Page 3 of 4 FirstFirst 1234 LastLast
Results 21 to 30 of 39

Thread: call for more open collaboration

  1. #21
    licensed to kill - FKA kev000
    Auto Apps:loading...
    tripzero's Avatar
    Join Date
    Aug 2006
    Location
    16.40618, 120.61106
    Posts
    2,560
    I've got a skeleton page up on the wiki: nGhost Starter's Kit. I'll add more later...
    Former author of LinuxICE, nghost, nobdy.
    Current author of Automotive Message Broker (AMB).
    Works on Tizen IVI. Does not represent anyone or anything but himself.

  2. #22
    Low Bitrate
    Join Date
    Mar 2000
    Location
    Northville, MI, USA
    Posts
    65
    In relation to the original point... I think cooperation on a project like this is great. It sounds like you have a group of people already helping with nGhost. The difficulty is convincing people who already have significant code already developed (SW2, etc) to stop developing what they are working on, and focus on a new project that probably doesn't do exactly what they want, or exactly how they would like it to work.

    In my case, my front end app is dominated by custom software used to control several vehicle functions via Phidgets. For example... the tire pressure is adjusted by the computer... the shock dampening is adjusted, the rear steering angle is adjusted, various vehicle temps are monitored, wheel slip is monitored... etc etc. I also have Mplayer for video/audio.. and a gps receiver that shows elevation/heading/speed.

    Trying to get all that functionality into someone else's app would probably be an equal amount of work as me writing it myself.

    Alot of people are making suggestions about that distro or how the frontend should be written (which is good).. but I think some goals should be laid out.

    First... It should be compatible with different hardware platforms.
    Second... it should be easily expanded
    Third... it should be easily maintained/updated.

    Now I'll throw in my 2 cents about a direction this might take. The frontend should be distro agnostic. At the end of the day, all I should need is a dependancy list.... because in 2 years from now there is going to be a new video card... new chipset... new network card.. etc. And if the custom distro thats developed specifically for the front end doesn't work... you will loose users.
    I also think it should run on X.. because there are so many other graphical apps that run on X which could be integrated with the frontend.. and why bother writing a graphics library when so many already exist?
    2 years from now.. when someone wants to add something to the app.. and they have been writing graphical apps for linux already... they are probably going to be able to do that faster if its written in GTK or QT.

    I think people are primarily using windows in their carpc because 1) There are some good frontends.. and 2) It works with their hardware.

    Sorry... long winded this morning.

    -Chris

  3. #23
    Variable Bitrate intuitionsys's Avatar
    Join Date
    Jul 2005
    Location
    Northern Virginia
    Posts
    293
    Quote Originally Posted by awsjr View Post
    I agree with you. Linux in general needs better direction. The Linux CarPC software is a prime example. I am not a developer but I am a designer with a little experience in interface design. So while I can’t contribute code I can make things pretty and test software. Maybe those of us interested should formalize things…I don’t know how but it seams like a good idea.
    Okay I'm back from a long weekend of no coding whatsoever I think what the developers absolutely need are people who can "pretty up" their work. There are some fantastic skins out there that I'd love to add to the Silverwolf2 library but I just don't have time to draw skins and code and do my regular job, etc., etc.

    I've already started work on completing the rest of the Silverwolf2 modules and it would be nice to release a new skin with the next alpha or beta release (hint hint...)
    Silverwolf 2 is dead.
    Silverwolf 3...?

  4. #24
    Low Bitrate
    Join Date
    Jan 2004
    Location
    Italy
    Posts
    68
    Quote Originally Posted by strandtc View Post
    I also think it should run on X.. because there are so many other graphical apps that run on X which could be integrated with the frontend.. and why bother writing a graphics library when so many already exist?
    2 years from now.. when someone wants to add something to the app.. and they have been writing graphical apps for linux already... they are probably going to be able to do that faster if its written in GTK or QT.
    Let me disagree with this point.
    My point of view starts from thinkin' of a carpc not like a desktop computer, but a system easy to control by touchscreen or other non-standard inpud device.
    This way, managing to control an external application via menu entries should be really hard and extremely dangerous if done while driving.
    So I prefer an application that integrates well in the frontend, without the need of change window/virtual desktop or minimize and maximize an app.
    I think a good linux application (not a frontend) is written dividing the core from the interface, so one could use commandline or any type of frontend...
    Just my 2 cents!

  5. #25
    Variable Bitrate intuitionsys's Avatar
    Join Date
    Jul 2005
    Location
    Northern Virginia
    Posts
    293
    Quote Originally Posted by strandtc View Post
    First... It should be compatible with different hardware platforms.
    Absolutely. Silverwolf up to v1.3 was so hardware specific it didn't get much attention.
    Quote Originally Posted by strandtc View Post
    Second... it should be easily expanded
    Absolutely again; mind you I think most frontends are expandable - to what degree and how easy that is is another question altogether.
    Quote Originally Posted by strandtc View Post
    Third... it should be easily maintained/updated.
    Silverwolf2 will be able to automate updating itself (next release), its skins, libraries, etc. but updating anything else is outside its scope.
    Quote Originally Posted by strandtc View Post
    Now I'll throw in my 2 cents about a direction this might take. The frontend should be distro agnostic. At the end of the day, all I should need is a dependancy list.... because in 2 years from now there is going to be a new video card... new chipset... new network card.. etc. And if the custom distro thats developed specifically for the front end doesn't work... you will loose users.
    I also think it should run on X.. because there are so many other graphical apps that run on X which could be integrated with the frontend.. and why bother writing a graphics library when so many already exist?
    2 years from now.. when someone wants to add something to the app.. and they have been writing graphical apps for linux already... they are probably going to be able to do that faster if its written in GTK or QT.

    I think people are primarily using windows in their carpc because 1) There are some good frontends.. and 2) It works with their hardware.
    No arguement! I would also add that alot of people are using Windows on their car PC because that's what they know. Here is where car-specific distros like linuxICE or slax might ease the pain of transition.
    Silverwolf 2 is dead.
    Silverwolf 3...?

  6. #26
    Constant Bitrate reece146's Avatar
    Join Date
    Sep 2004
    Location
    ON, Canada
    Posts
    126
    X Window can be set up to look like anything you want.

    Does this look like X?

    http://moose.ca/content/projects/davis/DAVIS_v03a.png


    If you really want to dump X I suggest doing something that is compatible with X or SDL or fbdev or whatever. Development of X apps via the network connection is invaluable - just export your display and you can work on the box anywhere. Also, when troubleshooting hook up the RJ45 and work from your desk instead of sitting in the driver's seat with some excuse of a keyboard.

    When it's time to run on the dedicated hardware move your interface over and you are off to the races.


    Collaboration is a tough nut to crack. If the "outsiders" don't have interest then they aren't going to pony up.

  7. #27
    Low Bitrate
    Join Date
    Jan 2004
    Location
    Italy
    Posts
    68
    Quote Originally Posted by reece146 View Post
    X Window can be set up to look like anything you want.

    Does this look like X?

    http://moose.ca/content/projects/davis/DAVIS_v03a.png
    Hey, very nice!
    I forgot the high flexibility of FVWM! But I remeber it's a bit hard to customize...
    After seeing the blog I wonder why to create a dedicated frontend, if we can just use a wm and existing applications! :P

    But, let me put a log in the fire : using external application you have to theme the frontend, the window manager and the library toolkit, instead of create a single frontend theme... less code or easy themeing? this is the question....

    bye
    SanzoGhenzo

  8. #28
    Constant Bitrate reece146's Avatar
    Join Date
    Sep 2004
    Location
    ON, Canada
    Posts
    126
    IMO creating a "proper" front end is the correct methodology to take. In my case I don't have the necessary skill set to write something from scratch in C/C++ and I don't have the time (or more truthfully, interest) to learn C/C++ to make an application of that scale.

    The biggest problem with taking the window manager approach is that it can be difficult to find a suite of applications that provide the functionality desired and that can be themed to suit your purposes.

    With my example above there are four things being themed:

    1.) Application (gqmpeg in this instance)
    2.) Window Manager (fvwm above)
    3.) GTK (all my chosen apps use GTK and/or Xlib directly)
    4.) X itself (.Xdefaults file)

    It gets tricky getting the apps to have the same look and feel also because some apps do not like to play nice with window manager hints and also some applications were written with a very rough idea of what a dialog box should look like, how to present items to the users, etc.

  9. #29
    licensed to kill - FKA kev000
    Auto Apps:loading...
    tripzero's Avatar
    Join Date
    Aug 2006
    Location
    16.40618, 120.61106
    Posts
    2,560
    In my case, my front end app is dominated by custom software used to control several vehicle functions via Phidgets. For example... the tire pressure is adjusted by the computer... the shock dampening is adjusted, the rear steering angle is adjusted, various vehicle temps are monitored, wheel slip is monitored... etc etc. I also have Mplayer for video/audio.. and a gps receiver that shows elevation/heading/speed.

    Trying to get all that functionality into someone else's app would probably be an equal amount of work as me writing it myself.
    I don't know how comfortable you would feel about opening up your code, but in my mind at least people will be able to look at your code and doing the work of combining functionality.

    2 years from now.. when someone wants to add something to the app.. and they have been writing graphical apps for linux already... they are probably going to be able to do that faster if its written in GTK or QT.
    You are probably right that by using QT or GTK makes it easier for others to contribute. I am doing 98% of the SDL coding in nGhost because it isn't an easy known "toolkit". However the flexibility SDL provides makes it all worth while. SDL runs natively in GNOME and KDE, with X or without X, and has great hardware driver support. It all comes down to preference really.
    Former author of LinuxICE, nghost, nobdy.
    Current author of Automotive Message Broker (AMB).
    Works on Tizen IVI. Does not represent anyone or anything but himself.

  10. #30
    Low Bitrate
    Join Date
    Mar 2004
    Posts
    89
    First... It should be compatible with different hardware platforms.
    Second... it should be easily expanded
    Third... it should be easily maintained/updated.
    I could not agree with you more, strandtc. In fact, I had tried steering some people in this direction a couple of years ago. They of course opted for the tightly integrated app, and today it is another piece of vaporware. Give me a few hours with a linux desktop and I can customize it to make it as touchscreen freindly as any of these simply by choosing the right combination of window-manager, toolbars, docking applications, etc...

    Look at OPIE and GPE for handhelds as an example. They look nothing like KDE and GTK, yet they are able to run many of the same apps with only a few modifications to make them suitable for PDA use.

    Going this route does not mean that your system has to look like a desktop either. Using X11 and a window manager does not mean that you have to have a "start menu" in the bottom left hand corner. If you want your application menu to be a full screen app that looks exactly like the one for nghost, you can do this.

    OpenMoko is a good example of something that runs on X11, yet looks nothing like a desktop. http://wiki.openmoko.org/wiki/Main_Page

Page 3 of 4 FirstFirst 1234 LastLast

Similar Threads

  1. Replies: 2
    Last Post: 08-09-2004, 02:49 PM
  2. Open Source project
    By NEO in forum Software & Software Development
    Replies: 86
    Last Post: 08-05-2004, 03:12 PM
  3. Questions for developers
    By SuperG35 in forum ME Archive
    Replies: 28
    Last Post: 08-26-2003, 10:12 AM
  4. Is ME open source?
    By SuperG35 in forum ME Archive
    Replies: 2
    Last Post: 08-22-2003, 02:42 PM
  5. still open source?
    By SilverJester in forum ME Archive
    Replies: 3
    Last Post: 03-03-2003, 09:52 PM

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •