You are not logged in.

#1 2011-03-06 08:04:21

smartass
Member
From: Prague, Czech Republic
Registered: 2011-02-25
Posts: 60

In search of a truly innovative Desktop

Hi, I'd like to bring this topic up on this forum, because I know there are many (former wink ) distro-hoppers and that's why I'd like to raise my question here.

After reading an article about Linux desktop environments mimicking other OS (let's face it, in some ways KDE resembles Vista and Gnome 2.30 resembles OSX a bit) and thus reducing the computing speed and uniqueness that Linux offers, I looked for alternatives. E17 looks promising, lightweight WMs are attractive too, but they all have one thing in common:

They all use windows as the main entity of the GUI:

http://en.wikipedia.org/wiki/Window_%28computing%29 wrote:

window is a visual area containing some kind of user interface. It usually has a rectangular shape that can overlap with the area of other windows. It displays the output of and may allow input to one or more processes.

And they display them flat:

http://en.wikipedia.org/wiki/Graphical_user_interface wrote:

The term GUI is historically restricted to the scope of two-dimensional display screens with display resolutions capable of describing generic information, in the tradition of the computer science research at the Palo Alto Research Center (PARC). The term GUI earlier might have been applicable to other high-resolution types of interfaces that are non-generic, such as videogames, or not restricted to flat screens, like volumetric displays.

So I'm asking you people here, have you on your Linux/GNU journey discovered a completely different GUI concept for which you don't need a volumetric screen?
Yes, CLI could be seen as the answer, but doesn't count this time (maybe except for ncurses or similar), because it has a rather restricted support for graphics output wink

For example this is what I found: http://s3d.berlios.de/ Still uses windows, but as a WM, it is very different.

Last edited by smartass (2011-03-06 13:02:48)


KISS my Arch, Willy Gates!

Offline

#2 2011-03-06 08:28:41

tlvb
Member
From: Sweden
Registered: 2008-10-06
Posts: 297
Website

Re: In search of a truly innovative Desktop

When text is the only thing to be displayed, using graphics seems quite unneccesary, using a 3d environment even more so,
as even most graphical content of today is flat. Cases can be made for all kinds of user interfaces, but the key is to use them
when the work at hand fits their managment style. Eg in the 3d example a representation of a web page would still only be
a flat window, no matter how deep you make the surroundings (or it would not render correctly).

Also, for checking out different paradigms, take a look at tiling window managers, but saying that the windows mac and linux
ui experiences are sort of the same is akin to stating that both bicycles and cars have wheels, of course they do, it has been a
method proven through time to be a very effective way of getting you places.

Last edited by tlvb (2011-03-06 08:30:03)


I need a sorted list of all random numbers, so that I can retrieve a suitable one later with a binary search instead of having to iterate through the generation process every time.

Offline

#3 2011-03-06 08:44:22

smartass
Member
From: Prague, Czech Republic
Registered: 2011-02-25
Posts: 60

Re: In search of a truly innovative Desktop

Good point smile But even though bicycles, carriages, cars, trains, etc. are used the most, man in his curiosity attempted to build a flying machine and after many helpless and pathetic attempts, we have airplanes, helicopters and space shuttles. I'm only asking you to dig in your memory for the open-source Wright brothers wink


KISS my Arch, Willy Gates!

Offline

#4 2011-03-06 08:51:37

dmz
Member
From: Sweden
Registered: 2008-08-27
Posts: 881
Website

Re: In search of a truly innovative Desktop

Ratpoison and Stumpwm does not use windows, but frames. This means you can have two applications in one frame, on top of each other, and simply flip the frame to switch to the other application.

Last edited by dmz (2011-03-06 08:52:33)

Offline

#5 2011-03-06 12:38:17

rwd
Member
Registered: 2009-02-08
Posts: 664

Re: In search of a truly innovative Desktop

There have been a few attempts at 3d desktops and now gesture recognition added to it. But for working with text I  don't think there is anything better than  a keyboard and text displayed in 2d.  So I suspect what will happen is not people usiing 3d desktops , but rather that  people will not use a desktop/laptop for certain activities anymore. Things like the fridge that will order groceries by itself, and   new uses for smartphones. Also more clippy-like software that thinks it knows what the user wants and recognizes facial expressions and context so that you need to press less buttons.

Last edited by rwd (2011-03-06 12:47:56)

Offline

#6 2011-03-06 13:02:15

smartass
Member
From: Prague, Czech Republic
Registered: 2011-02-25
Posts: 60

Re: In search of a truly innovative Desktop

I'm not saying 3D is a good choice for a flat screen, the example I gave was just to show something different, I'm just asking around for innovative GUIs people discovered.

The question is, whether there is or can be at all a non-window GUI concept, no matter how many dimensions (though 1 could be a little unusable).

@rwd: even though that may happen, because laziness is natural, I meant a GUI as in a way to communicate with your PC in order to use the speed and computing capacity of the computer to aid your mental processes, for example manipulating scientific data and analyzing them, looking at it from different frames of reference, etc. What I'm asking is whether there is a more efficient way to tell your computer to do things like that than by using window GUIs or CLI.

Last edited by smartass (2011-03-06 13:13:09)


KISS my Arch, Willy Gates!

Offline

#7 2011-03-06 14:18:10

tlvb
Member
From: Sweden
Registered: 2008-10-06
Posts: 297
Website

Re: In search of a truly innovative Desktop

I would say that in those cases the interaction method would be quite specific to the task at hand and difficult to propagate out to something as general
as what is currently the window manager/desktop environment.

Altough I have some times wished for something akin to pure data or labview on top of the regular gui, in order to bring the goodness and power that is shell programming into the gui world and to some extent merge them. (I don't know exactly where this would be useful, but I think it is a Good idea)


I need a sorted list of all random numbers, so that I can retrieve a suitable one later with a binary search instead of having to iterate through the generation process every time.

Offline

#8 2011-03-06 14:29:34

smartass
Member
From: Prague, Czech Republic
Registered: 2011-02-25
Posts: 60

Re: In search of a truly innovative Desktop

Hmm, it is true that CLI/Shell has some awesome powers, but I would argue that most of these powers come from the fact that it is CLI and not GUI (building complicated and "intelligent" command sequences and structures with many possible combinations of switches over point-and-click adventure with some keyboard shortcuts designated to usually only one action, possibly context-sensitive). So I don't see integrating CLI into GUI as something possible, because they are so different in interaction.

But if you disagree, I'm very interested in your ideas on how to accomplish that smile

Last edited by smartass (2011-03-06 14:39:46)


KISS my Arch, Willy Gates!

Offline

#9 2011-03-06 15:38:51

ngoonee
Forum Fellow
From: Between Thailand and Singapore
Registered: 2009-03-17
Posts: 7,354

Re: In search of a truly innovative Desktop

tlvb wrote:

Altough I have some times wished for something akin to pure data or labview on top of the regular gui, in order to bring the goodness and power that is shell programming into the gui world and to some extent merge them. (I don't know exactly where this would be useful, but I think it is a Good idea)

Have you actually USED labview? Its good for its specific purpose, but trying to use it as an actual programming language is just frustrating. Let's all play the 'click on this specific pixel' game smile


Allan-Volunteer on the (topic being discussed) mailn lists. You never get the people who matters attention on the forums.
jasonwryan-Installing Arch is a measure of your literacy. Maintaining Arch is a measure of your diligence. Contributing to Arch is a measure of your competence.
Griemak-Bleeding edge, not bleeding flat. Edge denotes falls will occur from time to time. Bring your own parachute.

Offline

#10 2011-03-06 15:58:10

smartass
Member
From: Prague, Czech Republic
Registered: 2011-02-25
Posts: 60

Re: In search of a truly innovative Desktop

@ngoonee: Luckily, Linux/GNU offers sooo many ways to accomplish one specific task. For example, my tutor uses makefiles instead of normal *.sh scripts, simply because he prefers the declaration of macros and targets over functions. It's quite cumbersome, but it works. So once again, Linux/GNU is about choice and I'm trying to discover new possibilities which may suit my tasks better.

Back to the main topic: I think ngoonee actually brought up an important issue:

ngoonee wrote:

Let's all play the 'click on this specific pixel' game

With the rising usage of touchscreens, there may be a need for a different approach to GUIs, because pressing normal buttons in/on windows with fingers (like trying to close a window - can be frustrating)  may not be precise enough
and if you use a different window geometry factor, everything will be bigger and easier to touch correctly, but at the same time most touchscreen devices are smaller than 15".


KISS my Arch, Willy Gates!

Offline

#11 2011-03-06 16:15:50

rwd
Member
Registered: 2009-02-08
Posts: 664

Re: In search of a truly innovative Desktop

smartass wrote:

@rwd: even though that may happen, because laziness is natural, I meant a GUI as in a way to communicate with your PC in order to use the speed and computing capacity of the computer to aid your mental processes

I don't think not using a keyboard and screen is lazyness, but rather specialization. I would like to be able to ask a computer more general things, as if I would ask another person, instead of having to specify every detail. That way I can spend time on more useful things. For the same reason that people rarely program in assembler anymore.

Last edited by rwd (2011-03-06 22:05:35)

Offline

#12 2011-03-06 16:34:28

smartass
Member
From: Prague, Czech Republic
Registered: 2011-02-25
Posts: 60

Re: In search of a truly innovative Desktop

@rwd: Well I'm afraid that this is still sci-fi, or at least not fully usable yet. Even if speech and face recognition got to that level, processing your general question/command would take a large portion of your computer's computing (I assume you have a computer with only a few GHz and GB of RAM wink) capacity, maybe it would take even longer than specifying more details and may not even yield the right results. For example Google has invested a lot of money into that Android voice search app, but it still provides only search results  which may not be 100% relevant to your question so you have to choose the answer anyway, even though Google is doing their best.

So let's stick to the ground and talk about present GUIs which may not be as automated, but really do what you want.
On the other hand rwd is right, GUIs sometimes lack the possibility to easily provide all details.

Last edited by smartass (2011-03-06 16:42:21)


KISS my Arch, Willy Gates!

Offline

#13 2011-03-06 17:18:37

tlvb
Member
From: Sweden
Registered: 2008-10-06
Posts: 297
Website

Re: In search of a truly innovative Desktop

ngoonee wrote:

Have you actually USED labview? Its good for its specific purpose, but trying to use it as an actual programming language is just frustrating. Let's all play the 'click on this specific pixel' game smile

Yes I have (as a lab assistant and developing with it on a summer job). I meant it should be done as an interface bridge as instead of doing 'put output in xwndfind $title $prop ->hbox->vbox->vbox->textentry' one could just drag an arrow to the textentry in question, but I think that an implementation should contain more power than that. Also labview's problems should not be taken to mean that graphical programming is dead, even though there are few examples of a succesful graphical programming languages...

Last edited by tlvb (2011-03-06 17:19:58)


I need a sorted list of all random numbers, so that I can retrieve a suitable one later with a binary search instead of having to iterate through the generation process every time.

Offline

#14 2011-03-06 17:41:20

ANOKNUSA
Member
Registered: 2010-10-22
Posts: 2,141

Re: In search of a truly innovative Desktop

smartass wrote:

Well I'm afraid that this is still sci-fi, or at least not fully usable yet.

I think you just hit on the crux of the argument against what you're looking for.  I think the closest thing to what you're asking for I've ever come across is Project Wonderland, and it isn't hard to figure out why that didn't catch on.  When it comes to information and communication technology, efficiency is tantamount to progress and innovation.  Having to click-and-drag a 3D virtual desktop around to find an open application, or walk an avatar over to a virtual water cooler to chat with someone online, or flip through a "shelf" of open containers clogging up screen space (see Project Looking Glass), might look cool. However, in the end it's just a waste of screen space, resources and time, and a massive amount of superfluous code to be written and maintained.  The appeal of say, a holographic 3D interface (a la the Jarvis interface in the Iron Man movie) is that it would be quicker to physically move application containers out of view than to use a mouse to minimize them.  Until that tech becomes available flat windows on the flat screens available to us are the most feasible option, and in the case of 3D holographic projection, flat containers for different projects/activities/apps would still be the most reasonable way to manage most tasks.

I'd also add that the uptick in touchscreen software development has more to do with the "cool" factor than true innovation, as present touchscreen technology lacks the precision and speed of buttons and keys, and sacrifices the speed of hardware-based calculation for the software-only approach.  This same "coolness" accounts for the use of MS Kinect on home computers, as well as the hideous "coffee table pc" that recently debuted.  Look at videos of these latter two technologies in action.  What are these folks doing?  In one case I found, motion capture video in the living room (totally improving our lives somehow roll); in most every other case, quick-and-dirty resizing of digital images, with no real precision.  Not exactly useful, if you ask me.  Of course if you're curious, Looking Glass and Wonderland are still available to play with, though neither is actively developed. smile

Offline

#15 2011-03-06 18:14:58

smartass
Member
From: Prague, Czech Republic
Registered: 2011-02-25
Posts: 60

Re: In search of a truly innovative Desktop

@tlvb: I agree that visual programming could be seen as a GUI concept, but it's aimed at constructing scenarios, not really interaction IMHO. But you're right that such features as redirecting resources, visual piping of programs would be a nice feature, for now we have drag 'n drop, which is somewhat close.

@anoknusa: Jeeez, I present one 3D oriented example and everyone now thinks I'm a 3D DE addict big_smile
Seriously, I used that example only to show an IMHO very useful and time sparing GUI for viewing networking relations, which could be a good starting point for a DE, because many core components of a Linux/GNU desktop are server-based (Xorg, cups,etc..)

Last edited by smartass (2011-03-06 18:32:02)


KISS my Arch, Willy Gates!

Offline

#16 2011-03-06 20:08:17

davidm
Member
Registered: 2009-04-25
Posts: 371

Re: In search of a truly innovative Desktop

I think much of the innovation yet to come has to do with turning the WM and associated applications less into a stupid series of instructions controlled by the user and more into an assistant who intelligently attempts to make predictive decisions on what to do in various cases.  Certainly strong AI would work best here (obviously rather Sci-fi atm) but isn't fully necessary to make something which is at least useful. 

Integration between applications and the WM is also another area and by this I do not merely mean how things such as windows are displayed and placed but more in depth integration. For example if historically 99% of the time after you receive an article in your feed reader containing a certain keyword in the title you open it up and save it, then it might be good and beneficial behavior for the WM and feed reader in question to work together based on your historical actions in those circumstances.  The WM should probably immediately bring that message to the forefront (depending on other complex factors) and perhaps the feed reader should just go ahead and automatically save the article and pop up a small notation in the corner informing you of this and allowing you to undo it with one click.

As for the 3D WM example, it's interesting and I see potential but at the moment I often find those implementations to be mostly eye candy.  We have to actually do things with the technology in a way which is useful to the user beyond merely looking pretty. Perhaps if we combine it with a more intelligent desktop we can do that.  For instance I can watch a movie in the background while doing some work. Once it's noticed that I am watching the movie more than working (maybe by watching eye movements) perhaps the display should adapt transparency, focus and size to favor the movie.  But if certain events occur in the other open applications perhaps the WM should do the reverse and favor the application.  Maybe sort of "cast" it by letting it fade in enough to where it is noticeable and if my eye movements seem to acknowledge it but then move back to the movie, then perhaps the WM should automatically fade it away again as far as transparency, size, and focus.  Just some ideas.  You get the picture.  Obviously there is A LOT of potential here and it's great to think about this stuff. smile

Last edited by davidm (2011-03-06 20:09:08)

Offline

#17 2011-03-06 20:51:09

smartass
Member
From: Prague, Czech Republic
Registered: 2011-02-25
Posts: 60

Re: In search of a truly innovative Desktop

@davidm: Eghm, no offense, but I hope nobody will send you to /dev/null (or worse - Windows) for that big_smile
The thing which I dislike about "intelligent learning" environments, is that if you suddenly break your habit and need to do something completely out of the ordinary, it may not be as easy as it would be without such a "helper", hopefully there would be an option to disable the helper temporarily. Another thing is that such helpers usually work only within a limited range of habits which they presume you don't combine, because no heuristic algorithm can analyze more complicated human thoughts (yet).

I guess that is more of a subject to personal taste, either to have a deterministic system which you configure specifically to your needs, or a machine that does most things for you automatically.

But you're right, integration is one of the keys to an effective environment.

Last edited by smartass (2011-03-06 20:58:34)


KISS my Arch, Willy Gates!

Offline

#18 2011-03-06 21:46:05

davidm
Member
Registered: 2009-04-25
Posts: 371

Re: In search of a truly innovative Desktop

Yes, my apologies, it was a bit wordy and quickly written. Windows?? No I use stumpwm and Xmonad. Not exactly a windows type of guy.  I just love the idea of making things more productive and useful.  Specifically in regards to being able to multi-task.

The point regarding an irregular action is valid but could be minimized by utilizing a thoughtful design which strives not to be intrusive yet still useful with a great majority of it's actions.  I think that is one of the keys in implementing it. Of course being able to disable it would be a given, but the goal would be to make it so that one would not want to do this. smile

Last edited by davidm (2011-03-06 21:46:43)

Offline

#19 2011-03-06 22:00:48

smartass
Member
From: Prague, Czech Republic
Registered: 2011-02-25
Posts: 60

Re: In search of a truly innovative Desktop

@davidm: And I apologize for questioning the purity of your soul big_smile

I actually use a dynamic helper with such behavior: preload
Doesn't have much to do with the GUI, but it's adaptive. Do you know of any other apps/daemons that dynamically adapt to your behavior? (some may call it non-KISS, but I'm just curious;))


KISS my Arch, Willy Gates!

Offline

#20 2011-03-06 22:15:44

rwd
Member
Registered: 2009-02-08
Posts: 664

Re: In search of a truly innovative Desktop

I can imagine that the pattern recognition logic that is now used for surveillance and serving ads could be used for userinterfaces too. Like in 'people who do x and y tend to prefer z too', and if you move around the userinterface in a certain way the computer knows that you're an experienced user that doesn't need hand-holding.

Offline

#21 2011-03-06 22:41:47

tlvb
Member
From: Sweden
Registered: 2008-10-06
Posts: 297
Website

Re: In search of a truly innovative Desktop

That is fine if you have only a few tasks to be done, but if switching tasks often or having many different tasks I think that one would start getting
irritated by the buttons moving around/dissappearing/appearing.
Eh, seems I only read the first sentence, ...detecting those 'pro' usage patterns would still be difficult, on par or worse than natural language,
these action having no predefined grammar that the ai could build on...

Last edited by tlvb (2011-03-06 22:46:44)


I need a sorted list of all random numbers, so that I can retrieve a suitable one later with a binary search instead of having to iterate through the generation process every time.

Offline

#22 2011-03-06 22:42:58

Awebb
Member
Registered: 2010-05-06
Posts: 6,275

Re: In search of a truly innovative Desktop

rwd wrote:

the computer knows that you're an experienced user that doesn't need hand-holding.

So it know's you're cool and throws you back to a shell and does nothing else? :-D

Online

#23 2011-03-06 22:48:10

rwd
Member
Registered: 2009-02-08
Posts: 664

Re: In search of a truly innovative Desktop

Awebb wrote:
rwd wrote:

the computer knows that you're an experienced user that doesn't need hand-holding.

So it know's you're cool and throws you back to a shell and does nothing else? :-D

exactly, And if this happens you'll be put back in a nice easy gui with colorful bottons:
Linux_Diskussion_by_hmdllr.jpg

Offline

#24 2011-03-06 22:58:45

davidm
Member
Registered: 2009-04-25
Posts: 371

Re: In search of a truly innovative Desktop

smartass wrote:

@davidm: And I apologize for questioning the purity of your soul big_smile

I actually use a dynamic helper with such behavior: preload
Doesn't have much to do with the GUI, but it's adaptive. Do you know of any other apps/daemons that dynamically adapt to your behavior? (some may call it non-KISS, but I'm just curious;))

I'm not aware of anything to the degree that I consider.  Just single purpose things which accomplish a trivial task.

<long>

In thinking I believe it would make sense to modularize it rather than to just build it fully into the application itself. The application would need to be able to at least send relevant messages to a main control module and would need to have this ability programmed it (e.g. for an IM app an Event_Id for a received message perhaps with different sub-IDs for the sender of the message).  Likewise the WM would need to receive and act on messages from the control module (to adjust focus, size, transparency, etc of a given app).  The control module would be responsible for acting on various events and deciding what to do based on input from the other components.  Obviously the control module is the most difficult part.  The application would need to generate and send a considerable amount of messages for given actions in order for the control module to be able to have enough data on which to dynamically adapt/"learn". The application would likely have to have an initial set of default recommendations associated with each main event_ID (e.g. message sent), perhaps part of a Sub_ID. But this would all need to be done in a way in which the control engine can work without knowing all the details behind the application.  It would ideally work solely based upon the messages it receives from the other components by deriving patterns.  Yes, hardly KISS but I guess that would be unavoidable with this sort of undertaking. There are lots of complications in actual implementation which would need to be addressed. It could be fun to attempt to build a proof of concept though. smile

</long>

Offline

#25 2011-03-06 23:19:31

davidm
Member
Registered: 2009-04-25
Posts: 371

Re: In search of a truly innovative Desktop

tlvb wrote:

That is fine if you have only a few tasks to be done, but if switching tasks often or having many different tasks I think that one would start getting
irritated by the buttons moving around/dissappearing/appearing.
Eh, seems I only read the first sentence, ...detecting those 'pro' usage patterns would still be difficult, on par or worse than natural language,
these action having no predefined grammar that the ai could build on...

Well as an example when I wrote the previous reply pidgin popped up a message asking me to write a deny authorization message in response to a spammer's request to get authorization on ICQ. This is due to my use of the bot-sentry plugin to automatically deny the request and generate a challenge/response.  There is apparently a limitation within pidgin that forces this box to pop up and it steals the focus each time.  One way to fix this would be to fix Podgin...yes. but please indulge me here. As an alternate couldn't such a system in theory realize that:

If both:

1) This particular message (associated Auth_Request) is received
2) The computer is not "idle"

then in 998/1000 of the last instances with no other pattern distinctions recognized I have done nothing but click OK immediately and so decide to not give the Pidgin focus in response to this event and instead (perhaps optionally) automatically initiate the normal action? Then it could for instance show a more unobtrusive message somewhere or allow me to stop the given action (feedback) from reoccurring with a single click.  The general idea being to add some central  intelligence to the computer besides that which exists between our ears.

It would be a challenge to design the proper algorithms to drive it, however in some cases such as this it is pretty clear cut.  Of course there would have to be some feedback mechanisms built in.  We might have a central log of actions taken (in addition to perhaps a brief notification shown depending on uncertainty) where a user could go back and tell the control module that it is "incorrect" (Say I suddenly decided to uninstall the bot-sentry plugin and wish to actually see these messages). The control engine would then have to work with that feedback in order to learn what it is doing wrong over time.  It wouldn't be perfect but at least helpful most of the time.

Last edited by davidm (2011-03-06 23:20:40)

Offline

Board footer

Powered by FluxBB