What's the plan for touch-based devices?

Hi folks,

What if any consideration has been given to the suitability of
HAIKU to touch-based devices like tablets and phones?
These seem to be the future of personal computing, if not workplace computing.
A few of these devices use Atom processors, like the FonePad 7 from Asus,
which is a 7-inch phone, and their upcoming FonePad Tab 6.

See: http://www.asus.com/Tablets_Mobile/ASUS_Fonepad_7_ME372CG/

I’ve spent a few years coding for iOS and I am firmly of the
opinion that an old-time desktop UI should not be adapted
to the touch-only realm, as some people in KDE-land appear to be trying to do.
Too much needs to change as one goes from mouse-based to finger-based UI’s.

Phones present a special problem because the GSM drivers
are usually closed-source binaries intended for Linux/Android, but
there may be a few phones out there that can be done with FLOSS
GSM drivers; I’m reminded of the OpenMoko project.


Well, the major issue is to move to a touch-friendly UI, not to just move to a touch-aware UI.
That means larger touch targets UI objects, gestures taking over some interface methods that are only visual (and often realestate consuming even when of no use), and so on. But it also means to rethink from ground the old WIMP paradigm that govern the way we accomplish tasks on a desktop user interface.

It’s a whole redesign of the UI.
Haiku R1 can’t and won’t do that. At best, Haiku R1 may be touch-aware, aka allow touch input (display, most probably) to be used too, beside the classic desktop input devices.

What is interesting here is that, in fact, there is still no good productive/desktop/massive multitasking touch UI out there. The existing ones are often a (more or less failed) attempt to adapt the mobile touch UI paradigm to a fixed desktop station usage. While it make perfectly sense to touch a device that you needs to hold with you, it don’t always make as much sense when it comes to use a device that is immobile, larger, possibly quite heavy. Even with large touch screen, there is an issue regarding the size of a gesture (full horizontal swipe on a 24’ is a move that a finger alone can’t do without a harm move too, while a single finger can travel the whole display of a mobile device.

I guess one first should seek the current research made on that field, because clearly there is many things to avoid and others to be unique regarding a better user experience. Like 3D, it’s not because it sounds cool that it’s productive. There is still no actual 3D desktop environment that make more than just a woah factor. Beside Hollywood, nobody actually found them a better experience. Desktop Touch interface seems at similar place now to me.

Some points:
A mobile UI isn’t just touch friendly but also designed for the many sensors on a phone. For example, flip over to change from landscape to portrait.

I can see a desktop UI as enhanced by touch here and there to supplement WIMP where it makes sense. But to make the entire UI touch friendly is solution in search of a problem.

It could make sense as part of a game API so that you can play an Angry Birds type game on your PC without a mouse.

I would like to see an open source alternative to Android. It’s great to hear how all these Be devs are on the android team and that they finally came up with something that everyone uses but it’s also sad that Android is known to be sluggish, not at all the best choice for musicians and has major problems with updates. Android has more to do with linux, java and Danger than anything BeOS or even BeIA had to offer.

My feeling is that the mouse is dead. The argument has long been settled. Personally I haven’t used one in years because the touchpad is so much more facile and reduces the risk of repetitive-stress injury compared to the mouse.

People in the PC world may not realize it but on the Mac, there are many very useful touchpad gestures that significantly reduce the number of steps to do work – on the desktop – and for most consumers it takes very little effort to convince them of the benefit. Some of these gestures are also on tablets and phones too of course.

As far as shifting to a touch-based UI goes, if programmers are going to spend their time and effort anyway to learn the Haiku/Be GUI’s API, why stick with the current old-fashioned desktop model and its API? It would be smarter to direct them toward a newer touch-based UI’s API that works with (A) touchscreens and (B) touchpads.

Uniqueness is not necessary. That’s like MS-DOS advocates saying “we need a windowing system but it has to be unique”. No, it doesn’t. It needs to be as usable as the other systems.

This is a paradigm shift, not a market competition.

I’m not speaking for haiku or the dev team, etc here but R1 is the important task, ahead of any new ui, etc. Adding touch support would be a nice step for support measures but think about this for a minute. Phones and those little cute tablets are being produced at a rate the PC Never was. It’s not an amazing factor to realize once you actually look at the industry and how the devices are being marketed, created and consumed. It’s not really a matter of “they are the future”, which is a farce btw. It’s more a matter of who uses what, and what is more productive for what group of users.

Every user has different requirements depending on the tasks they need / want to perform. Engineers, Architects, Researchrs / Scientists, etc will hardly use a tablet for their everyday work. It’s fine for off little tasks to supplement but not for their mainstay work. They are smaller, but they are by no means more powerful. Do not assume that everyone will work via a phone or tablet or touch device as it’s not a productive solution for every user group. For a general, everyday user that just needs to check their e-mail and play small games like Angry Birds, it works great. For those of us that have real work to do, the touch screen is not a viable option. I have a tablet, a good one at that. Same with a smartphone. I’ve explored it for myself and I can tell you with 100% certainty that it’s not going to replace my workstation, at all. My workstation doesn’t need to replace it’s main interface every so often because of material failures.

I’ve said it for years though, a pentium with enough ram is all people really need for their general use. They don’t need a “supercomputer” on their desk to browse the web, which is what makes them more attractive. You should also realize the material cost of producing so many units like this every year. It’s not sustainable just as much as the pc production isn’t. The small device production is worse though, because so many units are produced, we’re going to run out of those resources if it continues to the degree (or worse) than it is now. I’m fairly certain that every one of those devices produced are not actually being consumed and used. I ask you, do you need to replace your desktop every year? You shouldn’t have to. I haven’t in 5 years, so why would I want to move to a platform that is dominated by the need to be replaced every year or two (if the device lasts that long)? I know others who haven’t replace their computers in 15 years. Which, yes, that’s just insane… but they didn’t. Remember my work dictates that I need a powerful device, so it’s not so much a want as a need. iAnything doesn’t cut it, and neither does the android. There isn’t a one size fits all solution there.

I just want people to think about the big picture overall. You can’t just throw out those user groups and assume they will be forced to change. Not all of them will. There are also the gamers. I don’t know of one gamer that wants a tablet for their games. It’s almost heresy to them to even suggest that they play their fps or mmos or anything really on their tablets.

If it works for you that’s great, though, it’s not going to be everyone’s main computing device. I also think people forget, we’ve been here before with touch devices. It’s another of the same cycle. Mostly the same tech, just updated and refaced. As for haiku I wouldn’t push to drastically change the os to touch like that. I highly doubt it would shift there, thankfully so. Though after R1 and stability comes in, I don’t doubt that it would eventually be a supported platform, which is great. Though, it has to be stable in it’s current incarnation first.

First of all, I’m not AGAINST touch devices. What I’m saying here, bottom line is, it has to wait till a full stable release is done. I’m NOT attacking your standpoint. What I’m really saying here is “be patient”. I don’t doubt it will eventually come. I’m not arguing against your want for a touch interface. What I’m telling you is that it’s simply not for everyone. I have a tablet pad for art and design work. I’d love that to work right now. Though I realize it’s not as important as a stable and cohesive os. That has to come first. Then I’ll worry about the pad, and any other touch interface. Would I use a touch device to do my other work (including a touchpad mouse)? No. It’s simply not a productive enough interface for me. I also have an iPad and android tablet… Guess what I don’t EVER use… Either of them. No need to. No reason to, and never will have a viable reason to have them except for development and design work.

I’ve used OS/X, extensively… no thanks. Gestures have been around for a loooooong time and they existed before OS/X. So have touch interfaces btw. I used to have a super nintendo controller that was touch based… it sucked. Doesn’t mean they all do, just that particular input device sucks as a touch based device. If you like a touchpad mouse, GREAT! Keep it… use it… Last I checked a touchpad mouse should work ok, depending on the manufacturer, with haiku. Some of us prefer workstations to a macbook pro too. Does that mean I want to take away your macbook? No, it doesn’t.

What evidence exists that anyone will play games with HAIKU?

Read the forums and pay attention. The evidence is right there. There are more users requesting games and 3d acceleration and some won’t come over till that happens. Some are waiting for that to happen.

Will they switch to HAIKU? By your logic they will not.

Your misinterpreting what I’m actually saying. I work with engineers and architects directly, btw. Every one of them I’ve interviewed on the subject (touch interfaces and devices for their work) can’t stand touch devices and would prefer a mouse. Touch and gestures get in the way of their work when dealing with 3d modeling and manipulation, blueprinting, and yes, they’ve TRIED it with many touch devices. They usually end up cursing and plugging in their mouse again, if they can.

You mention people with old PCs who never change.

Yes, yes I did… because I know too many of these people who are STILL running windows xp and just waiting for something to come along and help out with that. Middle america doesn’t have the money to buy new devices that run at half the horsepower a desktop will, nor do they. These people would love haiku, I know a few of them already are using it now. Simple point is, it suits their needs and runs. They aren’t worried about anything else.

What you call a paradigm shift I’m just going to call recurring fad, again… Almost every tech for “touch” has been done, and done and done again. It’s been done so much that those who remember it, are over it for the most part. Didn’t work for them then, it’s not going to work for them now and that’s not my opinion. To me, it’s a matter of doing it right. There have been so many interface devices created for a computer (of any type), it’s staggering. Touch isn’t the latest and greatest of these either. I don’t know if you remember or not, but kiosks in stores and malls were being installed in the early 90s that were 100% touch driven. Much like there are now. That was really the beginning there. It’s taken the better part of 30 years for them to get it to this point.

So don’t take my response as an attack, because it’s not… I know the devs are crazy busy working on R1… If you love OS/X so much, please stay with it and use it. I don’t care. I do care that the os is stable and functional before I care if any touch device is ever implemented. That’s the pragmatist in me and that’s not meant to offend you or anyone else either.

Just “be patient”… That’s all. I have no problem with haiku including a touch interface, or even to go as far as including an interface shift for a touch device. I do care that it’s optional and stays optional and doesn’t become the only interface to haiku.

Every user has different requirements depending on the tasks they need / want to perform. Engineers, Architects, Researchrs / Scientists, etc will hardly use a tablet for their everyday work.[/quote]

This seems like a non sequitur. You mentioned a number of groups who don’t want touch devices as if that were justification for not adapting HAIKU to touch input.

But it isn’t.

What evidence do you have that engineers, architects, and scientists plan to use Haiku, or would use it if they knew about it? I would think those groups would require compatibility with mainstream software including X Windows (imperfect as it is).

You mention gamers. What evidence exists that anyone will play games with HAIKU?

You mention people with old PCs who never change. Will they switch to HAIKU? By your logic they will not.

So if none of the groups you presented are going to use HAIKU, you’re not offering a strong argument against adapting HAIKU to touch input.

Also, while my original post mentioned phones and tablets, not every touch based device is a tablet or phone. There is a paradigm shift happening toward touch input. My Macbook Pro is a touch-based device, because I’ve never used a mouse with it. I rarely even click. The touchpad as it is used on OS/X is as sophisticated if not more as the touch gestures of the iPad. (If you’re not an OS/X user you may have no idea how facile it is, so go to an Apple store and try it out.)

i very much prefer physical controls to touchscreens for music. mild arthritis means an unyielding touchscreen is pure torture before long.

not that anything’s stopping anyone from making a touch-enabled app (i.e. tunetracker).

Musicians will probably favor a touch based UI. I think it may even be part of TuneTracker.

There is no plan for touch based devices.

Not only it would imply a huge amount of work, but there is also the problem, that enough people consider the haiku ui “perfect”. I remember when stippi was proposing a new and more functional deskbar , and people where like screaming: no don’t touch it, it’s already perfect. etc

What some people don’t understand is that touch-screen doesn’t need to be used all the time as a touchscreen, just when you needed it, and use a mouse when a mouse feels better. That’s why there is in fact no reason to be against touch-screens.

That a new UI can’t be done before R1 is somehow clear, but I have the impression that also after R1 it’s nearly impossible. And one of reasons is that also the willingness is missing, and that some people are stuck in the 90’s when it comes to UI.

One observation is that the hardware seems to start to become all the time smaller and thinner, and I guess this trend will continue as long as possible.
One day you might ask you, why to carry around, or have in the house a quite big device, if you can have a small glasses that can be used as the display.
And why not let the phone be a computer too (e.g. ubuntu touch) , and use it as a phone, but when needed you connected it to glasses, or you connect it to a beamer (projector).
For most people the power of a smartphone would be enough to be also used as a “regular computer”. The others who more power, can buy a computer with more power but without screen (like the apple mini) , and connect this one to the glasses or beamer too or any other display. [And yes, if you need to you connect bluetooth keyboard and mouse too].
The point is: when you are mobile, in most cases you don’t need the power of a high-end-desktop-pc. And if it happens that you need a lot of power (like 3D CAD applications etc) then you often are not “mobile”, you are mostly at home or at your office.

That means, I guess, that the desktop will still be available for longer time, but the laptop has changes to be transformed over time to something more “invisible” (like a phone+glasses).

For all this changes haiku will not be prepared, because at takes many years, and for big changes you have to start preparing early. Ubuntu is already now preparing with Ubuntu Touch , for the time when the tablet, phone, pc will be (or might be) the same.

I think a UI of the future would be one that is adaptable. Adaptable to the screen sizes, but also adaptable to the location, perhaps even the mood of the user [working/productive mood, party mood] (yes, kde activities are something similar)… etc…

Don’t get me wrong, I’m not asking all this from the Haiku devs, by far not. I just say, how I guess that the computer/os of the future will be (or seems to become) (or partially is already).
I also don’t like at all the people popping up with all kind of ideas, but are not willing to invest own efforts in making it possible.

Also “google now” makes sense to me.

Perhaps what haiku would need, is a completely new and unique feature, something remarkable. This could bring new users and developers. But yes, nobody has time, at also no ideas. But perhaps it could be one idea, to make something like a contest, and let people come up with various ideas, that don’t exist yet.


I simply don’t believe you’re making a good argument.

You’ve basically stated your preference and you claimed to know the preferences of other individuals that you spoke to. To me, that’s too small a sample set with which to plan software development. Even if one generalizes to classes of people, let’s say CAD or CAE users, what reason do you have to think they will want to give up e.g. AutoCAD and switch to Haiku, a platform for which there is no equivalent program?

As for the gamers wanting to use Haiku, that’s simply unbelievable. Gamers prefer gaming systems with high-end hardware from big companies that have 1000’s of games written for them. If you believe gamers when they say they’ll use Haiku if you do XYZ, you’re letting them lead you by the nose.


If I could summarize what you said:

  1. There are different UI modalities: touch-screen, touch pad, voice, touch sensors e.g. on glasses, keyboard and mouse.
  2. There are places to display a UI: laptop screen, tablet screen, phone screen, large desktop display e.g. for CAD, projected display, TV etc.
  3. A touch UI does not prevent mouse input or make a 1990's desktop experience less possible (I agree).
  4. Haiku at present provides nothing but the 1990's desktop experience. This is also true of GNU/Linux (except for the Android variant).

I think the short-term goal should be maximum coverage of the above inputs and outputs.
The earlier that work begins the less that has to be thrown out later.

As for something that would make Haiku remarkable: how about a focus on privacy? You mentioned Google, and I’ve mentioned Apple. These are companies that are complicit in government spying on their own customers’ daily communications. We must assume that worse offenses have been committed e.g. allowing use of backdoors to get into individuals’ computers. Most likely Ubuntu is doing the same since their CEO is a former spyware writer. All of the recent media revelations about spying provide a good reason to have a safer alternative. Perhaps another thread could be started for the privacy topic.

I simply don’t care what you think about my so called “argument”. Guess what, it’s not really an argument I’m proposing. Read cipri’s post… He hit the nail right on the head. Study his post because he’s right…

It may not ever happen. If it does, awesome. If not, eh. If you want to do a poll, great. By all means, please proceed. Let the community here know your results. Start a thread about the polling subject and have the community help you on the poll subject as well. Let the developers give you their 2 cents on the subject.

I don’t think you really understand what I’m getting at. Haiku wasn’t meant to be a tablet, phone or anything but desktop os. I imagine it will support a touchpad mouse… Outside of that, it’s uncertain.

When I say gamer, I’m referring to PC Gamers. Which equates to a windows gamer. Whether you like it or not. They play consoles second. Yes, they have high end machines. No, they don’t like windows but will use it to play their latest battlefield, cod, dota, etc, etc. I’m fairly certain there is a large portion of the world’s pc gamers that don’t care. They will go where their games are, that simple. I am a gamedev, and I get a ton of email asking when a game is going to be on x or y console, or if we’ve looked into other operating systems like linux or haiku. Believe me or not, I don’t care. Again, it’s inconsequential to me. I also run a software firm for 3rd party applications and interfaces for engineering and architecture. I’m well aware of what the industry is actually using and what the preferences are. There is also Catia… outside of autodesk entirely, and there are even more than that.

Do something with your poll. I’m not going to stop you, nor am I against it. I would actually encourage you to do so. Just don’t go around and get pissy with people when they don’t agree with your viewpoint, that’s all.

If your waiting for haiku to become to next big thing for touch, your going to be waiting a very long time.

This may twist your brain a bit, but I do agree with you and cipri. A mouse can still be used with a touch interface, and believe it or not vice versa. It just has to be done intelligently is all. My original point was, do not stomp the current users just to get a touch interface. It’s that simple. I also agree with your point of better security and privacy. Though, just realize that most of the time, the users also pose a risk to themselves because of their simple actions. If they aren’t careful, it’ll be a lot harder to protect them.

I still don’t see how touch screens can work on a desktop. But for laptops designed with touch in mind (e.g. the psion netbook http://www.toptenreviews.com/i/rev/misc/articles/7996/5-close-relativ-5.jpg) or tablets I think taking a leaf from Psions EPOC OS could be good. There have been many advances in touch interfaces since, some of them no doubt useful, but EPOC is a very elegant middle ground between a functional productivity UI and the touch paradigm.

Phones are a different matter, but I can’t really see a place for haiku in phones.

As always when this old canard is wheeled out, I say: “For people who’d like to judge themselves if [people where like screaming: no don’t touch it, it’s already perfect. etc] or if people stated their reasons for liking many of the current Deskbar’s features, here’s the thread from back then.”
I’m not saying there aren’t some who think the GUI is perfect as is, but implying those are the people driving Haiku’s development is wrong in my perception.


I have seen for example axeld, when I made a preview of the documentviewer and some devs (like axeld) was against the ui in that application.

I don’t know how many like the haiku ui, but to me it was a surprize that when also devs where against doing improvements to the deskbar … etc.

I don’t know if those are the people driving haiku’s development, but seems there are people in haiku’s leadership who can stop it.

I have the impression that the people who grow up with haiku’s ui, are happy with it.

But perhaps, haiku should target “new people” who don’t have a connection to beos, otherwise how is haiku going to survive (we all have a biologic clock)?

I guess, if haiku wants to improve, it needs a lot of support, but for that you also many users, who can support it.

At the moment, I even don’t know how haiku is doing. Does it have more users than 3 years ago? How is it growing, or is it falling?

What you can learn from seth godin, is that if you need to be remarkable.
And sice the UI is the most visible, it has a big impact to many (especially new people).

The first time I tried haiku, i was thinking, what a uncomfortable UI, and old-fashion-look.

I understand that the ui should perhaps not change before R1. But I don’t have any confidence, that after R1 the ui is going to be get an big update. Because it doesn’t look to be an interesting subject to the devs.

I guess, and hope the R1 is not that far away, but still there are not discussions in respect to plans to the future ui. I guess this are good hints, that the ui is not really something urgent to haiku, even after R1.

Honestly, I think the devs may be more concerned about a wildly moving target, than they are a ui reface. Just don’t know on their end. Some improvements would be nice, and welcome. Though again, stability is king. Can’t be stable if everything is in constant flux.

People change their goals over time. A project that doesn’t adapt to new circumstances is not proceeding wisely.

One thing people are doing nowadays is applying the idea of CSS to static GUIs, so that if the screen is of one type and/or size, use one set of CSS e.g. specifying a button to be 40 by 20 pixels, but if it’s a touch-sensitive display, use a different CSS file containing a set of finger-size dimensions e.g. 72x37 pixels for a button, and maybe different text choices.

Similarly if you have a multitouch input, pinch-zoom about a point is possible, but if not, Meta-equals for zoom in, meta-minus for zoom out and these are about the center of the document.

The reason for putting things like this in a file is that you can adapt an existing program to new inputs and outputs without always changing C++ code.

Agreed. I think the UI of Haiku looks very nice in terms of color choices, fonts, and proportions… for a desktop UI, but as a desktop UI it’s still stuck in the past.

Could be worse. The typical Linux UI IMHO looks gaudy and works horribly. I look at KDE and I’m reminded of the Spanish adjective sobreproducida, usually applied to women who use too much makeup.

Seth Godin:

i first booted haiku a couple months ago. the ui struck me as clean, uncluttered, and colorful – as opposed to the post-2001 trend of increasingly monochromatic presentation of technology. looking for further simplicity, i moved from gnome 3 to kde 4.7 to openbox before booting haiku and finding an interface similar to the latter, but with the addition of dynamic tiling out of the box. that was kind of incredible and new – under linux wms, i tend to manually tile windows, which is tedious to set up, and i’d been pining after dynamic tiling for quite some time (never got xmonad running in manjaro linux, nor have i had any luck running dynamic tiling scripts written for openbox) – it’s kind of great, and from what i can gather, pretty new. anyone clamoring for a more modern interface, which i understand as a minimalist aesthetic that was all the rage in the ‘80s (back when computer guis couldn’t emulate it; back in steve jobs’ wilder years) can totally use qt or java apps.

fixes to the gui i’d appreciate are things like larger handles for replicants, tiling stacked groups of windows, maybe better multitouch trackpad support (two-fingered scrolling works maybe every other boot), and then everything else is kind of app specific (cortex needs some love and documentation).

really, though, the gui’s the last thing in a long list of things that needs to be worked on right now, especially as extensively as is being proposed in this thread.