Re: New Rasterman Image...
2008/10/2 The Rasterman Carsten Haitzler [EMAIL PROTECTED]: On Thu, 02 Oct 2008 00:59:20 -0700 Steve Mosher [EMAIL PROTECTED] babbled: aha! a decent frequency corpus (a few thousand words). i'll merge this with the default us dict (and remove the small one as now that's useless). If you are still looking for a word corpus I found this the other day: http://en.wiktionary.org/wiki/Wiktionary:Frequency_lists Looks like they come from a number of sources including Project Gutenberg. Robert ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Hi there! I've got one question... Since this image is using a much richer theme things like kinetic scrolling of lists get really slow. I tried to lower the device's resolution and it was much faster! The only problem: There is something wrong with the colors when using 'xrandr -s 240x320'... Is there any way to fix that? -- Sebastian ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Marco Trevisan (Treviño) wrote: In the enlightenment archives there's a new Rasterman image: New updates from Rasterman (ehy, why don't you add RSS feeds to your site?! :P): http://www.rasterman.com/. To underline a new image [1] and some scripts to get the same by compiling [2]. In his files folder there's also a nice mockup for the GTA03 hw [3]! [1] http://download.enlightenment.org/misc/Illume/Freerunner/2008-10-12/ [2] http://trac.enlightenment.org/e/browser/trunk/TMP/oe [3] http://www.rasterman.com/files/gta03-idea-1.png -- Treviño's World - Life and Linux http://www.3v1n0.net/ ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Wed, Oct 15, 2008 at 09:12:44PM +0200, Marco Trevisan (Treviño) wrote: Marco Trevisan (Treviño) wrote: In the enlightenment archives there's a new Rasterman image: New updates from Rasterman (ehy, why don't you add RSS feeds to your site?! :P): http://www.rasterman.com/. To underline a new image [1] and some scripts to get the same by compiling [2]. In his files folder there's also a nice mockup for the GTA03 hw [3]! [1] http://download.enlightenment.org/misc/Illume/Freerunner/2008-10-12/ [2] http://trac.enlightenment.org/e/browser/trunk/TMP/oe [3] http://www.rasterman.com/files/gta03-idea-1.png Have you tried it already? Does it work? (ie, does it make/receive calls? can you hear and be listented to?) Rui -- Frink! Today is Pungenday, the 69th day of Bureaucracy in the YOLD 3174 + No matter how much you do, you never do enough -- unknown + Whatever you do will be insignificant, | but it is very important that you do it -- Gandhi + So let's do it...? ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
El mié, 15-10-2008 a las 20:31 +0100, Rui Miguel Silva Seabra escribió: On Wed, Oct 15, 2008 at 09:12:44PM +0200, Marco Trevisan (Treviño) wrote: Marco Trevisan (Treviño) wrote: In the enlightenment archives there's a new Rasterman image: New updates from Rasterman (ehy, why don't you add RSS feeds to your site?! :P): http://www.rasterman.com/. To underline a new image [1] and some scripts to get the same by compiling [2]. In his files folder there's also a nice mockup for the GTA03 hw [3]! [1] http://download.enlightenment.org/misc/Illume/Freerunner/2008-10-12/ [2] http://trac.enlightenment.org/e/browser/trunk/TMP/oe [3] http://www.rasterman.com/files/gta03-idea-1.png Have you tried it already? I have tried, it has an tree apps an alarm, a screenshotter and a terminal. looks very pretty and lovely, I'm still guessin why OM has no agree with raster on what the phone software has to look like Does it work? (ie, does it make/receive calls? can you hear and be listented to?) You have to install somethings to arrive to do this, is more a proof of concept of what a Illume can look like in the Neo than a intention of have a working phone. I dream a day that I can aply this directly with an opkg install to a working official OM distro. BTW, regarding http://www.rasterman.com/files/gta03-idea-1.png I like a lot all the design less one detail, raster let's start again a qvga vs vga flame? :) Rui ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Wed, 15 Oct 2008 21:49:34 +0200 David Samblas [EMAIL PROTECTED] babbled: El mié, 15-10-2008 a las 20:31 +0100, Rui Miguel Silva Seabra escribió: On Wed, Oct 15, 2008 at 09:12:44PM +0200, Marco Trevisan (Treviño) wrote: Marco Trevisan (Treviño) wrote: In the enlightenment archives there's a new Rasterman image: New updates from Rasterman (ehy, why don't you add RSS feeds to your site?! :P): http://www.rasterman.com/. To underline a new image [1] and some scripts to get the same by compiling [2]. In his files folder there's also a nice mockup for the GTA03 hw [3]! [1] http://download.enlightenment.org/misc/Illume/Freerunner/2008-10-12/ [2] http://trac.enlightenment.org/e/browser/trunk/TMP/oe [3] http://www.rasterman.com/files/gta03-idea-1.png Have you tried it already? I have tried, it has an tree apps an alarm, a screenshotter and a terminal. looks very pretty and lovely, I'm still guessin why OM has no agree with raster on what the phone software has to look like i just have zero phone software on it. no agreement one way or another here - i don't work for OM, so no need to agree or disagree. i removed all of frameworkd etc. because it was causing major performance problems. it alos means the image got a fair bit smaller. it boots baster too as less hammering the device on startup. as such frameworkd has yet to actually handle an incoming or outgoing call for me on my sim card so i haven't focused on this. right now i have other things to focus on (widget set, ui integration and adaption to different resolutions etc.). zhone and frameworkd just get in my way. Does it work? (ie, does it make/receive calls? can you hear and be listented to?) You have to install somethings to arrive to do this, is more a proof of concept of what a Illume can look like in the Neo than a intention of have a working phone. I dream a day that I can aply this directly with an opkg install to a working official OM distro. BTW, regarding http://www.rasterman.com/files/gta03-idea-1.png I like a lot all the design less one detail, raster let's start again a qvga vs vga flame? :) that's an old design i did. before om had any designs of its own. it's just useless junk right now. Rui ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)[EMAIL PROTECTED] ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Wed, Oct 15, 2008 at 09:49:34PM +0200, David Samblas wrote: El mié, 15-10-2008 a las 20:31 +0100, Rui Miguel Silva Seabra escribió: On Wed, Oct 15, 2008 at 09:12:44PM +0200, Marco Trevisan (Treviño) wrote: Marco Trevisan (Treviño) wrote: In the enlightenment archives there's a new Rasterman image: New updates from Rasterman (ehy, why don't you add RSS feeds to your site?! :P): http://www.rasterman.com/. To underline a new image [1] and some scripts to get the same by compiling [2]. In his files folder there's also a nice mockup for the GTA03 hw [3]! [1] http://download.enlightenment.org/misc/Illume/Freerunner/2008-10-12/ [2] http://trac.enlightenment.org/e/browser/trunk/TMP/oe [3] http://www.rasterman.com/files/gta03-idea-1.png Have you tried it already? I have tried, it has an tree apps an alarm, a screenshotter and a terminal. looks very pretty and lovely, I'm still guessin why OM has no agree with raster on what the phone software has to look like Does it work? (ie, does it make/receive calls? can you hear and be listented to?) You have to install somethings to arrive to do this, is more a proof of concept of what a Illume can look like in the Neo than a intention of have a working phone. I dream a day that I can aply this directly with an opkg install to a working official OM distro. I know... I don't mind installing stuff to have phone work, just some hints for the bare bones minimum definition of work :) Right now I receive and make calls, but it's totally silent. I tried fiddling with alsamixer, but to no avail :| Tomorrow is a work day, I'll need to reinstall Om2008.9 soon if I can't hear/speak :| Rui -- You are what you see. Today is Pungenday, the 69th day of Bureaucracy in the YOLD 3174 + No matter how much you do, you never do enough -- unknown + Whatever you do will be insignificant, | but it is very important that you do it -- Gandhi + So let's do it...? ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Wed, Oct 15, 2008 at 11:10:59PM +0100, Rui Miguel Silva Seabra wrote: Tomorrow is a work day, I'll need to reinstall Om2008.9 soon if I can't hear/speak :| Didn't last till tomorrow, I just had a most dreadful White Screen of Death *sigh* Rui -- Or not. Today is Pungenday, the 69th day of Bureaucracy in the YOLD 3174 + No matter how much you do, you never do enough -- unknown + Whatever you do will be insignificant, | but it is very important that you do it -- Gandhi + So let's do it...? ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
No sound is a FSO problem I believe. It's fixed, but I think the fix is only in sources. We'll have to wait for updated frameworkd and zhone on downloads.freesmartphone.org On Thu, Oct 16, 2008 at 12:26 AM, Rui Miguel Silva Seabra [EMAIL PROTECTED]wrote: On Wed, Oct 15, 2008 at 11:10:59PM +0100, Rui Miguel Silva Seabra wrote: Tomorrow is a work day, I'll need to reinstall Om2008.9 soon if I can't hear/speak :| Didn't last till tomorrow, I just had a most dreadful White Screen of Death *sigh* Rui -- Or not. Today is Pungenday, the 69th day of Bureaucracy in the YOLD 3174 + No matter how much you do, you never do enough -- unknown + Whatever you do will be insignificant, | but it is very important that you do it -- Gandhi + So let's do it...? ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Nice, thanks. I guess I'll try again later :) On Thu, Oct 16, 2008 at 12:45:42AM +0200, Leonti Bielski wrote: No sound is a FSO problem I believe. It's fixed, but I think the fix is only in sources. We'll have to wait for updated frameworkd and zhone on downloads.freesmartphone.org On Thu, Oct 16, 2008 at 12:26 AM, Rui Miguel Silva Seabra [EMAIL PROTECTED]wrote: On Wed, Oct 15, 2008 at 11:10:59PM +0100, Rui Miguel Silva Seabra wrote: Tomorrow is a work day, I'll need to reinstall Om2008.9 soon if I can't hear/speak :| Didn't last till tomorrow, I just had a most dreadful White Screen of Death *sigh* -- Kallisti! Today is Pungenday, the 69th day of Bureaucracy in the YOLD 3174 + No matter how much you do, you never do enough -- unknown + Whatever you do will be insignificant, | but it is very important that you do it -- Gandhi + So let's do it...? ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Carsten Haitzler (The Rasterman) ha scritto: On Thu, 02 Oct 2008 20:52:00 +0200 Marco Trevisan (Treviño) [EMAIL PROTECTED] babbled: So this is a little utility I wrote [1] to check the frequency of each word and writing back a new dictionary with frequency data. To run it you need php-cli (I guess v5 or above), set the given options, do php words-popularity.php and wait the work to be finished! :P It could be a long work, but it should give good results. yes. it would. who wants to run it? :) I've done it for about 42 words. Divinding the work in 5 shells went quite fine and took few hours, but now Google blocked it. I didn't know that I wasn't allowed to do it :/. I figure we should change our source :P. nb. i checked illume's kbd code - it does have issues with utf8 keysequences in sorted dicts. if you have any it'll fail to keep looking for more words so you need to remove anything utf8 from your dict :( yes - i know. bad. i need to address this. and the change in dict format i am sure 1. makes this now simple, 2. compresses the dict, 3. speeds it up, 4. solves this problem. :) but i just need to do it - no time right now :( Yes, I do agree with this. Using a better compressed format would increase the performances allowing to add more words. I think that the qtopia dawg format is a good example for this. I just hope you'll find some time for it soon :P. -- Treviño's World - Life and Linux http://www.3v1n0.net/ ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Cédric Berger wrote: On Tue, Oct 7, 2008 at 08:10, The Rasterman Carsten Haitzler [EMAIL PROTECTED] wrote: Ummm. Nice try, but see the results below. I ran this remotely on my workstation at work, since it's right on a 20mb fiber... Hopefully I'll be allowed to use Google again by the time I get in to work tomorrow... ;) hmm. looks like we need a botnet to distribute the queries! :) each bot does 1000 of them... :) Or maybe it is possible to ask google to be allowed to do such a task ? Maybe, as reported in their license, it would be possible since I don't think that we really use their data (i.e. search results them selves) but only the number of the results they have for a single word, and this shouldn't be nothing of bad for them, considering also that the script ask only a result per page and that the search time is always really low. Anyone could contact the G? BTW we could also using another search engine as source... -- Treviño's World - Life and Linux http://www.3v1n0.net/ ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Joel Newkirk wrote: On Tue, 7 Oct 2008 10:52:13 +1100, Carsten Haitzler (The Rasterman) [EMAIL PROTECTED] wrote: On Thu, 02 Oct 2008 20:52:00 +0200 Marco Trevisan (Treviño) [EMAIL PROTECTED] babbled: So this is a little utility I wrote [1] to check the frequency of each word and writing back a new dictionary with frequency data. To run it you need php-cli (I guess v5 or above), set the given options, do php words-popularity.php and wait the work to be finished! :P It could be a long work, but it should give good results. yes. it would. who wants to run it? :) Ummm. Nice try, but see the results below. I ran this remotely on my workstation at work, since it's right on a 20mb fiber... Hopefully I'll be allowed to use Google again by the time I get in to work tomorrow... ;) j [1328/98568] [1329/98568] [1330/98568] Unknown Popularity of Aeneid's [1331/98568] Unknown Popularity of Aeolus [1332/98568] Unknown Popularity of Aeolus's [1333/98568] Unknown Popularity of aeon [1334/98568] Unknown Popularity of aeon's [1335/98568] Unknown Popularity of aeons [1336/98568] Unknown Popularity of aerate [1337/98568] Unknown Popularity of aerated [1338/98568] Unknown Popularity of aerates [1339/98568] Unknown Popularity of aerating [1340/98568] Unknown Popularity of aeration [1341/98568] Unknown Popularity of aeration's [1342/98568] Unknown Popularity of aerator [1343/98568] Unknown Popularity of aerator's [1344/98568] Unknown Popularity of aerators [1345/98568] After which browsing to www.google.com results in: Google Error As said above, after that I was able to collect informations for about 42 words, Google blocked it since bigG doesn't allow to use these batch searches (I didn't know :P). The only way to do it like I did is asking the permission to Google, I guess. -- Treviño's World - Life and Linux http://www.3v1n0.net/ ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Tue, 7 Oct 2008 10:52:13 +1100, Carsten Haitzler (The Rasterman) [EMAIL PROTECTED] wrote: On Thu, 02 Oct 2008 20:52:00 +0200 Marco Trevisan (Treviño) [EMAIL PROTECTED] babbled: So this is a little utility I wrote [1] to check the frequency of each word and writing back a new dictionary with frequency data. To run it you need php-cli (I guess v5 or above), set the given options, do php words-popularity.php and wait the work to be finished! :P It could be a long work, but it should give good results. yes. it would. who wants to run it? :) Ummm. Nice try, but see the results below. I ran this remotely on my workstation at work, since it's right on a 20mb fiber... Hopefully I'll be allowed to use Google again by the time I get in to work tomorrow... ;) j [1328/98568] [1329/98568] [1330/98568] Unknown Popularity of Aeneid's [1331/98568] Unknown Popularity of Aeolus [1332/98568] Unknown Popularity of Aeolus's [1333/98568] Unknown Popularity of aeon [1334/98568] Unknown Popularity of aeon's [1335/98568] Unknown Popularity of aeons [1336/98568] Unknown Popularity of aerate [1337/98568] Unknown Popularity of aerated [1338/98568] Unknown Popularity of aerates [1339/98568] Unknown Popularity of aerating [1340/98568] Unknown Popularity of aeration [1341/98568] Unknown Popularity of aeration's [1342/98568] Unknown Popularity of aerator [1343/98568] Unknown Popularity of aerator's [1344/98568] Unknown Popularity of aerators [1345/98568] After which browsing to www.google.com results in: Google Error We're sorry... ... but your query looks similar to automated requests from a computer virus or spyware application. To protect our users, we can't process your request right now. We'll restore your access as quickly as possible, so try again soon. In the meantime, if you suspect that your computer or network has been infected, you might want to run a virus checker or spyware remover to make sure that your systems are free of viruses and other spurious software. If you're continually receiving this error, you may be able to resolve the problem by deleting your Google cookie and revisiting Google. For browser-specific instructions, please consult your browser's online support center. If your entire network is affected, more information is available in the Google Web Search Help Center. We apologize for the inconvenience, and hope we'll see you again on Google. To continue searching, please type the characters you see below: If you can read this, you do not have images enabled. Please enable images in order to proceed. ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Tue, 7 Oct 2008 02:07:38 -0400 Joel Newkirk [EMAIL PROTECTED] babbled: On Tue, 7 Oct 2008 10:52:13 +1100, Carsten Haitzler (The Rasterman) [EMAIL PROTECTED] wrote: On Thu, 02 Oct 2008 20:52:00 +0200 Marco Trevisan (Treviño) [EMAIL PROTECTED] babbled: So this is a little utility I wrote [1] to check the frequency of each word and writing back a new dictionary with frequency data. To run it you need php-cli (I guess v5 or above), set the given options, do php words-popularity.php and wait the work to be finished! :P It could be a long work, but it should give good results. yes. it would. who wants to run it? :) Ummm. Nice try, but see the results below. I ran this remotely on my workstation at work, since it's right on a 20mb fiber... Hopefully I'll be allowed to use Google again by the time I get in to work tomorrow... ;) hmm. looks like we need a botnet to distribute the queries! :) each bot does 1000 of them... :) j [1328/98568] [1329/98568] [1330/98568] Unknown Popularity of Aeneid's [1331/98568] Unknown Popularity of Aeolus [1332/98568] Unknown Popularity of Aeolus's [1333/98568] Unknown Popularity of aeon [1334/98568] Unknown Popularity of aeon's [1335/98568] Unknown Popularity of aeons [1336/98568] Unknown Popularity of aerate [1337/98568] Unknown Popularity of aerated [1338/98568] Unknown Popularity of aerates [1339/98568] Unknown Popularity of aerating [1340/98568] Unknown Popularity of aeration [1341/98568] Unknown Popularity of aeration's [1342/98568] Unknown Popularity of aerator [1343/98568] Unknown Popularity of aerator's [1344/98568] Unknown Popularity of aerators [1345/98568] After which browsing to www.google.com results in: Google Error We're sorry... ... but your query looks similar to automated requests from a computer virus or spyware application. To protect our users, we can't process your request right now. We'll restore your access as quickly as possible, so try again soon. In the meantime, if you suspect that your computer or network has been infected, you might want to run a virus checker or spyware remover to make sure that your systems are free of viruses and other spurious software. If you're continually receiving this error, you may be able to resolve the problem by deleting your Google cookie and revisiting Google. For browser-specific instructions, please consult your browser's online support center. If your entire network is affected, more information is available in the Google Web Search Help Center. We apologize for the inconvenience, and hope we'll see you again on Google. To continue searching, please type the characters you see below: If you can read this, you do not have images enabled. Please enable images in order to proceed. ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)[EMAIL PROTECTED] ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Tue, Oct 7, 2008 at 2:19 PM, Eildert Groeneveld [EMAIL PROTECTED] wrote: Raster's keyboard is nice! So I tried his Image and installed fso as a phone stack. Further I like tangoGPS and also installed it together with fso-gpsd. However, tangoGPS does not show up on the desktop. Any ideas? applications.menu should have change. In /usr/share/applications, replace Application by Applications for the Categories section. ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Tue, 7 Oct 2008 14:19:58 +0200, Eildert Groeneveld [EMAIL PROTECTED] wrote: Raster's keyboard is nice! So I tried his Image and installed fso as a phone stack. Further I like tangoGPS and also installed it together with fso-gpsd. However, tangoGPS does not show up on the desktop. Any ideas? Further, on booting zhone comes up without the illume desktop. Only issueing /etc/init.d/xserver-nodm restart rectifies the situation. Did I screw up something somewhere? Lastly, what repositorities do you guys use for updates? Is there a way to use both fso and om? cheers Eildert Yeah, I ran into that with TangoGPS. Go into /usr/share/applications and edit tangogps.desktop - change the 'Categories' entry to 'Office' and the icon will appear. Zhone auto-starts, which is likely what we want since in this setup it doesn't register to GSM network and work as a phone without Zhone running. Drop-down the Top Shelf and you can select 'Home' to see the desktop, or select one of the three empty squares from the four on the top-center/right to switch to a different desktop that doesn't have Zhone on it... I'd be very leery of mixing feeds... I've got mine using http://downloads.freesmartphone.org/fso-unstable. (Although I still don't have any sound but the alarm, will be looking into that more later today) j ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: Making sound work (was Re: New Rasterman Image...)
El mar, 07-10-2008 a las 13:24 +0200, David Samblas escribió: El mar, 07-10-2008 a las 12:15 +0100, David Pottage escribió: On Mon, October 6, 2008 12:11 am, Richy wrote: I somehow managed to get it to work to some extend. The phone isn' ringin, just vibrating, but I can talk and hear during a call.. make sure you have gstreamer installed, also I added a lot of packages, with alsa, gst, openmoko-state, etc. Didn't work until reboot. (And due to the Neos booting-time I don't want to reboot after each package just to check...) Now, some PIM-software and this is near perfect :-) I am also having trouble with sound. I can hear the other party during calls, but apart from the The phone loudspeaker appears not to work. I don't get any ring tones, and I can't hear music from media plays such as pythm. I have installed pymixer, and have maxed out all the sliders but still nothing. I have also tried installing the snd_pcm_oss kernel module out of FSO, but it will not load due to undefined symbols. Any ideas. I am running the latest FDOM (Based on ASU 2008.9). have you managed to put the raster new illume in to FDOM??? :) :) I want to see it? an a step by step shell commands used will be even more cool :) ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
New Rasterman Image...
Raster's keyboard is nice! So I tried his Image and installed fso as a phone stack. Further I like tangoGPS and also installed it together with fso-gpsd. However, tangoGPS does not show up on the desktop. Any ideas? Further, on booting zhone comes up without the illume desktop. Only issueing /etc/init.d/xserver-nodm restart rectifies the situation. Did I screw up something somewhere? Lastly, what repositorities do you guys use for updates? Is there a way to use both fso and om? cheers Eildert ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Tue, Oct 7, 2008 at 08:10, The Rasterman Carsten Haitzler [EMAIL PROTECTED] wrote: Ummm. Nice try, but see the results below. I ran this remotely on my workstation at work, since it's right on a 20mb fiber... Hopefully I'll be allowed to use Google again by the time I get in to work tomorrow... ;) hmm. looks like we need a botnet to distribute the queries! :) each bot does 1000 of them... :) Or maybe it is possible to ask google to be allowed to do such a task ? ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: Making sound work (was Re: New Rasterman Image...)
El mar, 07-10-2008 a las 12:15 +0100, David Pottage escribió: On Mon, October 6, 2008 12:11 am, Richy wrote: I somehow managed to get it to work to some extend. The phone isn' ringin, just vibrating, but I can talk and hear during a call.. make sure you have gstreamer installed, also I added a lot of packages, with alsa, gst, openmoko-state, etc. Didn't work until reboot. (And due to the Neos booting-time I don't want to reboot after each package just to check...) Now, some PIM-software and this is near perfect :-) I am also having trouble with sound. I can hear the other party during calls, but apart from the The phone loudspeaker appears not to work. I don't get any ring tones, and I can't hear music from media plays such as pythm. I have installed pymixer, and have maxed out all the sliders but still nothing. I have also tried installing the snd_pcm_oss kernel module out of FSO, but it will not load due to undefined symbols. Any ideas. I am running the latest FDOM (Based on ASU 2008.9). have you managed to put the raster new illume in to FDOM??? :) :) I want to see it? an a step by ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Making sound work (was Re: New Rasterman Image...)
On Mon, October 6, 2008 12:11 am, Richy wrote: I somehow managed to get it to work to some extend. The phone isn' ringin, just vibrating, but I can talk and hear during a call.. make sure you have gstreamer installed, also I added a lot of packages, with alsa, gst, openmoko-state, etc. Didn't work until reboot. (And due to the Neos booting-time I don't want to reboot after each package just to check...) Now, some PIM-software and this is near perfect :-) I am also having trouble with sound. I can hear the other party during calls, but apart from the The phone loudspeaker appears not to work. I don't get any ring tones, and I can't hear music from media plays such as pythm. I have installed pymixer, and have maxed out all the sliders but still nothing. I have also tried installing the snd_pcm_oss kernel module out of FSO, but it will not load due to undefined symbols. Any ideas. I am running the latest FDOM (Based on ASU 2008.9). -- David Pottage Error compiling committee.c To many arguments to function. ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Mon, 6 Oct 2008 14:46:56 +0200, gromez [EMAIL PROTECTED] wrote: On Mon, Oct 6, 2008 at 2:32 PM, gromez [EMAIL PROTECTED] wrote: On Mon, Oct 6, 2008 at 1:11 AM, Richy [EMAIL PROTECTED] wrote: I somehow managed to get it to work to some extend. The phone isn' ringin, just vibrating, but I can talk and hear during a call.. make sure you have gstreamer installed, also I added a lot of packages, with alsa, gst, openmoko-state, etc. Didn't work until reboot. (And due to the Neos booting-time I don't want to reboot after each package just to check...) By default, it seems there's no /dev/dsp so sound can't work. Trying to find the missing package ... Well, it's not a package but snd_pcm_oss module must be loaded. [EMAIL PROTECTED]:~# modprobe snd_pcm_oss Well, I'm still unable to get a peep out of my FreeRunner with the new Rasterman image, except for the alarm. This is with Zhone and frameworkd, plus several alsa and oss packages. I'm using the FSO unstable feed. lsmod tells me: Module Size Used by ipt_LOG 5472 1 xt_tcpudp 2880 2 xt_state1952 3 iptable_filter 2208 1 iptable_nat 5860 0 nf_nat 16118 1 iptable_nat nf_conntrack_ipv4 14344 5 iptable_nat xt_conntrack2368 0 nf_conntrack 58312 5 xt_state,iptable_nat,nf_nat,nf_conntrack_ipv4,xt_conntrack ip_tables 10344 2 iptable_filter,iptable_nat x_tables 11812 6 ipt_LOG,xt_tcpudp,xt_state,iptable_nat,xt_conntrack,ip_tables bnep 12288 2 snd_pcm_oss39872 0 snd_mixer_oss 14720 1 snd_pcm_oss snd_soc_neo1973_gta02_wm8753 7944 0 snd_soc_s3c24xx_i2s 3896 1 snd_soc_neo1973_gta02_wm8753 snd_soc_s3c24xx 4480 1 snd_soc_neo1973_gta02_wm8753 snd_soc_wm8753 29152 1 snd_soc_neo1973_gta02_wm8753 snd_soc_core 27200 3 snd_soc_neo1973_gta02_wm8753,snd_soc_s3c24xx,snd_soc_wm8753 snd_pcm73253 3 snd_pcm_oss,snd_soc_s3c24xx,snd_soc_core snd_timer 20164 1 snd_pcm snd_page_alloc 6120 1 snd_pcm snd46132 7 snd_pcm_oss,snd_mixer_oss,snd_soc_neo1973_gta02_wm8753,snd_soc_wm8753,snd_soc_core,snd_pcm,snd_timer rfcomm 36404 2 ohci_hcd 22060 0 hidp 14784 0 l2cap 21156 13 bnep,rfcomm,hidp hci_usb13980 0 bluetooth 54400 7 bnep,rfcomm,hidp,l2cap,hci_usb Any suggestions on what I've missed? j ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
I think you need to install gst-sid or something like that. I don't have my freerunner here at the moment. Just look at the filetype of the ringtone and opkg list | grep filetype watch out for gst/gstreamer ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Where does FSO stand in regard to rasters releases? Is FSO bleeding edge on Illume and co? I would like to have them in FSO. Since this is the distro i trust the most right now for delivering me the newest and greatest stuff that will also work on my gta01. Marco Trevisan (Treviño) wrote: In the enlightenment archives there's a new Rasterman image: - openmoko-illume-image-glibc-ipk--20080929-om-gta02.rootfs.jffs2 [1] It looks so cool with the new illume theme [2] [3] [4] [5] [6], but practically it has no telephony related software installed in since it seems a simple OE built with these scripts [7], so (I figure) that you're free to put in both the FSO and/or the Om2008.8 stack, just install the needed packages! Bye. [1] http://download.enlightenment.org/misc/Illume/ [2] http://forum.telefoninux.org/index.php/topic,533.0/topicseen.html [3] http://www.rasterman.com/files/illume-01.ogg [4] http://www.rasterman.com/files/illume-02.ogg [5] http://www.rasterman.com/files/illume-03.ogg [6] http://www.rasterman.com/files/illume-04.ogg [7] http://www.rasterman.com/files/oe-setup.tar.gz -- Drucken Sie diese Mail bitte nur auf Recyclingpapier aus. Please print this mail only on recycled paper. ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Mon, Oct 6, 2008 at 1:11 AM, Richy [EMAIL PROTECTED] wrote: I somehow managed to get it to work to some extend. The phone isn' ringin, just vibrating, but I can talk and hear during a call.. make sure you have gstreamer installed, also I added a lot of packages, with alsa, gst, openmoko-state, etc. Didn't work until reboot. (And due to the Neos booting-time I don't want to reboot after each package just to check...) By default, it seems there's no /dev/dsp so sound can't work. Trying to find the missing package ... ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Mon, Oct 6, 2008 at 2:32 PM, gromez [EMAIL PROTECTED] wrote: On Mon, Oct 6, 2008 at 1:11 AM, Richy [EMAIL PROTECTED] wrote: I somehow managed to get it to work to some extend. The phone isn' ringin, just vibrating, but I can talk and hear during a call.. make sure you have gstreamer installed, also I added a lot of packages, with alsa, gst, openmoko-state, etc. Didn't work until reboot. (And due to the Neos booting-time I don't want to reboot after each package just to check...) By default, it seems there's no /dev/dsp so sound can't work. Trying to find the missing package ... Well, it's not a package but snd_pcm_oss module must be loaded. [EMAIL PROTECTED]:~# modprobe snd_pcm_oss ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Mon, 6 Oct 2008 14:46:56 +0200, gromez [EMAIL PROTECTED] wrote: On Mon, Oct 6, 2008 at 2:32 PM, gromez [EMAIL PROTECTED] wrote: On Mon, Oct 6, 2008 at 1:11 AM, Richy [EMAIL PROTECTED] wrote: I somehow managed to get it to work to some extend. The phone isn' ringin, just vibrating, but I can talk and hear during a call.. make sure you have gstreamer installed, also I added a lot of packages, with alsa, gst, openmoko-state, etc. Didn't work until reboot. (And due to the Neos booting-time I don't want to reboot after each package just to check...) By default, it seems there's no /dev/dsp so sound can't work. Trying to find the missing package ... Well, it's not a package but snd_pcm_oss module must be loaded. [EMAIL PROTECTED]:~# modprobe snd_pcm_oss Fantastic - I'll be trying this out ASAP - so long as I can get phone calls ringing and working, I think I want to stick for now with the Raster image plus FSO. echo snd_pcm_oss /etc/modutils.c/snd_pcm_oss update-modules will resolve that more permanently than modprobe'ing, auto-inserting that module (and dependancies) on boot. j ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
2008/10/1 The Rasterman Carsten Haitzler [EMAIL PROTECTED]: eg: Press | Guess+dist e e+0 w+1 r+2 d+2 s+1 r r+0 t+1 e+2 f+1 g+2 d+3 k k+0 l+1 o+2 i+3 j+3 d d+0 f+1 s+1 e+1 c+1 r+2 w+2 I assume this takes into account the keyboard layout, so if I use dvorak or some random layout, it'll do the right thing. Right? ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Mon, 6 Oct 2008 14:01:14 -0400 Chris Wright [EMAIL PROTECTED] babbled: 2008/10/1 The Rasterman Carsten Haitzler [EMAIL PROTECTED]: eg: Press | Guess+dist e e+0 w+1 r+2 d+2 s+1 r r+0 t+1 e+2 f+1 g+2 d+3 k k+0 l+1 o+2 i+3 j+3 d d+0 f+1 s+1 e+1 c+1 r+2 w+2 I assume this takes into account the keyboard layout, so if I use dvorak or some random layout, it'll do the right thing. Right? yes. of course. i was just using qwerty as the example - it doesnt care what the layout is. a key is an object in a location on the canvas that emits a string (converted to a keystroke) or a direct keystroke (f1, escape, tab, etc.). only keys that emit strings are used for dictionary lookup - those that emit a keystroke end any string composition and then just emit exactly that keystroke. when it looks for nearby keys it hunts through everything in the layout and just bases it on geometry. so it's generic. do a dvorak .kbd file and it'll just work. -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)[EMAIL PROTECTED] ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Thu, 02 Oct 2008 20:52:00 +0200 Marco Trevisan (Treviño) [EMAIL PROTECTED] babbled: Carsten Haitzler (The Rasterman) wrote: it does have concept of frequency orf words. i just dont have any DATA for that. the dict format handles is: word1 word2 word3 OR word1 20 word2 434 word3 1 I was thinking to a way to automatize this a while ago, but I wrote something just now... The basic idea is that of using the google number of results for each word and using this value as a frequency number (well, I know these numbers are often too much great, so I guess that they should be re-analyzed and lowered but I had no time to do this now :P). So this is a little utility I wrote [1] to check the frequency of each word and writing back a new dictionary with frequency data. To run it you need php-cli (I guess v5 or above), set the given options, do php words-popularity.php and wait the work to be finished! :P It could be a long work, but it should give good results. yes. it would. who wants to run it? :) nb. i checked illume's kbd code - it does have issues with utf8 keysequences in sorted dicts. if you have any it'll fail to keep looking for more words so you need to remove anything utf8 from your dict :( yes - i know. bad. i need to address this. and the change in dict format i am sure 1. makes this now simple, 2. compresses the dict, 3. speeds it up, 4. solves this problem. :) but i just need to do it - no time right now :( PS: I've used php since I run it both on my PC and on a server (dividing the work) where I've ssh access but in which I can run by command line just a little subset of languages, and php is one of this. [1] http://3v1n0.tuxfamily.org/openmoko/words-popularity.phps -- Treviño's World - Life and Linux http://www.3v1n0.net/ ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)[EMAIL PROTECTED] ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Hey, I'm trying to get FSO working beneath the Raster-Image. I installed zhone. That drew in a lot of dependencies, which is fine. However, I couldn't connect to DBUS, even though dbus is installed, but frameworkd wasn't. (Why itsn't it a dependency of of zhone?) Now, I don't have any sound in zhone. No ring tones, I can't hear anyone talking. Rasters alarm is working though. what package am I missing here? Richard ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Fri, 3 Oct 2008 09:34:28 +1000 Carsten Haitzler (The Rasterman) [EMAIL PROTECTED] wrote: On Thu, 02 Oct 2008 09:37:58 -0700 Steve Mosher [EMAIL PROTECTED] babbled: I used to have a bunch of them when I was doing a NLG ( natural Language generation) pet project. I sent you a link to US names as well. from the US census. For personal dictionaries, people could just run a simple word frequency analysis on their archived email,( there are GPL programs that do this I believe, but its dead easy to write yourself) and import their email contacts into the database. in fact that is the idea of the 3 dictionaries the keyboard has. it has system (which is base language - eg english), personal (any words they type in at all go in here - they inherit frequency they had before but now gain in count as they get used more), and.. generated dictionary - ~/.e/e/dicts-dynamic/data.dic - this file is expected to be regularly generated from the users sms's, emails, contact list etc. containing words from their every-day activity - so that friend with a strange name... gets their name into the dictionary pool this way. :) Is there a way to tie into bash_completion when we are on a terminal? That should be fun :-) Regards, Daniel Willmann signature.asc Description: PGP signature ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Sun, 5 Oct 2008 22:01:30 +0800 Daniel Willmann [EMAIL PROTECTED] babbled: Is there a way to tie into bash_completion when we are on a terminal? That should be fun :-) none really. that info is private to the shell - which is private (and insulated) from the wm in turn. Regards, Daniel Willmann -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)[EMAIL PROTECTED] ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Raster, I don't think I was exactly clear on using the email database. What I was thinking is that you could run a simple program on your desktop computer and do a frequency analysis of the mails ( and contacts) that you have on your PC. Same with chat logs. Then merge that (somehow) into the database on the phone, say in the personal. Daniel Willmann wrote: On Fri, 3 Oct 2008 09:34:28 +1000 Carsten Haitzler (The Rasterman) [EMAIL PROTECTED] wrote: On Thu, 02 Oct 2008 09:37:58 -0700 Steve Mosher [EMAIL PROTECTED] babbled: I used to have a bunch of them when I was doing a NLG ( natural Language generation) pet project. I sent you a link to US names as well. from the US census. For personal dictionaries, people could just run a simple word frequency analysis on their archived email,( there are GPL programs that do this I believe, but its dead easy to write yourself) and import their email contacts into the database. in fact that is the idea of the 3 dictionaries the keyboard has. it has system (which is base language - eg english), personal (any words they type in at all go in here - they inherit frequency they had before but now gain in count as they get used more), and.. generated dictionary - ~/.e/e/dicts-dynamic/data.dic - this file is expected to be regularly generated from the users sms's, emails, contact list etc. containing words from their every-day activity - so that friend with a strange name... gets their name into the dictionary pool this way. :) Is there a way to tie into bash_completion when we are on a terminal? That should be fun :-) Regards, Daniel Willmann ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Hi Richard Did you install frameworkd from fso repository and its dependencies? Worked for me. Markus On Sun, 5 Oct 2008 15:55:33 +0200 Richy [EMAIL PROTECTED] wrote: Hey, I'm trying to get FSO working beneath the Raster-Image. I installed zhone. That drew in a lot of dependencies, which is fine. However, I couldn't connect to DBUS, even though dbus is installed, but frameworkd wasn't. (Why itsn't it a dependency of of zhone?) Now, I don't have any sound in zhone. No ring tones, I can't hear anyone talking. Rasters alarm is working though. what package am I missing here? Richard ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Sun, 5 Oct 2008 19:38:14 +0200, Markus Huggler [EMAIL PROTECTED] wrote: Hi Richard Did you install frameworkd from fso repository and its dependencies? Worked for me. Markus On Sun, 5 Oct 2008 15:55:33 +0200 Richy [EMAIL PROTECTED] wrote: Hey, I'm trying to get FSO working beneath the Raster-Image. I installed zhone. That drew in a lot of dependencies, which is fine. However, I couldn't connect to DBUS, even though dbus is installed, but frameworkd wasn't. (Why itsn't it a dependency of of zhone?) Now, I don't have any sound in zhone. No ring tones, I can't hear anyone talking. Rasters alarm is working though. what package am I missing here? Richard I'm in a similar place. Starting with Raster's new image (pretty and lightweight, BTW) I added zhone and frameworkd, but I have no sound within zhone, no audio either direction in calls. Is there some other package required that handles the audio states, or what? j ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Sun, 05 Oct 2008 09:28:19 -0700 Steve Mosher [EMAIL PROTECTED] babbled: Raster, I don't think I was exactly clear on using the email database. What I was thinking is that you could run a simple program on your desktop computer and do a frequency analysis of the mails ( and contacts) that you have on your PC. Same with chat logs. Then merge that (somehow) into the database on the phone, say in the personal. i know - i don't think that's a great idea. it'd be full of typos! :) if it's going to pollute the dictionary with typos... they should be the user's own! :) btw - i added the first/last names in and man did completion quality go down. too many names. a lot of them can be strange and look nonsensical when you actually use the dict with it all in - it suggests lots of garbage (so now suggestion lists are huge and most of them are entire junk as they are some bizarre name). Daniel Willmann wrote: On Fri, 3 Oct 2008 09:34:28 +1000 Carsten Haitzler (The Rasterman) [EMAIL PROTECTED] wrote: On Thu, 02 Oct 2008 09:37:58 -0700 Steve Mosher [EMAIL PROTECTED] babbled: I used to have a bunch of them when I was doing a NLG ( natural Language generation) pet project. I sent you a link to US names as well. from the US census. For personal dictionaries, people could just run a simple word frequency analysis on their archived email,( there are GPL programs that do this I believe, but its dead easy to write yourself) and import their email contacts into the database. in fact that is the idea of the 3 dictionaries the keyboard has. it has system (which is base language - eg english), personal (any words they type in at all go in here - they inherit frequency they had before but now gain in count as they get used more), and.. generated dictionary - ~/.e/e/dicts-dynamic/data.dic - this file is expected to be regularly generated from the users sms's, emails, contact list etc. containing words from their every-day activity - so that friend with a strange name... gets their name into the dictionary pool this way. :) Is there a way to tie into bash_completion when we are on a terminal? That should be fun :-) Regards, Daniel Willmann ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)[EMAIL PROTECTED] ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
I somehow managed to get it to work to some extend. The phone isn' ringin, just vibrating, but I can talk and hear during a call.. make sure you have gstreamer installed, also I added a lot of packages, with alsa, gst, openmoko-state, etc. Didn't work until reboot. (And due to the Neos booting-time I don't want to reboot after each package just to check...) Now, some PIM-software and this is near perfect :-) On Sun, Oct 5, 2008 at 11:53 PM, Joel Newkirk [EMAIL PROTECTED] wrote: On Sun, 5 Oct 2008 19:38:14 +0200, Markus Huggler [EMAIL PROTECTED] wrote: Hi Richard Did you install frameworkd from fso repository and its dependencies? Worked for me. Markus On Sun, 5 Oct 2008 15:55:33 +0200 Richy [EMAIL PROTECTED] wrote: Hey, I'm trying to get FSO working beneath the Raster-Image. I installed zhone. That drew in a lot of dependencies, which is fine. However, I couldn't connect to DBUS, even though dbus is installed, but frameworkd wasn't. (Why itsn't it a dependency of of zhone?) Now, I don't have any sound in zhone. No ring tones, I can't hear anyone talking. Rasters alarm is working though. what package am I missing here? Richard I'm in a similar place. Starting with Raster's new image (pretty and lightweight, BTW) I added zhone and frameworkd, but I have no sound within zhone, no audio either direction in calls. Is there some other package required that handles the audio states, or what? j ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Mon, 6 Oct 2008 01:11:39 +0200 Richy [EMAIL PROTECTED] babbled: I somehow managed to get it to work to some extend. The phone isn' ringin, just vibrating, but I can talk and hear during a call.. make sure you have gstreamer installed, also I added a lot of packages, with alsa, gst, openmoko-state, etc. Didn't work until reboot. (And due to the Neos booting-time I don't want to reboot after each package just to check...) Now, some PIM-software and this is near perfect :-) coo. btw - i will eventually one by one suck in frameworkd and other apps - but i want to be very high quality about this - only include in the base what absolutely works or works well - or what i'm directly developing at the time (as i then have direct ability time etc. to make it work well). for now it's just a start. needs much more polish. -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)[EMAIL PROTECTED] ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Spelling issue Noted. RE: first and last names: I suppose it was worth the try?; Was it primarily the last names (US is a nation of immigrants) that gave the most trouble? or first names as well? The last names file is huge in comparision to the first name files, FWIW. and Looking at the list of names and graphing frequency versus rank, it looks like it follows a power law distribution (http://en.wikipedia.org/wiki/Power_law) to get to 80% of each gender you need : 313 male names, 811 female names). The top 100 male names is 60% of all male names. The top 100 female names is 43% of all female names. last names, the top 100 are only 18% of the population. FWIW, word frequency is known to follow a power law distribution, so I'm not surprised that proper names appear to follow a power law. Ok, no more backseat driving. Hope all is well Carsten Haitzler (The Rasterman) wrote: On Sun, 05 Oct 2008 09:28:19 -0700 Steve Mosher [EMAIL PROTECTED] babbled: Raster, I don't think I was exactly clear on using the email database. What I was thinking is that you could run a simple program on your desktop computer and do a frequency analysis of the mails ( and contacts) that you have on your PC. Same with chat logs. Then merge that (somehow) into the database on the phone, say in the personal. i know - i don't think that's a great idea. it'd be full of typos! :) if it's going to pollute the dictionary with typos... they should be the user's own! :) btw - i added the first/last names in and man did completion quality go down. too many names. a lot of them can be strange and look nonsensical when you actually use the dict with it all in - it suggests lots of garbage (so now suggestion lists are huge and most of them are entire junk as they are some bizarre name). Daniel Willmann wrote: On Fri, 3 Oct 2008 09:34:28 +1000 Carsten Haitzler (The Rasterman) [EMAIL PROTECTED] wrote: On Thu, 02 Oct 2008 09:37:58 -0700 Steve Mosher [EMAIL PROTECTED] babbled: I used to have a bunch of them when I was doing a NLG ( natural Language generation) pet project. I sent you a link to US names as well. from the US census. For personal dictionaries, people could just run a simple word frequency analysis on their archived email,( there are GPL programs that do this I believe, but its dead easy to write yourself) and import their email contacts into the database. in fact that is the idea of the 3 dictionaries the keyboard has. it has system (which is base language - eg english), personal (any words they type in at all go in here - they inherit frequency they had before but now gain in count as they get used more), and.. generated dictionary - ~/.e/e/dicts-dynamic/data.dic - this file is expected to be regularly generated from the users sms's, emails, contact list etc. containing words from their every-day activity - so that friend with a strange name... gets their name into the dictionary pool this way. :) Is there a way to tie into bash_completion when we are on a terminal? That should be fun :-) Regards, Daniel Willmann ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Looks excellent! Can't seem to get suspend to work though. Is it because it's missing 'apm' or something? Also, would it be hard to port this to 2009.9? Is it more than a change of edje files? -Nick Marco Trevisan (Treviño) wrote: In the enlightenment archives there's a new Rasterman image: - openmoko-illume-image-glibc-ipk--20080929-om-gta02.rootfs.jffs2 [1] It looks so cool with the new illume theme [2] [3] [4] [5] [6], but practically it has no telephony related software installed in since it seems a simple OE built with these scripts [7], so (I figure) that you're free to put in both the FSO and/or the Om2008.8 stack, just install the needed packages! Bye. [1] http://download.enlightenment.org/misc/Illume/ [2] http://forum.telefoninux.org/index.php/topic,533.0/topicseen.html [3] http://www.rasterman.com/files/illume-01.ogg [4] http://www.rasterman.com/files/illume-02.ogg [5] http://www.rasterman.com/files/illume-03.ogg [6] http://www.rasterman.com/files/illume-04.ogg [7] http://www.rasterman.com/files/oe-setup.tar.gz ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Thu, 02 Oct 2008 17:00:53 +1000 nickd [EMAIL PROTECTED] babbled: Looks excellent! Can't seem to get suspend to work though. Is it because it's missing 'apm' or something? Also, would it be hard to port this to 2009.9? Is it more than a change of edje files? its got it - thew FSO x builds dont allow aux and power buttons to be keys in x - so e can never get them (and no - i'm not going to do awful hacks like open kernel input devices directly when it should be punted through x). there is apm - but for some reason for me it wont wake up from suspend - no idea why. i havent really cared. thats a kernel problem and not somethnig i am wanting to focus on :( -Nick Marco Trevisan (Treviño) wrote: In the enlightenment archives there's a new Rasterman image: - openmoko-illume-image-glibc-ipk--20080929-om-gta02.rootfs.jffs2 [1] It looks so cool with the new illume theme [2] [3] [4] [5] [6], but practically it has no telephony related software installed in since it seems a simple OE built with these scripts [7], so (I figure) that you're free to put in both the FSO and/or the Om2008.8 stack, just install the needed packages! Bye. [1] http://download.enlightenment.org/misc/Illume/ [2] http://forum.telefoninux.org/index.php/topic,533.0/topicseen.html [3] http://www.rasterman.com/files/illume-01.ogg [4] http://www.rasterman.com/files/illume-02.ogg [5] http://www.rasterman.com/files/illume-03.ogg [6] http://www.rasterman.com/files/illume-04.ogg [7] http://www.rasterman.com/files/oe-setup.tar.gz ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)[EMAIL PROTECTED] ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Thu, Oct 2, 2008 at 05:57, The Rasterman Carsten Haitzler [EMAIL PROTECTED] wrote: [...] you will know its not there when its not offered as a correction in the short or long list. you get to delete what u typed and type again carefully. you only need to do this once. then it goes into the dict. [...] seriously - as an experiment, when you type a word and it gets the guess wrong pop up the full list and at the top is EXACTLY what you typed. [..] What I really miss is having what I really typed be immediatly available as an option (ie not having to pop up a list). Personnaly I would even prefer by far (but might be an option) that what I type is what is selected by default, and a computed word is shown in the box that I can choose instead, if I want... and even writing english text messages, it is too often I have to enter unknown names, locations, emails, whatever ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
raster the list I sent you was from the British National Corpus, but should be a good start. Carsten Haitzler (The Rasterman) wrote: On Wed, 1 Oct 2008 21:05:53 -0600 Ori Pessach [EMAIL PROTECTED] babbled: I understand what it's doing. It's not doing it well. I tried it for shell i disagree. it works like a charm for me - as per my previous mail - i can use it while walking down the street. more than i can say for pretty much any other virtual keyboard i have available to me. input, and it was an unusable mess. I tried it for text messaging, and it why someone would use a language dictionary-based corrective keyboard for shell input beats me! in this case i call silly user - using a motorcycle to deliver elephants line :) use the terminal keyboard. use a stylus. thats what it was meant for. :) was an unusable mess. It has no model of the likelihood of erroneous input it does. it absolutely does. maybe your fingers are incredibly off-center? here is the algorithm (and if u don't believe me - code is there to be read): it stores a press POINT (x,y). it looks for all keys whose center point is WITHIN f distance of x,y (f being the fuzz value - the .kbd file for the qwerty Default keyboard is 135 units wide, with fuzz radius of 20, so that's about 1/3rd of the keyboard that it searches through for a likely match). likelihood factors (distance) per key found is allocated based on distance (0 == most likely, 0 less likely the greater the value). each press is done this way EXCEPT if u hold for 0.25 sec then drag to select a key explicitly in zoom mode - then the ONLY key available for that word slot is that letter selected given a distance of 0. as you type all permutations of letters are searched and put into a list - with each permutation given a distance metric based on the letters used (simply addition of the distances). now this is combined with the dictionary's frequency metric (multiplied by an inverse) so the more likely the word is to be used the lower its distance becomes. words are sorted from most to least likely based on this metric then listed with most likely in the middle of the list, leas likely to the left/right ends - which you may not see. the vertical list lists all matches from most to least likely (top to bottom) with 1 exception - EXACTLY what u typed it as the top. it absolutely has a fairly good idea of likelihood of error and likelihood of usage of a word etc. etc. eg: Press | Guess+dist e e+0 w+1 r+2 d+2 s+1 r r+0 t+1 e+2 f+1 g+2 d+3 k k+0 l+1 o+2 i+3 j+3 d d+0 f+1 s+1 e+1 c+1 r+2 w+2 so erkd has distance 0 = but its not a word in the dictionary at all, so thrown out. rwkd has distance 1, but not a word, srkd same, etkd, efkd, erld, etc. etc. in the end it produces a list where most likely world ends up the word with other options too - and this is a much simplified list. mostly the list for candidate letters per input letter is about 10-12 letters. so u have 12*12*12*12 permutations for a 4 letter word - of which a fraction of that space is legitimate words. each permutation has a likelihood value based on press distance and on frequency of usage of that word in language in general in the dictionary. mind you - i AM talking about illume's keyboard, its algorithms as is in the image i built. if you use something else i cannot comment as it's something else. (relatively low) and instead appears to look for the word with the closest minimum edit distance to the user's input. This is nuts. I have never - it's not - as the edit distance is the likelihood of error. you likely press the key you want - or near it. thus keys near where you pressed are more likely than those further away. to limit search distance only up to a certain distance is searched. chances are that you do this: fingerprint: ___ /~~~\ |~~~| |~~~| \x/ where x is the pressure point reported on the touchscreen. the only info the touchscreen reports is the pressure point - nothing else. you think u press somewhere else, but don't. you know what u pressed bu what key pops up that lets u know pretty well how good your pressing of the screen is. this is just a hardware limit of a resistive touchscreen. the point of greatest pressure is used - not the middle point of the area in which skin contacts the screen. get the gpe-sketchbook and try press with the flat of your finger and see just of far off your press point is. it may surprise you. as i said - it does have all the model and code and even data to do proper correction based on many factors. i do NOT have a dictionary with frequency info for all of english - there is a small english dict (5000 words) with some frequency info in it i managed to gather, but its very small. if you don't believe me - read the code, or do better. patches accepted, but i think the problem is just that the dictionary
Re: New Rasterman Image...
Quoting Joel Newkirk [EMAIL PROTECTED]: On Wed, 1 Oct 2008 20:32:34 -0600, Ori Pessach [EMAIL PROTECTED] wrote: I have a question about the predictive keyboard shown in one of those movies snip --Ori Pessach Intended, you just have to understand what it's doing. Unlike T9 and other predictive inputs that are familiar from cellphones, the keyboard predictive process looks at neighboring keys, 'predicting' that periodically you may hit the wrong one. (if you're using your finger that's pretty likely, especially if they're as wide as mine) At first look I thought this keyboard was useless. Coming back to it later I started to see how it worked, but still switched to a full dvorak layout with stylus all the time. Seeing this thread I took another look ... stylus ... finger...! Now I love it! I can mash all over the keyboard with my fat fingers and it makes the right words available almost every time! No more squinting in poor light trying to get exactly the right key with the stylus. I think I'll be using this predictive keyboard a lot more. IanS ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Thu, 2 Oct 2008 05:12:44 -0400 Ian Stephen [EMAIL PROTECTED] babbled: Quoting Joel Newkirk [EMAIL PROTECTED]: On Wed, 1 Oct 2008 20:32:34 -0600, Ori Pessach [EMAIL PROTECTED] wrote: I have a question about the predictive keyboard shown in one of those movies snip --Ori Pessach Intended, you just have to understand what it's doing. Unlike T9 and other predictive inputs that are familiar from cellphones, the keyboard predictive process looks at neighboring keys, 'predicting' that periodically you may hit the wrong one. (if you're using your finger that's pretty likely, especially if they're as wide as mine) At first look I thought this keyboard was useless. Coming back to it later I started to see how it worked, but still switched to a full dvorak layout with stylus all the time. Seeing this thread I took another look ... stylus ... finger...! Now I love it! I can mash all over the keyboard with my fat fingers and it makes the right words available almost every time! No more squinting in poor light trying to get exactly the right key with the stylus. I think I'll be using this predictive keyboard a lot more. and he started doing the monster mash! mash mash mash! :) now you've got the spirit! :) yes. if you want to hen-peck with a stylus - use terminal (or use a terminal where a dictionary makes no sense). if you want to write english and mach about with your fingers! use default (alpha) layout and let the dictionary fix up your monster mash. :) mash more and more over time and its idea of what words are more important that others to match will improve over time as you use it. of course words not in the dictionary can't be de-mashed, so just hold and drag in zoom moved to enter them just once (per word not in the dict)... (have the patience) and thereafter... mash away!. -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)[EMAIL PROTECTED] ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Thu, 02 Oct 2008 00:59:20 -0700 Steve Mosher [EMAIL PROTECTED] babbled: aha! a decent frequency corpus (a few thousand words). i'll merge this with the default us dict (and remove the small one as now that's useless). Hey raster, How's it going. I promised you some frequency data a while back. http://ucrel.lancs.ac.uk/bncfreq/flists.html http://ucrel.lancs.ac.uk/bncfreq/lists/1_2_all_freq.txt there are others as well Carsten Haitzler (The Rasterman) wrote: On Wed, 1 Oct 2008 21:05:53 -0600 Ori Pessach [EMAIL PROTECTED] babbled: I understand what it's doing. It's not doing it well. I tried it for shell i disagree. it works like a charm for me - as per my previous mail - i can use it while walking down the street. more than i can say for pretty much any other virtual keyboard i have available to me. input, and it was an unusable mess. I tried it for text messaging, and it why someone would use a language dictionary-based corrective keyboard for shell input beats me! in this case i call silly user - using a motorcycle to deliver elephants line :) use the terminal keyboard. use a stylus. thats what it was meant for. :) was an unusable mess. It has no model of the likelihood of erroneous input it does. it absolutely does. maybe your fingers are incredibly off-center? here is the algorithm (and if u don't believe me - code is there to be read): it stores a press POINT (x,y). it looks for all keys whose center point is WITHIN f distance of x,y (f being the fuzz value - the .kbd file for the qwerty Default keyboard is 135 units wide, with fuzz radius of 20, so that's about 1/3rd of the keyboard that it searches through for a likely match). likelihood factors (distance) per key found is allocated based on distance (0 == most likely, 0 less likely the greater the value). each press is done this way EXCEPT if u hold for 0.25 sec then drag to select a key explicitly in zoom mode - then the ONLY key available for that word slot is that letter selected given a distance of 0. as you type all permutations of letters are searched and put into a list - with each permutation given a distance metric based on the letters used (simply addition of the distances). now this is combined with the dictionary's frequency metric (multiplied by an inverse) so the more likely the word is to be used the lower its distance becomes. words are sorted from most to least likely based on this metric then listed with most likely in the middle of the list, leas likely to the left/right ends - which you may not see. the vertical list lists all matches from most to least likely (top to bottom) with 1 exception - EXACTLY what u typed it as the top. it absolutely has a fairly good idea of likelihood of error and likelihood of usage of a word etc. etc. eg: Press | Guess+dist e e+0 w+1 r+2 d+2 s+1 r r+0 t+1 e+2 f+1 g+2 d+3 k k+0 l+1 o+2 i+3 j+3 d d+0 f+1 s+1 e+1 c+1 r+2 w+2 so erkd has distance 0 = but its not a word in the dictionary at all, so thrown out. rwkd has distance 1, but not a word, srkd same, etkd, efkd, erld, etc. etc. in the end it produces a list where most likely world ends up the word with other options too - and this is a much simplified list. mostly the list for candidate letters per input letter is about 10-12 letters. so u have 12*12*12*12 permutations for a 4 letter word - of which a fraction of that space is legitimate words. each permutation has a likelihood value based on press distance and on frequency of usage of that word in language in general in the dictionary. mind you - i AM talking about illume's keyboard, its algorithms as is in the image i built. if you use something else i cannot comment as it's something else. (relatively low) and instead appears to look for the word with the closest minimum edit distance to the user's input. This is nuts. I have never - it's not - as the edit distance is the likelihood of error. you likely press the key you want - or near it. thus keys near where you pressed are more likely than those further away. to limit search distance only up to a certain distance is searched. chances are that you do this: fingerprint: ___ /~~~\ |~~~| |~~~| \x/ where x is the pressure point reported on the touchscreen. the only info the touchscreen reports is the pressure point - nothing else. you think u press somewhere else, but don't. you know what u pressed bu what key pops up that lets u know pretty well how good your pressing of the screen is. this is just a hardware limit of a resistive touchscreen. the point of greatest pressure is used - not the middle point of the area in which skin contacts the screen. get the gpe-sketchbook and try press with the flat of your finger and see just of far off your press point is. it may surprise you. as i said -
Re: New Rasterman Image...
El jue, 02-10-2008 a las 03:18 +0200, Marco Trevisan (Treviño) escribió: In the enlightenment archives there's a new Rasterman image: - openmoko-illume-image-glibc-ipk--20080929-om-gta02.rootfs.jffs2 [1] It looks so cool with the new illume theme [2] [3] [4] [5] [6], but practically it has no telephony related software installed in since it seems a simple OE built with these scripts [7], so (I figure) that you're free to put in both the FSO and/or the Om2008.8 stack, just install the needed packages! Any way to doing this inverted? from a already build 2008.9 stack add this beautifull front end? Bye. [1] http://download.enlightenment.org/misc/Illume/ [2] http://forum.telefoninux.org/index.php/topic,533.0/topicseen.html [3] http://www.rasterman.com/files/illume-01.ogg [4] http://www.rasterman.com/files/illume-02.ogg [5] http://www.rasterman.com/files/illume-03.ogg [6] http://www.rasterman.com/files/illume-04.ogg [7] http://www.rasterman.com/files/oe-setup.tar.gz ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Any way to doing this inverted? from a already build 2008.9 stack add this beautifull front end? i'm interested in this last option... ;-) Rasterman illume us! d On Thu, Oct 2, 2008 at 1:11 PM, David Samblas [EMAIL PROTECTED] wrote: El jue, 02-10-2008 a las 03:18 +0200, Marco Trevisan (Treviño) escribió: In the enlightenment archives there's a new Rasterman image: - openmoko-illume-image-glibc-ipk--20080929-om-gta02.rootfs.jffs2 [1] It looks so cool with the new illume theme [2] [3] [4] [5] [6], but practically it has no telephony related software installed in since it seems a simple OE built with these scripts [7], so (I figure) that you're free to put in both the FSO and/or the Om2008.8 stack, just install the needed packages! Any way to doing this inverted? from a already build 2008.9 stack add this beautifull front end? Bye. [1] http://download.enlightenment.org/misc/Illume/ [2] http://forum.telefoninux.org/index.php/topic,533.0/topicseen.html [3] http://www.rasterman.com/files/illume-01.ogg [4] http://www.rasterman.com/files/illume-02.ogg [5] http://www.rasterman.com/files/illume-03.ogg [6] http://www.rasterman.com/files/illume-04.ogg [7] http://www.rasterman.com/files/oe-setup.tar.gz ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Carsten Haitzler (The Rasterman) wrote: On Wed, 1 Oct 2008 21:05:53 -0600 Ori Pessach [EMAIL PROTECTED] babbled: I understand what it's doing. It's not doing it well. I tried it for shell i disagree. it works like a charm for me - as per my previous mail - i can use it while walking down the street. more than i can say for pretty much any other virtual keyboard i have available to me. I do agree with raster... I'm using it with no problems and always without the stylus. Since when I've bought the Freerunner I've used the stylus less than 3-4 times. I always tried to do all with my fingers and generally I got it; mostly thanks to Carsten's work. Of course, there are some things that could be improved (i.e. parsing big dictionaries) but it's all generally so much usable! look at the personal dict file. ~/.e/e/dicts-dynamic/personal.dic it saves usage frequency. About this, did you see my reported bug #2049? And more... When the personal addictions are stored there? Since it doesn't seem to grow as soon as I type (or am I simply too blind? :P). Thanks... -- Treviño's World - Life and Linux http://www.3v1n0.net/ ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Thu, 02 Oct 2008 17:10:27 +0200 Marco Trevisan (Treviño) [EMAIL PROTECTED] babbled: Carsten Haitzler (The Rasterman) wrote: On Wed, 1 Oct 2008 21:05:53 -0600 Ori Pessach [EMAIL PROTECTED] babbled: I understand what it's doing. It's not doing it well. I tried it for shell i disagree. it works like a charm for me - as per my previous mail - i can use it while walking down the street. more than i can say for pretty much any other virtual keyboard i have available to me. I do agree with raster... I'm using it with no problems and always without the stylus. Since when I've bought the Freerunner I've used the stylus less than 3-4 times. I always tried to do all with my fingers and generally I got it; mostly thanks to Carsten's work. Of course, there are some things that could be improved (i.e. parsing big dictionaries) but it's all generally so much usable! yeah. saw your trac bug. i know. it's just out of my focus at the moment. i'll eventually cycle back around to it - and i'll not fix the bugs - i'll probably do some radical surgery to the dict format and enigne. :) look at the personal dict file. ~/.e/e/dicts-dynamic/personal.dic it saves usage frequency. About this, did you see my reported bug #2049? And more... When the personal addictions are stored there? Since it doesn't seem to grow as soon as I type (or am I simply too blind? :P). i saw the one with duplicates - that shouldn't happen - but there likely is some bug that makes it - somewhere... shouldn't. it saves - just not immediately. you need to keep the dict idle for 5+ seconds - then it flushes and saves. that means not adding any new words or modifying them for 5 seconds (ie take a pause from typing and it'll flush and save). -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)[EMAIL PROTECTED] ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Lists of US names. first and last http://www.census.gov/genealogy/names/names_files.html Cédric Berger wrote: On Thu, Oct 2, 2008 at 05:57, The Rasterman Carsten Haitzler [EMAIL PROTECTED] wrote: [...] you will know its not there when its not offered as a correction in the short or long list. you get to delete what u typed and type again carefully. you only need to do this once. then it goes into the dict. [...] seriously - as an experiment, when you type a word and it gets the guess wrong pop up the full list and at the top is EXACTLY what you typed. [..] What I really miss is having what I really typed be immediatly available as an option (ie not having to pop up a list). Personnaly I would even prefer by far (but might be an option) that what I type is what is selected by default, and a computed word is shown in the box that I can choose instead, if I want... and even writing english text messages, it is too often I have to enter unknown names, locations, emails, whatever ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Cédric Berger wrote: What I really miss is having what I really typed be immediatly available as an option (ie not having to pop up a list). Personnaly I would even prefer by far (but might be an option) that what I type is what is selected by default, and a computed word is shown in the box that I can choose instead, if I want... and even writing english text messages, it is too often I have to enter unknown names, locations, emails, whatever Well, I can agree in this. I think that the selected word/char should be set as default (i.e. I often write in Italian thinks like l'many-vocal-starting-word, since adding any word matching to [clmstv]'many-vocal-starting-word would increase a lot the dictionary I do prefer using the ' as a non-dictionary char so when I've to write a word like that I, for example, press l, then ' and finally the word. However this doesn't work always since the l is quite never selected by default also if I press exactly on it; so I've to press l, select it in the list, do a backspace and finally writing the word. You can figure that it's not so good :P). -- Treviño's World - Life and Linux http://www.3v1n0.net/ ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
I used to have a bunch of them when I was doing a NLG ( natural Language generation) pet project. I sent you a link to US names as well. from the US census. For personal dictionaries, people could just run a simple word frequency analysis on their archived email,( there are GPL programs that do this I believe, but its dead easy to write yourself) and import their email contacts into the database. ( speeling mistakes might require some work, like the one I just did) If you had access to archived chats or chat logs you could pick up things like LOL, PITA, etc, or logs of SMS. There are some studies on word frequency in SMS but I havent found a online resource. Carsten Haitzler (The Rasterman) wrote: On Thu, 02 Oct 2008 00:59:20 -0700 Steve Mosher [EMAIL PROTECTED] babbled: aha! a decent frequency corpus (a few thousand words). i'll merge this with the default us dict (and remove the small one as now that's useless). Hey raster, How's it going. I promised you some frequency data a while back. http://ucrel.lancs.ac.uk/bncfreq/flists.html http://ucrel.lancs.ac.uk/bncfreq/lists/1_2_all_freq.txt there are others as well Carsten Haitzler (The Rasterman) wrote: On Wed, 1 Oct 2008 21:05:53 -0600 Ori Pessach [EMAIL PROTECTED] babbled: I understand what it's doing. It's not doing it well. I tried it for shell i disagree. it works like a charm for me - as per my previous mail - i can use it while walking down the street. more than i can say for pretty much any other virtual keyboard i have available to me. input, and it was an unusable mess. I tried it for text messaging, and it why someone would use a language dictionary-based corrective keyboard for shell input beats me! in this case i call silly user - using a motorcycle to deliver elephants line :) use the terminal keyboard. use a stylus. thats what it was meant for. :) was an unusable mess. It has no model of the likelihood of erroneous input it does. it absolutely does. maybe your fingers are incredibly off-center? here is the algorithm (and if u don't believe me - code is there to be read): it stores a press POINT (x,y). it looks for all keys whose center point is WITHIN f distance of x,y (f being the fuzz value - the .kbd file for the qwerty Default keyboard is 135 units wide, with fuzz radius of 20, so that's about 1/3rd of the keyboard that it searches through for a likely match). likelihood factors (distance) per key found is allocated based on distance (0 == most likely, 0 less likely the greater the value). each press is done this way EXCEPT if u hold for 0.25 sec then drag to select a key explicitly in zoom mode - then the ONLY key available for that word slot is that letter selected given a distance of 0. as you type all permutations of letters are searched and put into a list - with each permutation given a distance metric based on the letters used (simply addition of the distances). now this is combined with the dictionary's frequency metric (multiplied by an inverse) so the more likely the word is to be used the lower its distance becomes. words are sorted from most to least likely based on this metric then listed with most likely in the middle of the list, leas likely to the left/right ends - which you may not see. the vertical list lists all matches from most to least likely (top to bottom) with 1 exception - EXACTLY what u typed it as the top. it absolutely has a fairly good idea of likelihood of error and likelihood of usage of a word etc. etc. eg: Press | Guess+dist e e+0 w+1 r+2 d+2 s+1 r r+0 t+1 e+2 f+1 g+2 d+3 k k+0 l+1 o+2 i+3 j+3 d d+0 f+1 s+1 e+1 c+1 r+2 w+2 so erkd has distance 0 = but its not a word in the dictionary at all, so thrown out. rwkd has distance 1, but not a word, srkd same, etkd, efkd, erld, etc. etc. in the end it produces a list where most likely world ends up the word with other options too - and this is a much simplified list. mostly the list for candidate letters per input letter is about 10-12 letters. so u have 12*12*12*12 permutations for a 4 letter word - of which a fraction of that space is legitimate words. each permutation has a likelihood value based on press distance and on frequency of usage of that word in language in general in the dictionary. mind you - i AM talking about illume's keyboard, its algorithms as is in the image i built. if you use something else i cannot comment as it's something else. (relatively low) and instead appears to look for the word with the closest minimum edit distance to the user's input. This is nuts. I have never - it's not - as the edit distance is the likelihood of error. you likely press the key you want - or near it. thus keys near where you pressed are more likely than those further away. to limit search distance only up to a certain distance is searched. chances are that you do this: fingerprint: ___ /~~~\ |~~~| |~~~| \x/ where x is the
Re: New Rasterman Image...
On Thu, 02 Oct 2008 09:37:58 -0700, Steve Mosher [EMAIL PROTECTED] wrote: I used to have a bunch of them when I was doing a NLG ( natural Language generation) pet project. I sent you a link to US names as well. from the US census. For personal dictionaries, people could just run a simple word frequency analysis on their archived email,( there are GPL programs that do this I believe, but its dead easy to write yourself) and import their email contacts into the database. ( speeling mistakes might require some work, like the one I just did) If you had access to archived chats or chat logs you could pick up things like LOL, PITA, etc, or logs of SMS. There are some studies on word frequency in SMS but I havent found a online resource. I was thinking of a quickie program to scan the ~/.bash_history on my desktop and generate frequency data for command names... ;) Take everything up to the first space, strip off any path, count and insert in dict. (unfortunately the command history on the FR is by default VERY VERY short, I've not investigated how to extend it) More useful would be if someone can scare up a thorough list of common SMS shorthand, like 'cul8r' and what-not. (don't know how that'd work out though, with numeric characters embedded) j ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
I tried it. Very nice raster! Good work. Alarm is fantastic :) Now we need only a dialer, contacts and messages compatible with FSO :) ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community -- View this message in context: http://n2.nabble.com/New-Rasterman-Image...-tp1132929p1134577.html Sent from the Openmoko Community mailing list archive at Nabble.com. ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
Carsten Haitzler (The Rasterman) wrote: it does have concept of frequency orf words. i just dont have any DATA for that. the dict format handles is: word1 word2 word3 OR word1 20 word2 434 word3 1 I was thinking to a way to automatize this a while ago, but I wrote something just now... The basic idea is that of using the google number of results for each word and using this value as a frequency number (well, I know these numbers are often too much great, so I guess that they should be re-analyzed and lowered but I had no time to do this now :P). So this is a little utility I wrote [1] to check the frequency of each word and writing back a new dictionary with frequency data. To run it you need php-cli (I guess v5 or above), set the given options, do php words-popularity.php and wait the work to be finished! :P It could be a long work, but it should give good results. PS: I've used php since I run it both on my PC and on a server (dividing the work) where I've ssh access but in which I can run by command line just a little subset of languages, and php is one of this. [1] http://3v1n0.tuxfamily.org/openmoko/words-popularity.phps -- Treviño's World - Life and Linux http://www.3v1n0.net/ ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
I suppose I could compile a list of the academic books done on SMS, contact the authors, and see if they are willing to cough up the data. For oneline dictionaries with no frequency data you have several dictionaries, like http://www.netlingo.com/emailsh.cfm Have not found any downloadable lists with freq data. Joel Newkirk wrote: On Thu, 02 Oct 2008 09:37:58 -0700, Steve Mosher [EMAIL PROTECTED] wrote: I used to have a bunch of them when I was doing a NLG ( natural Language generation) pet project. I sent you a link to US names as well. from the US census. For personal dictionaries, people could just run a simple word frequency analysis on their archived email,( there are GPL programs that do this I believe, but its dead easy to write yourself) and import their email contacts into the database. ( speeling mistakes might require some work, like the one I just did) If you had access to archived chats or chat logs you could pick up things like LOL, PITA, etc, or logs of SMS. There are some studies on word frequency in SMS but I havent found a online resource. I was thinking of a quickie program to scan the ~/.bash_history on my desktop and generate frequency data for command names... ;) Take everything up to the first space, strip off any path, count and insert in dict. (unfortunately the command history on the FR is by default VERY VERY short, I've not investigated how to extend it) More useful would be if someone can scare up a thorough list of common SMS shorthand, like 'cul8r' and what-not. (don't know how that'd work out though, with numeric characters embedded) j ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Thu, 02 Oct 2008 09:37:58 -0700 Steve Mosher [EMAIL PROTECTED] babbled: I used to have a bunch of them when I was doing a NLG ( natural Language generation) pet project. I sent you a link to US names as well. from the US census. For personal dictionaries, people could just run a simple word frequency analysis on their archived email,( there are GPL programs that do this I believe, but its dead easy to write yourself) and import their email contacts into the database. in fact that is the idea of the 3 dictionaries the keyboard has. it has system (which is base language - eg english), personal (any words they type in at all go in here - they inherit frequency they had before but now gain in count as they get used more), and.. generated dictionary - ~/.e/e/dicts-dynamic/data.dic - this file is expected to be regularly generated from the users sms's, emails, contact list etc. containing words from their every-day activity - so that friend with a strange name... gets their name into the dictionary pool this way. :) ( speeling mistakes might require some work, like the one I just did) If you had access to archived chats or chat logs you could pick up things like LOL, PITA, etc, or logs of SMS. There are some studies on word frequency in SMS but I havent found a online resource. yup. and as above. the idea of the generated dictionary would be to capture these live as usage happens :) there's only so much hunting of data i can (and will) do as a developer. it's a never-ending game of gather more data into dictionaries. i need to put in the mechanisms by which this is possible (already done) and leave it up to the vast pool of users to fill that in for each language, country, region etc. :) Carsten Haitzler (The Rasterman) wrote: On Thu, 02 Oct 2008 00:59:20 -0700 Steve Mosher [EMAIL PROTECTED] babbled: aha! a decent frequency corpus (a few thousand words). i'll merge this with the default us dict (and remove the small one as now that's useless). Hey raster, How's it going. I promised you some frequency data a while back. http://ucrel.lancs.ac.uk/bncfreq/flists.html http://ucrel.lancs.ac.uk/bncfreq/lists/1_2_all_freq.txt there are others as well Carsten Haitzler (The Rasterman) wrote: On Wed, 1 Oct 2008 21:05:53 -0600 Ori Pessach [EMAIL PROTECTED] babbled: I understand what it's doing. It's not doing it well. I tried it for shell i disagree. it works like a charm for me - as per my previous mail - i can use it while walking down the street. more than i can say for pretty much any other virtual keyboard i have available to me. input, and it was an unusable mess. I tried it for text messaging, and it why someone would use a language dictionary-based corrective keyboard for shell input beats me! in this case i call silly user - using a motorcycle to deliver elephants line :) use the terminal keyboard. use a stylus. thats what it was meant for. :) was an unusable mess. It has no model of the likelihood of erroneous input it does. it absolutely does. maybe your fingers are incredibly off-center? here is the algorithm (and if u don't believe me - code is there to be read): it stores a press POINT (x,y). it looks for all keys whose center point is WITHIN f distance of x,y (f being the fuzz value - the .kbd file for the qwerty Default keyboard is 135 units wide, with fuzz radius of 20, so that's about 1/3rd of the keyboard that it searches through for a likely match). likelihood factors (distance) per key found is allocated based on distance (0 == most likely, 0 less likely the greater the value). each press is done this way EXCEPT if u hold for 0.25 sec then drag to select a key explicitly in zoom mode - then the ONLY key available for that word slot is that letter selected given a distance of 0. as you type all permutations of letters are searched and put into a list - with each permutation given a distance metric based on the letters used (simply addition of the distances). now this is combined with the dictionary's frequency metric (multiplied by an inverse) so the more likely the word is to be used the lower its distance becomes. words are sorted from most to least likely based on this metric then listed with most likely in the middle of the list, leas likely to the left/right ends - which you may not see. the vertical list lists all matches from most to least likely (top to bottom) with 1 exception - EXACTLY what u typed it as the top. it absolutely has a fairly good idea of likelihood of error and likelihood of usage of a word etc. etc. eg: Press | Guess+dist e e+0 w+1 r+2 d+2 s+1 r r+0 t+1 e+2 f+1 g+2 d+3 k k+0 l+1 o+2 i+3 j+3 d d+0 f+1 s+1 e+1 c+1 r+2 w+2 so erkd has distance 0 = but its not a word in the dictionary at all, so thrown out. rwkd has distance 1, but not a word, srkd same, etkd, efkd, erld, etc. etc. in the
New Rasterman Image...
In the enlightenment archives there's a new Rasterman image: - openmoko-illume-image-glibc-ipk--20080929-om-gta02.rootfs.jffs2 [1] It looks so cool with the new illume theme [2] [3] [4] [5] [6], but practically it has no telephony related software installed in since it seems a simple OE built with these scripts [7], so (I figure) that you're free to put in both the FSO and/or the Om2008.8 stack, just install the needed packages! Bye. [1] http://download.enlightenment.org/misc/Illume/ [2] http://forum.telefoninux.org/index.php/topic,533.0/topicseen.html [3] http://www.rasterman.com/files/illume-01.ogg [4] http://www.rasterman.com/files/illume-02.ogg [5] http://www.rasterman.com/files/illume-03.ogg [6] http://www.rasterman.com/files/illume-04.ogg [7] http://www.rasterman.com/files/oe-setup.tar.gz -- Treviño's World - Life and Linux http://www.3v1n0.net/ ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Wed, Oct 1, 2008 at 7:18 PM, Marco Trevisan (Treviño) [EMAIL PROTECTED]wrote: In the enlightenment archives there's a new Rasterman image: - openmoko-illume-image-glibc-ipk--20080929-om-gta02.rootfs.jffs2 [1] It looks so cool with the new illume theme [2] [3] [4] [5] [6], but practically it has no telephony related software installed in since it seems a simple OE built with these scripts [7], so (I figure) that you're free to put in both the FSO and/or the Om2008.8 stack, just install the needed packages! Bye. [1] http://download.enlightenment.org/misc/Illume/ [2] http://forum.telefoninux.org/index.php/topic,533.0/topicseen.html [3] http://www.rasterman.com/files/illume-01.ogg [4] http://www.rasterman.com/files/illume-02.ogg [5] http://www.rasterman.com/files/illume-03.ogg [6] http://www.rasterman.com/files/illume-04.ogg [7] http://www.rasterman.com/files/oe-setup.tar.gz I have a question about the predictive keyboard shown in one of those movies - what on earth is the justification for providing the following predictions - User input is on the left, prediction is on the left: w - a wo - so wor - wot worl - work world - ? (I'm not sure what was the prediction, but at this point it's moot... It was probably wrong, anyway.) The qtopia keyboards do the same braindead thing. They're perfectly unusable as a result. The handwriting recognition engine appears to be wired into the same prediction engine, resulting in letters that were recognized correctly to be randomly replaced with incorrect letters. Is this a bug, or is that the intended behavior? --Ori Pessach ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Wed, 1 Oct 2008 20:32:34 -0600, Ori Pessach [EMAIL PROTECTED] wrote: I have a question about the predictive keyboard shown in one of those movies - what on earth is the justification for providing the following predictions - User input is on the left, prediction is on the left: w - a wo - so wor - wot worl - work world - ? (I'm not sure what was the prediction, but at this point it's moot... It was probably wrong, anyway.) The qtopia keyboards do the same braindead thing. They're perfectly unusable as a result. The handwriting recognition engine appears to be wired into the same prediction engine, resulting in letters that were recognized correctly to be randomly replaced with incorrect letters. Is this a bug, or is that the intended behavior? --Ori Pessach Intended, you just have to understand what it's doing. Unlike T9 and other predictive inputs that are familiar from cellphones, the keyboard predictive process looks at neighboring keys, 'predicting' that periodically you may hit the wrong one. (if you're using your finger that's pretty likely, especially if they're as wide as mine) If you pay close attention, the actual user input is helko worlc, and yes it suggests 'hello' and 'world' as the prime suspects. As a developer and hacker I hate it. For text messaging it seems it would be useful, presuming that the user won't always have (or choose to use) a stylus. j ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Wed, Oct 1, 2008 at 8:57 PM, Joel Newkirk [EMAIL PROTECTED] wrote: On Wed, 1 Oct 2008 20:32:34 -0600, Ori Pessach [EMAIL PROTECTED] wrote: I have a question about the predictive keyboard shown in one of those movies - what on earth is the justification for providing the following predictions - User input is on the left, prediction is on the left: w - a wo - so wor - wot worl - work world - ? (I'm not sure what was the prediction, but at this point it's moot... It was probably wrong, anyway.) The qtopia keyboards do the same braindead thing. They're perfectly unusable as a result. The handwriting recognition engine appears to be wired into the same prediction engine, resulting in letters that were recognized correctly to be randomly replaced with incorrect letters. Is this a bug, or is that the intended behavior? --Ori Pessach Intended, you just have to understand what it's doing. Unlike T9 and other predictive inputs that are familiar from cellphones, the keyboard predictive process looks at neighboring keys, 'predicting' that periodically you may hit the wrong one. (if you're using your finger that's pretty likely, especially if they're as wide as mine) If you pay close attention, the actual user input is helko worlc, and yes it suggests 'hello' and 'world' as the prime suspects. As a developer and hacker I hate it. For text messaging it seems it would be useful, presuming that the user won't always have (or choose to use) a stylus. I understand what it's doing. It's not doing it well. I tried it for shell input, and it was an unusable mess. I tried it for text messaging, and it was an unusable mess. It has no model of the likelihood of erroneous input (relatively low) and instead appears to look for the word with the closest minimum edit distance to the user's input. This is nuts. I have never - literally - gotten the word I typed in. In the common use case, of a user who enters a correct word, it invariably get it wrong. Understanding what it's doing doesn't make it less of a nuisance. --Ori Pessach ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Wed, 1 Oct 2008 22:57:10 -0400 Joel Newkirk [EMAIL PROTECTED] babbled: On Wed, 1 Oct 2008 20:32:34 -0600, Ori Pessach [EMAIL PROTECTED] wrote: I have a question about the predictive keyboard shown in one of those movies - what on earth is the justification for providing the following predictions - User input is on the left, prediction is on the left: w - a wo - so wor - wot worl - work world - ? (I'm not sure what was the prediction, but at this point it's moot... It was probably wrong, anyway.) The qtopia keyboards do the same braindead thing. They're perfectly unusable as a result. The handwriting recognition engine appears to be wired into the same prediction engine, resulting in letters that were recognized correctly to be randomly replaced with incorrect letters. Is this a bug, or is that the intended behavior? --Ori Pessach Intended, you just have to understand what it's doing. Unlike T9 and other predictive inputs that are familiar from cellphones, the keyboard predictive process looks at neighboring keys, 'predicting' that periodically you may hit the wrong one. (if you're using your finger that's pretty likely, especially if they're as wide as mine) If you pay close attention, the actual user input is helko worlc, and yes it suggests 'hello' and 'world' as the prime suspects. As a developer and hacker I hate it. For text messaging it seems it would be useful, presuming that the user won't always have (or choose to use) a stylus. bingo. thus the terminal keyboard layout exists - it doesnt correct/predict or suggest. you get what you press, no more, no less, and immediately. so i'm trying to conquer both worlds. i have actually been using the corrective keyboard for testing - the freerunenr has a 2.8 screen. in practice thanks to the bevels on the sides its effective usable area is about 2.5. and on that tiny surface area i can usably type in full english sentences with very few mistakes WHILE walking down the street. thats much better than i can do with a stylus and a non-corrective keyboard on windows ce/mobile/pocketpc, on qtopia's stylus keyboard on my roker e6 and on the matchbox qwerty keyboard etc. etc. for entering sms/email/human stuff its great. if the word is not in your dict - it will be listed always at the top of the long-suggestions list (tap the top-left arrow). it will then be added to your dict. as you use words their Frequency count goes up and is stored. it learns what words you use a lot and tends to offer them as predicted corrections much more accurately once you have been using it for a bit. it is guessing what you meant to type based on the you probably pressed the key you wanted or somewhere near it - maybe a key or 2 away, per letter in the word, so let me search a dictionary and look to see what you likely may have meant. as you type its looking these up and suggesting words. it won't suggest a word other than one in the dictionary file OR one in your personal dictionary (one it's learnt), or EXACTLY what u pressed. depending on how many things match you may get more in that list than can be displayed above the keyboard, so the top-left arrow brings up a full scrollable list of all matches in that case. what the qtopia keyboard does (i can't comment as well as the illume one) and illume do (which was heavily inspired by qtopia's keyboard - and kudos to the tolls for it. of course some things i really dont like about it and i've tried to address them in illume's keyboard - but qtopia's is very good!) is expect you to be entering english text (eg an email, sms, some notes etc.) and correct. both have modes to FORCE a letter to be used (in illume press and hold for 0.25 sec and the zoom box pops up - qtopia has a similar thing). in illume any letters you zoomed in on will not be corrected. they will be taken as explicit selections and thus reduce the search space. illume provides multiple layouts as FILES - so you can extend them and provide more of your own for yoru langauge or personal preferences. the terminal layout just makes the illume keyboard be a dumb stylus keyboard where 1 press on 1 key == exactly that key stroke sent to the app. nothing more or less. -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)[EMAIL PROTECTED] ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Wed, Oct 1, 2008 at 9:11 PM, The Rasterman Carsten Haitzler [EMAIL PROTECTED] wrote: On Wed, 1 Oct 2008 22:57:10 -0400 Joel Newkirk [EMAIL PROTECTED] babbled: On Wed, 1 Oct 2008 20:32:34 -0600, Ori Pessach [EMAIL PROTECTED] wrote: I have a question about the predictive keyboard shown in one of those movies - what on earth is the justification for providing the following predictions - User input is on the left, prediction is on the left: w - a wo - so wor - wot worl - work world - ? (I'm not sure what was the prediction, but at this point it's moot... It was probably wrong, anyway.) The qtopia keyboards do the same braindead thing. They're perfectly unusable as a result. The handwriting recognition engine appears to be wired into the same prediction engine, resulting in letters that were recognized correctly to be randomly replaced with incorrect letters. Is this a bug, or is that the intended behavior? --Ori Pessach Intended, you just have to understand what it's doing. Unlike T9 and other predictive inputs that are familiar from cellphones, the keyboard predictive process looks at neighboring keys, 'predicting' that periodically you may hit the wrong one. (if you're using your finger that's pretty likely, especially if they're as wide as mine) If you pay close attention, the actual user input is helko worlc, and yes it suggests 'hello' and 'world' as the prime suspects. As a developer and hacker I hate it. For text messaging it seems it would be useful, presuming that the user won't always have (or choose to use) a stylus. bingo. thus the terminal keyboard layout exists - it doesnt correct/predict or suggest. you get what you press, no more, no less, and immediately. so i'm trying to conquer both worlds. i have actually been using the corrective keyboard for testing - the freerunenr has a 2.8 screen. in practice thanks to the bevels on the sides its effective usable area is about 2.5. and on that tiny surface area i can usably type in full english sentences with very few mistakes WHILE walking down the street. thats much better than i can do with a stylus and a non-corrective keyboard on windows ce/mobile/pocketpc, on qtopia's stylus keyboard on my roker e6 and on the matchbox qwerty keyboard etc. etc. for entering sms/email/human stuff its great. if the word is not in your dict - it will be listed always at the top of the long-suggestions list (tap the top-left arrow). it will then be added to your dict. as you use words their Frequency count goes up and is stored. it learns what words you use a lot and tends to offer them as predicted corrections much more accurately once you have been using it for a bit. it is guessing what you meant to type based on the you probably pressed the key you wanted or somewhere near it - maybe a key or 2 away, per letter in the word, so let me search a dictionary and look to see what you likely may have meant. as you type its looking these up and suggesting words. it won't suggest a word other than one in the dictionary file OR one in your personal dictionary (one it's learnt), or EXACTLY what u pressed. depending on how many things match you may get more in that list than can be displayed above the keyboard, so the top-left arrow brings up a full scrollable list of all matches in that case. what the qtopia keyboard does (i can't comment as well as the illume one) and illume do (which was heavily inspired by qtopia's keyboard - and kudos to the tolls for it. of course some things i really dont like about it and i've tried to address them in illume's keyboard - but qtopia's is very good!) is expect you to be entering english text (eg an email, sms, some notes etc.) and correct. both have modes to FORCE a letter to be used (in illume press and hold for 0.25 sec and the zoom box pops up - qtopia has a similar thing). in illume any letters you zoomed in on will not be corrected. they will be taken as explicit selections and thus reduce the search space. illume provides multiple layouts as FILES - so you can extend them and provide more of your own for yoru langauge or personal preferences. the terminal layout just makes the illume keyboard be a dumb stylus keyboard where 1 press on 1 key == exactly that key stroke sent to the app. nothing more or less. Making the user compensate for the algorithm's mistakes before it made them doesn't sound very useful. Sounds like it assumes that the user has intimate knowledge of the correction algorithm, which you have, and I don't. Maybe that's why it works for you and drives me absolutely batty. (I'm talking about the mode where the user is expected to force a letter before she has a chance to find out that it's going to be corrected in an unhelpful way.) --Ori Pessach ___ Openmoko community mailing list
Re: New Rasterman Image...
On Wed, 1 Oct 2008 21:05:53 -0600 Ori Pessach [EMAIL PROTECTED] babbled: I understand what it's doing. It's not doing it well. I tried it for shell i disagree. it works like a charm for me - as per my previous mail - i can use it while walking down the street. more than i can say for pretty much any other virtual keyboard i have available to me. input, and it was an unusable mess. I tried it for text messaging, and it why someone would use a language dictionary-based corrective keyboard for shell input beats me! in this case i call silly user - using a motorcycle to deliver elephants line :) use the terminal keyboard. use a stylus. thats what it was meant for. :) was an unusable mess. It has no model of the likelihood of erroneous input it does. it absolutely does. maybe your fingers are incredibly off-center? here is the algorithm (and if u don't believe me - code is there to be read): it stores a press POINT (x,y). it looks for all keys whose center point is WITHIN f distance of x,y (f being the fuzz value - the .kbd file for the qwerty Default keyboard is 135 units wide, with fuzz radius of 20, so that's about 1/3rd of the keyboard that it searches through for a likely match). likelihood factors (distance) per key found is allocated based on distance (0 == most likely, 0 less likely the greater the value). each press is done this way EXCEPT if u hold for 0.25 sec then drag to select a key explicitly in zoom mode - then the ONLY key available for that word slot is that letter selected given a distance of 0. as you type all permutations of letters are searched and put into a list - with each permutation given a distance metric based on the letters used (simply addition of the distances). now this is combined with the dictionary's frequency metric (multiplied by an inverse) so the more likely the word is to be used the lower its distance becomes. words are sorted from most to least likely based on this metric then listed with most likely in the middle of the list, leas likely to the left/right ends - which you may not see. the vertical list lists all matches from most to least likely (top to bottom) with 1 exception - EXACTLY what u typed it as the top. it absolutely has a fairly good idea of likelihood of error and likelihood of usage of a word etc. etc. eg: Press | Guess+dist e e+0 w+1 r+2 d+2 s+1 r r+0 t+1 e+2 f+1 g+2 d+3 k k+0 l+1 o+2 i+3 j+3 d d+0 f+1 s+1 e+1 c+1 r+2 w+2 so erkd has distance 0 = but its not a word in the dictionary at all, so thrown out. rwkd has distance 1, but not a word, srkd same, etkd, efkd, erld, etc. etc. in the end it produces a list where most likely world ends up the word with other options too - and this is a much simplified list. mostly the list for candidate letters per input letter is about 10-12 letters. so u have 12*12*12*12 permutations for a 4 letter word - of which a fraction of that space is legitimate words. each permutation has a likelihood value based on press distance and on frequency of usage of that word in language in general in the dictionary. mind you - i AM talking about illume's keyboard, its algorithms as is in the image i built. if you use something else i cannot comment as it's something else. (relatively low) and instead appears to look for the word with the closest minimum edit distance to the user's input. This is nuts. I have never - it's not - as the edit distance is the likelihood of error. you likely press the key you want - or near it. thus keys near where you pressed are more likely than those further away. to limit search distance only up to a certain distance is searched. chances are that you do this: fingerprint: ___ /~~~\ |~~~| |~~~| \x/ where x is the pressure point reported on the touchscreen. the only info the touchscreen reports is the pressure point - nothing else. you think u press somewhere else, but don't. you know what u pressed bu what key pops up that lets u know pretty well how good your pressing of the screen is. this is just a hardware limit of a resistive touchscreen. the point of greatest pressure is used - not the middle point of the area in which skin contacts the screen. get the gpe-sketchbook and try press with the flat of your finger and see just of far off your press point is. it may surprise you. as i said - it does have all the model and code and even data to do proper correction based on many factors. i do NOT have a dictionary with frequency info for all of english - there is a small english dict (5000 words) with some frequency info in it i managed to gather, but its very small. if you don't believe me - read the code, or do better. patches accepted, but i think the problem is just that the dictionary has no frequency info by default (a matter of simple lack of data) or how you press the screen. i suggest you pay close attention to how you type and see. yes the black word (in the black box) may not be always the word u want - but its most often that
Re: New Rasterman Image...
On Wed, 1 Oct 2008 21:35:40 -0600 Ori Pessach [EMAIL PROTECTED] babbled: Making the user compensate for the algorithm's mistakes before it made them doesn't sound very useful. Sounds like it assumes that the user has intimate knowledge of the correction algorithm, which you have, and I don't. Maybe that's why it works for you and drives me absolutely batty. (I'm talking about the mode where the user is expected to force a letter before she has a chance to find out that it's going to be corrected in an unhelpful way.) it's impossible to know. how do u know what the word int he end is going to be - if its something you haven't got in your dictionary and can't match to? that is why the full match list exists with exactly what you typed. then it goes into the dict and can be matched and as you use it more is more likely to be matched. if u are on an even smaller screen or on a very bumpy bus ride you can use hold+drag (zoom mode) to carefully enter a specific letter or word. if the word is not in the dict - this is a way to enter it. you will know its not there when its not offered as a correction in the short or long list. you get to delete what u typed and type again carefully. you only need to do this once. then it goes into the dict. btw - you DO know u can PRESS on the list of matches on the top - not just stroke right to accept the default most likely one? and you do know of the full match list? (top-left)? it doesnt assume intimate knowledge of the algorithm btw. you assume that it needs it - if u keep pressing the wrong keys hold+drag+zoom is a way to fix it. if u have a stylus and u are standing still the correction stuff is useless anyway - so u just don't use it. just press away with stylus as if it were any normal stylus keyboard (as u will always enter the correct word and thus it will always be exactly what you typed). seriously - as an experiment, when you type a word and it gets the guess wrong pop up the full list and at the top is EXACTLY what you typed. look at it then go wow! it managed to correct THAT into the list of following words! wow! it's astounding how it can turn a mangled mess into a legible word. it works! -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)[EMAIL PROTECTED] ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Thu, 2 Oct 2008 13:11:21 +1000, Carsten Haitzler (The Rasterman) [EMAIL PROTECTED] wrote: for entering sms/email/human stuff its great. if the word is not in your dict - it will be listed always at the top of the long-suggestions list (tap the top-left arrow). it will then be added to your dict. as you use words their Frequency count goes up and is stored. it learns what words you use a lot and tends to offer them as predicted corrections much more accurately once you have been using it for a bit. Can you offer the ability to weight the personal dictionary? IE, either on write or on read, scale the frequency by a configurable factor? Force any entries there to come up first regardless of frequency? And for that matter, how does it handle the case where the built-in dict has no frequency data, but the user dict does? (I presume this is the 'default' situation) j ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community
Re: New Rasterman Image...
On Thu, 2 Oct 2008 00:19:50 -0400 Joel Newkirk [EMAIL PROTECTED] babbled: On Thu, 2 Oct 2008 13:11:21 +1000, Carsten Haitzler (The Rasterman) [EMAIL PROTECTED] wrote: for entering sms/email/human stuff its great. if the word is not in your dict - it will be listed always at the top of the long-suggestions list (tap the top-left arrow). it will then be added to your dict. as you use words their Frequency count goes up and is stored. it learns what words you use a lot and tends to offer them as predicted corrections much more accurately once you have been using it for a bit. Can you offer the ability to weight the personal dictionary? IE, either on write or on read, scale the frequency by a configurable factor? Force any entries there to come up first regardless of frequency? And for that matter, how does it handle the case where the built-in dict has no frequency data, but the user dict does? (I presume this is the 'default' situation) the personal dict has frequency in it - and it always takes preference. any word you use will go into the personal dict as its weight will have changed (the system dict has no weights so all entries are assumed 1). so from then on personal dict is king. each word has a weight. words with higher weights are more likely to match - but that also depends how far off your typing is too. in generally you will find that the more you use a word the more it is either the default correction choice or that it is one of the very top candidates and is instantly visible (if its short enough). generally the longer the word the fewer dict candidates so it tends to get it right more reliably anyway. just try it! be patient - teach it things it doesn't know (names and proper names may not all be there as well as slang and shortenings like 2day and 4ever sms-speak), and as you use it and select the corrected word you want it will get much better at guessing your habits. it definitely does. i've tried it out a lot! -- - Codito, ergo sum - I code, therefore I am -- The Rasterman (Carsten Haitzler)[EMAIL PROTECTED] ___ Openmoko community mailing list community@lists.openmoko.org http://lists.openmoko.org/mailman/listinfo/community