complicated configuration:

Section "InputDevice"
                Identifier "Gaming Mouse"
                Driver "evdev"
                Option "Name" "Logitech USB Gaming Mouse"
EndSection

Now getting touchpad on my laptop to do all the cool stuff (up and back scroll buttions, iPod-like circular scrolling, etc) was a little more involved. I set a udev rule to make the device name explicit and I had to find the configuration entry on the net and cut-and-paste...

udev rule created in "new" file /etc/udev/rules.d/10-input.rules:

BUS="serio", SYSFS{description}="i8042 Aux Port", KERNEL="event?", SYMLINK="input/alps"

Section "InputDevice"
        Identifier "Alps"
        Driver "synaptics"
        Option "Device" "/dev/input/alps"
        Option "CorePointer"
        Option "Protocol" "auto-dev"
        Option "LeftEdge" "130"
        Option "RightEdge" "840"
        Option "TopEdge" "130"
        Option "BottomEdge" "640"
        Option "FingerLow" "12"
        Option "FingerHigh" "15"
        Option "MaxTapTime" "180"
        Option "MaxTapMove" "200"
        Option "MaxDoubleTapTime" "100"
        Option "EmulateMidButtonTime" "75"
        Option "VertScrollDelta" "20"
        Option "HorizScrollDelta" "20"
        Option "MinSpeed" "0.60"
        Option "MaxSpeed" "1.10"
        Option "AccelFactor" "0.030"
        Option "UpDownScrolling" "1"
        Option "CircularScrolling" "1"
        Option "CircScrollDelta" "0.1"
        Option "CircScrollTrigger" "2"
        Option "SHMConfig" "on"
        Option "Emulate3Buttons" "on"
EndSection

This is on ubuntu with the current "stock" 2.6.15 kernel and Xorg packages.

It has to be said...

[Anonymous Coward]

My octopus has been after me for this feature for years.

Re:It has to be said...

[Haley's Comet]

yeah, but my pet fly is PISSED because not a single distribution has support for multiple monitor cards out-of-the-box!

Sounds neat, but...

[pain]

there are already applications that support multiple cursors. Take a software audio mixer for example. With a decent usb mixer controller you have so much input devices available that are mappable to one or two dimensional sliders to control your musique.

But sure. Multiple input devices in X would be great. The new API could be used to model the above application.

But don't forget to support my usb mixer input device.

Re:Oh yeah?

[Anonymous Coward]

Actually multi-player xpong sounds a lot better than 6 people dragging shit around on a table.

That's just my take.

Tactiva/Tactipad

[rthille]

This opens up X to the possibility of supporting something like the Tactipad. That's cool.

But I still think it's going to be awhile before I see the Tactipad at Best Buy...

[OT] Re:Reminds me of Anakin

[soren42]

Wow... Kim Jong Il as Darth Vader... I totally get it.

I can't believe I didn't make that connection alright... forget Offtopic... should be modded "Insightful".

Works as well as "Mr. Burns as Goa'uld"...

"Quote from an Albert Brooks stand-up routine.

[Anonymous Coward]

After the character takes a hit off of a joint: "I'm gonna buy me a Chevy with TWO STEERING WHEELS!"

Reminds me of Settlers

[Anonymous Coward]

http://en.wikipedia.org/wiki/The_Settlers

In 1993 already, you could play this DOS game splitscreen against your buddy with mouse and keyboard or mouse and mouse, e.g. one in a serial and the other one in a PS/2 port.

multi pointer x sever

[Anonymous Coward]

more like

multi pointless x server!

A little bit closer

[Schraegstrichpunkt]

We're now a little bit closer to Star Trek-style touch screens...

Geeze... talk about 1980's

[neo]

The Amiga was multi-cursor from the start. I used to play video games with my brother that used both mouse inputs. (Bethesda Softwares Football and Hockey games used it.) And now... after all these years... someone figured out how do to it in X-Windows.

Sigh.

If I am not mistaken...

[cr0sh]

...Didn't Douglas Englebert demonstrate such dual mouse input in his demo in 1968? I know there was dual-handed input (a mouse in one hand and a chorded-keyboard in the other), in addition to footpedals - but I could have sworn there was also a dual-mouse input system to allow two-handed resize and rotation of windows and other widgets.

Also, dual-input (if not dual cursors) handling by X is available today - I once hooked a serial mouse and a PS/2 mouse up at the same time to one of my Linux boxen, and by tweaking the xconfig just right I was able to use one mouse for the cursor, but read and utilize either mouse from inside Python (I was playing with an OpenGL app I was developing where I needed both inputs)...

Multiple Sessions

[cskrat]

What I would like to see is a way to 2 mice, 2 keyboards and 2 monitors to the same system to allow 2 different people to log in to the same machine with their own desktops. This could be useful in situations where you have multiple terminals sitting right next to each other since you would only have to purchase and maintain one system per two terminals. If you consider that most retail video cards support two independant displays and dual core processors are hitting the mainstream, the hardware is more than ready to cope with the demand but the software still has that annoying assumption that only one user can log on locally while all other sessions must log in over the network.

Think about public libraries or school computer labs where the disks are restored to a pristine state every night and nobody should be running any programs that are demanding enough to block another user. I could see this as being a viable alternative to the use of thin clients in the right situations.

Re:Multiple Sessions

[bcmm]

Can't find a link, but this was covered on /. a year or two ago. Some people made a cheap desktop for education in Africa or something, with 2 video cards and dual keyboards and mice (USB presumably). It acted more or less like a machine with two thin clients; you could log in independantly on either of the heads.

I don't think they had to do anything very special; AFAIK they were using pretty standard x86 hardware and (two seperate) stock Xorg servers (using 2 config files). Dunno how they did it, but I know that seperate mice (and presumably keyboards) have seperate files in /dev/, and at least the nVidia drivers allow you to specify the hardware address of the card to be used.

Wrists

[IdleByte]

Great, now both my wrists will go bad....

Finally, I can finger paint!

[Anonymous Coward]

Screw these styluses and crap, now I can finger paint on the computer!

wow, 2 mouse X....... That's Idea is so old!

[goslackware]

Douglas Engelbart had this back in the 60's. Take a look at his demo from back then. http://www.archive.org/details/AlanKeyD1987 http://www.archive.org/details/AlanKeyD1987_2 About: smalltalk (what squeak is) http://www.archive.org/details/DanIngal1989 Hopefully Alan Kay and his team will save us again with the squeak system, think of squeak as a whole new OS, but much more. I know that OpenCroquet project within squeak supports multi mouse www.opencroquet.org ~goslackware

Re:Oh yeah?

[CastrTroy]

It would be really nice if it supported multiple keyboards too. You could have a single meeting room, each seat with a mouse and keyboard, and when anybody wanted to put something on the screen, they could use their own mouse and keyboard. You wouldn't need a switch, and multiple people could be putting stuff on the screen at the same time. With USB you can theoretically support up to 128 devices, so there's a lot of different uses for this. Another good use would be 3d Design. Instead of buying a $700 spaceball, you could purchase 3 trackballs, and just use those.