View source for Wish List
From Openmoko
You do not have permission to edit this page, for the following reasons:
You can view and copy the source of this page:
Templates used on this page:
Return to Wish List.
You do not have permission to edit this page, for the following reasons:
You can view and copy the source of this page:
Templates used on this page:
Return to Wish List.
This article is a place to collect various thoughts about the future of the OpenMoko software platform.
Our goal should be a completely painless setup for somebody wanting to develop using OpenMoko
No extra configuration required.
People like to see plugins for
Glade code generation is deprecated, so we don't want to use it. The Gtk+ powers told me that the plan is to have gtk 2.12 (out early 2007) with support for GtkBuilder, a libglade derivative which breaks a bit the XML definition in order to support all the new widgets and properties; as soon as it's in the other ui builders will add support for this format. See also the relevant bug entry
Infrastructure for developers with
In the future there could be complete, unofficial "product images" that are created by the community, for example maybe one that incorporates only free software (in the GNU or OSI sense). Or images build with a particular niche market in mind -- a student for example.
There are many useful options that now can be used to full capacity:
Something like HexInput would be very interesting to try. Especially versions of the QUONG layout. Different variations for different languages will be needed.
Another possibility would be to use Dasher along with the touchpad interface.
I'd like to see U-Boot being enhanced to support a boot menu that allows to boot a complete image from the MMC. This way we have a simple way for part-time developers (aka people actually using their phone or people not having a separate phone for development) to experiment with new OpenMoko releases or custom images.
Please see the games page.
It would be really great to be able to read :
in both landscape and portrait
It would be really neat to be able to print over either bluetooth or USB. I can imagine wanting to print:
Cups contains a bluetooth printing backend, so (in theory) once you have your data in postscript format, you could hand it to cups and it'll do the rest. In practice, it depends on
NOTE: GTK+'s printing support seems to be very immature in 2.6 (which we need to use for some time). Gtk+ 2.10 contains much better printing support -- once we can use this, it should be more easy. |
There's always the possibility to render postscript ourselves, but this is not a piece of cake -- in general, printing is much harder than one would imagine.
Further details:
The Access group is probably coming out with their Linux platform any time soon. One of the components is a PalmOS emulator which I'd like to see working on OpenMoko as well. There are literally millions of PalmOS apps.
After reading something about too much features in mobile phones I thought about the following. It would be nice to have the possibility to choose from several profiles when starting the phone the first time. Every user profile is defined for special kinds of user behaviour. Let me define some example profiles:
etc. In addition the user should be able to create his own profiles. The advantage is that the user gets not frustrated by thousands of menus. What do you think about such a feature?
Allow Bluetooth to automatically turn off after loosing connectivity and to automatically turn back on based upon GPS location.
A Bluetooth device is configured for automatic reacquisition based on the following profiles:
Each target device is configured as follows:
Examples devices include: computers
The location and range of the target device is determined via training. Periodically, the current GPS coordinates and Bluetooth signal strength are logged. Additionally, connectivity loss events are logged. An algorithm uses these logs to determine the device location and range.
Connection attempts are made when in a configurable proximity to the device. The first attempt when entering the proximity and further attempts at a configurable interval.
Example devices include: automobiles
Mobile devices are configured to have two types of locations:
A car is mobile, ideally, when you leave your car, the phone should note the car's location when connectivity is lost and then attempt to reacquire the car when you return to the location of the car.
As mobile devices may have multiple users, it is not sufficient to always use the last known location. In this case, the device may additionally have multiple homes. For example, a car might have as its homes: home garage and work parking lot.
Using the microphone to detect ambient noise the ringtone volume could be adjusted automatically.
Shut up a ringing phone, without rejecting the call.
Another alternative might be to use microphone to recognize when the user gives an audible "Shhh!" command. This could prove difficult to determine with the simultaneous ringing, and possible in-pocket shuffling noises.
Button to temporarily disable microphone while talking.
With a "silent mode" timeout, there is no need to turn the ringer back on if you previously know how long will be the film/meeting/...
I arrives to home and there is some "@home" things in the to-do list, the Context based to-do list reminds me that.
Like the one laptop per child (OLPC) interface, keep a number in the status bar that represents a count of other openmoko or compatible bluetooth devices in the area. Allow for the spontaneous initiation of a chatroom or multiplayer game or file trading with any moko in the area.
Once there is good TCP/IP connectivity on this phone, integration with corportate email/calendar/todo/etc servers would be a big advantage... near-real-time automatic email downloads and automatic bi-directional syncing are productivity boosters that you have to experience to appreciate. It turns your phone from a 'nice gadget to fiddle with' to a natural-feeling extension of your day-to-day life.
An application that would allow the user to define their own vibration patterns, and possibly link them to audio files. Recording would be done in real time initiated with a "Record" button, optionally playing the associated sound file in sync with recording). While recording, the user would press and hold a button to define the timing and duration of vibration. The user would press "Stop" when finished. Vibration patterns would have the option of being looped(would terminate at some global ringtone length maximum).
One simple suggested vibration file format would be a sort of run-length encoding: First byte defines the length of a "time-slice" in milliseconds, which would determine the overall tempo(actually the inverse of tempo). The next byte would define the number of time-slices to leave the vibration on, and then another byte for how long to pause after. Continue alternating these on/off bytes until the entire pattern is defined.
Let OpenMoko devices connect to one another via bluetooth and hold a conversation. Depending on the effective range between two devices, this might be a very silly thing to do, so maybe it's only purpose is as a fun toy application.
Provide a method to use the touchscreen as input device for a nearby desktop machine. Could connect over USB or bluetooth.
ANARM would be an application for handling all event-based audible notifications from an OpenMoko device.
Have a deeper connection to the car than just handsfree speakerphone. For instance a transciever with challenge/response systems to open, possibly even start the car. Possibly go as far as OBD connection to monitor car status on screen/log for later.
Python bindings seem to be a commonly requested feature.
User:Mickey says, "They are kind of usable on the Nokia 770, but it's at the lower end of being bearable. We should keep this in mind -- Gtk+ already comes with Python Bindings, so we "just" would need to wrap libmoko*. I would prefer to leave this to the community do though, since it doesn't make sense to start wrapping the API until we have a stable API -- and I can imagine it will take us a couple of months after going open until we can start with stabilizing the libmoko API."
There is a whole skilled C++ community coming from the Qtopia and Opie projects. If we would consider basing OpenMoko C++ Bindings on Gtkmm, then we could drag these guys in.
The Trolltech folks have a great widget library. I'd like to interface OpenMoko with Qt4, so that we can write Qt4 applications for the phone which don't look alienated.
The Maemo folks have created a successful standard for Webpad applications. I'd like to have a set of MaemoMoko and MokoMaemo wrapper classes that allow me add support for running OpenMoko applications on Maemo and vice versa. Perhaps we can get help from the Nokia OSS folks for that.
wxWidgets is a cross-platform application framework that's very popular (I'd say, #3 after Qt and Gtk+). On Linux, wxWidgets uses Gtk+ to implement the widgets. It shouldn't be hard to add support for the additional OpenMoko classes to wxWidgets hence supporting the native OpenMoko look and feel for wxWidgets applications.
wxWidgets team wants OpenMoko classes too and we (wxWidgets) plan to include this project as one of our ideas for GSoC 2007
SDL is the game developer library. There are tons of SDL games out there. We should add OpenMoko support into SDL.