Hildon Input Method Finally Opened
I knew having the GNOME wiki Recent Changes feed in Google Reader was a good idea... I just noticed a number of pages being created under the HildonInputMethod tree. Hooray, the Hildon Input Method is finally open source!
NP: We Made It For You, The Boats
Posted by MJ Ray at Fri Sep 7 19:05:38 2007:
So what is the Hildon Input Method? Your link links to mostly videos and I STFW'd, but found no FM to R.
Posted by Ross at Fri Sep 7 19:16:33 2007:
HIM is the magic which lets you use a virtual keyboard or handwriting recognition on the Nokia 770 and N800. It's a pretty complicated layer of magic. :)
Posted by Rusty at Sun Sep 9 00:18:33 2007:
The video needs work by the way, text flow is entirely too fast for reading during playback, and entirely to short for someone to frame forward through if that were even an option.
In any case I was able to glean that the input method is that used on Nokia web tablets to take user text input and hand it off to the application or gui. Also that their intent is to open source it, with proposed hooks into GTK and X. If that's the message so far, with the understanding that parts of it are being open sourced already, a text writeup would be a better use of the wiki.
From a video, I would like to see an example of the user experience, and what the method does to convert that experience into the input the platform is expecting. It might make sense to have a tutorial showing how to make a plug-in for a new keyboard layout, or a filter to take some form of keyboard input to be passed on to an application, say a mnemonic converter to convert 1800-flowers to 1800-3569377 for Skype-out or another phone application. (technically this is as complicated as using a pipeline of 'tr [A-Z] [a-z] | tr [abcdefghijklmnopqrstuvwxyz] [22233344455566677778889999]'
Other interesting plug-ins might be speech to text conversion. Probably a lot more work as the method appears to be screen focused right now. But the simple experience might be four on-screen buttons (or eight, four across the bottom, four across the top, the top row in reverse order to the bottom, eg:
---------------
[4] [3] [2] [1]
[1] [2] [3] [4]
---------------
)
With button 1 being restart, 2 be record/stop, 3 be playback recorded text, 4 be enter. 3 is really send recording through a speech to text app, then send output of that as input to text to speech and playback that through the audio system. where 4 is normally send the result to the application, though there may be some variation to handle things like punctuation, mouse movement, etc. Button 1 throws anything in the buffer away, and would most likely be used if 3 comes up with garbage.
The 8 button layout is so that a user doesn't have to know which is the op of the screen, and which is the bottom. Potentially helpful for a blind user.
I doubt that either of these is beyond the capabilities of a god programmer to hook into the Hildon Method, though the latter may involve tools that your average developer may not work with often.
Posted by MJ Ray at Sun Nov 4 12:43:38 2007:
In any case I was able to glean that the input method is that used on Nokia web tablets to take user text input and hand it off to the application or gui. Also that their intent is to open source it, with proposed hooks into GTK and X. If that's the message so far, with the understanding that parts of it are being open sourced already, a text writeup would be a better use of the wiki.
From a video, I would like to see an example of the user experience, and what the method does to convert that experience into the input the platform is expecting. It might make sense to have a tutorial showing how to make a plug-in for a new keyboard layout, or a filter to take some form of keyboard input to be passed on to an application, say a mnemonic converter to convert 1800-flowers to 1800-3569377 for Skype-out or another phone application. (technically this is as complicated as using a pipeline of 'tr [A-Z] [a-z] | tr [abcdefghijklmnopqrstuvwxyz] [22233344455566677778889999]'
Other interesting plug-ins might be speech to text conversion. Probably a lot more work as the method appears to be screen focused right now. But the simple experience might be four on-screen buttons (or eight, four across the bottom, four across the top, the top row in reverse order to the bottom, eg:
---------------
[4] [3] [2] [1]
[1] [2] [3] [4]
---------------
)
With button 1 being restart, 2 be record/stop, 3 be playback recorded text, 4 be enter. 3 is really send recording through a speech to text app, then send output of that as input to text to speech and playback that through the audio system. where 4 is normally send the result to the application, though there may be some variation to handle things like punctuation, mouse movement, etc. Button 1 throws anything in the buffer away, and would most likely be used if 3 comes up with garbage.
The 8 button layout is so that a user doesn't have to know which is the op of the screen, and which is the bottom. Potentially helpful for a blind user.
I doubt that either of these is beyond the capabilities of a god programmer to hook into the Hildon Method, though the latter may involve tools that your average developer may not work with often.
