User:Nicola.mfb
From Openmoko
Nicola.mfb (Talk | contribs) |
Nicola.mfb (Talk | contribs) m |
||
| Line 3: | Line 3: | ||
I'm Nicola, born on 1973 in Paduli a little country of southern Italy. | I'm Nicola, born on 1973 in Paduli a little country of southern Italy. | ||
| − | I'm | + | I'm graduated in Informatic science, and since 1997 I'm working as system manager for the University Federico II of Naples. |
When I arrived I found about 40 servers, equipped with some proprietary Unix (and not Unix) OS, time after time I migrated all them to Linux, today about 400 servers are free, and 99% of them have community distro. | When I arrived I found about 40 servers, equipped with some proprietary Unix (and not Unix) OS, time after time I migrated all them to Linux, today about 400 servers are free, and 99% of them have community distro. | ||
| − | My mission is Linux everywhere, I use it on all my desktop, servers, wireless AP, ipaq, and finally on my phone too. | + | My mission is Linux everywhere, I use it on all my desktop, servers, wireless AP, hp ipaq, and finally on my phone too. |
| − | While not struggling over all my devices I play accordion, flute and piano in an ethnic | + | While not struggling over all my devices I play accordion, flute and piano in an ethnic popular music band, or take photos around the world. |
| − | On this page you may see what I'm thinking about the freerunner and what I'm doing for | + | On this page you may see what I'm thinking about the freerunner and what I'm doing for it!. |
=Distros= | =Distros= | ||
==2007.12== | ==2007.12== | ||
| − | I do not know it, when my freerunner arrived on september 2008 it was already deprecated so I flashed ASU after a couple of minutes. I did an error, | + | I do not know it, when my freerunner arrived on september 2008 it was already deprecated so I flashed ASU after a couple of minutes. I did an error, I lost the knowledge of some important features and ideas I was able to see only some month ago with Hackable:1, and that are inspiring me now. |
==Qtopia== | ==Qtopia== | ||
| − | Qtopia was the first distro I installed on the SD, it was FAST and complete, but | + | Qtopia was the first distro I installed on the SD, it was FAST and complete, but cannot run X11 applications. |
==ASU== | ==ASU== | ||
| − | A mix between E, QT and GTK, the idea was to provide a subset of qtopia on the X server | + | A mix between E, QT and GTK, the idea was to provide a subset of the mature qtopia stack on the X server together with all other X11 apps. So I swapped and put Qtopia on the flash and ASU on the SD card to have more space to play with it. |
==FSO== | ==FSO== | ||
| − | "All development should go to FSO, ASU is deprecated" was the rule, and as I was excited to write my apps for my phone I deleted Qtopia, put ASU on flash and FSO on SD. | + | "All development should go to FSO, ASU is deprecated" was the rule, and as I was excited to write my apps for my phone with simple dbus calls so I deleted Qtopia, put ASU on the flash and FSO on the SD and resumed my OE knowledge. |
| − | There is not much to say on FSO images, | + | There is not much to say on FSO images, it's not a really distro, but a way to carry and use the FSO frameworkd. |
==SHR== | ==SHR== | ||
| − | In the mean while, SHR | + | In the mean while, SHR was born to recycle the 2007.12 code porting it to FSO, it was growing and now it's quite all based on E* software. It's a very promising distro, but a lot of users are annoyed because it frequently breaks. The release of a stable milestone was announced several months ago, but we are still waiting, developers sometime freeze the unstable branch to a testing snapshot tree calling user to test and report feedbacks, and after a bit they delete it and freezes again. This happened a lot of time while peoples are asking for a stable distro. |
| − | SHR application are based on python, so you actually have python apps, interacting with a python ophonekitd, interacting with a python frameworkd. The second problem is that flash/sd is slow, and phone functionalities are splitted between several apps, the load time of each of them is high, the result is a very slow device. Please to not blame me! I respect the | + | SHR application are based on python, so you actually have python apps, interacting with a python ophonekitd, interacting with a python frameworkd. The second problem is that flash/sd is slow, and phone functionalities are splitted between several apps, the load time of each of them is high, the result is a very slow device. Please to not blame me! I respect the effort of SHR developers, but think that some decision must be revisited above all when we may see in the past better solutions! |
=OM2009= | =OM2009= | ||
| − | OM2009 was initially the new offical OM distro, as OM stopped developing actually it's based on volounter work. It has two strong points over SHR, the first is that it uses more legacy software to try effectively a stabilization for release, an example is the kernel as the last have serious problems on wifi, the second is that the main application, Paroli, is a semi monolitic python app, so the python load time is limited only on startup, and the big part of functionality are ready to use because they are in ram or is loaded | + | OM2009 was initially the new offical OM distro, as OM stopped developing actually it's based on volounter work. It has two strong points over SHR, the first is that it uses more legacy software to try effectively a stabilization for release, an example is the kernel as the last have serious problems on wifi, the second is that the main application, Paroli, is a semi monolitic python app, so the python load time is limited only on startup, and the big part of functionality are ready to use because they are in ram or is loaded easyly with plugins. |
=Gentoo= | =Gentoo= | ||
| − | My preferred distro on laptop and servers, but it's too slow on the device, the lack of precompiled binaries and freerunner customization make it usable only after some weeks of emerging/ | + | My preferred distro on laptop and servers, but it's too slow on the device, the lack of precompiled binaries and freerunner customization make it usable only after some weeks of emerging/twicking. |
=Hackable:1= | =Hackable:1= | ||
| − | A very interesting distro, it uses hardware directly, stable underliyng system (debian lenny), and | + | A very interesting distro, it uses hardware directly, stable underliyng system (debian lenny), and revisited and improved 2007.12 code. No python in core parts, so very it's fast, it has some important features and ideas that I think it'absourd we losed in ASU/SHR/OM2009, for example the possibility to launch a stay on top window by holding the aux menu to kill applications and switch fullscreen windows! |
Try actually to close a fullscreen application on distro that use aux to lock and power to suspend! | Try actually to close a fullscreen application on distro that use aux to lock and power to suspend! | ||
Luckily SHR has restored the use of power button to launch a dialog that make you kill applications or poweroff the system. | Luckily SHR has restored the use of power button to launch a dialog that make you kill applications or poweroff the system. | ||
| − | The only problem of hackable is that it uses hardware and not FSO, so many applications FSO based does not work, developers has to drive the hardware without a central mediator | + | The only problem of hackable is that it uses hardware and not FSO, so many applications FSO based does not work, developers has to drive the hardware without a central mediator. For example if two applications power on the BT at startup and power off it at exit, you cannot close the first and use the second, while the FSO resource system handles that for you. And the last, it uses an old 2.6.24 kernel, but actually it seems more stable than recents! |
=Debian SID= | =Debian SID= | ||
My actually preferred distro, has quite recent kernel and fso, tons of precompiled ready to use software, the best distro to try and test software and to develop new ones! One only problem, it's not targeted for freerunner (of course) so you have to customize it, and it upgrades slowly, but this may be not a defect :) | My actually preferred distro, has quite recent kernel and fso, tons of precompiled ready to use software, the best distro to try and test software and to develop new ones! One only problem, it's not targeted for freerunner (of course) so you have to customize it, and it upgrades slowly, but this may be not a defect :) | ||
| Line 38: | Line 38: | ||
Not tested yet | Not tested yet | ||
=A stable device?= | =A stable device?= | ||
| − | After one year I may say my freerunner is *not* a stable platform, I | + | After one year I may say my freerunner is *not* a stable platform, I encountered all the problem you may read on the list archives: |
* SD partition corruption | * SD partition corruption | ||
* WSOD | * WSOD | ||
| Line 48: | Line 48: | ||
* a lot of kernel ops, device freezing | * a lot of kernel ops, device freezing | ||
| − | So the real problem is that the kernel | + | So the real problem is that the kernel is not rock solid. Some chip specification are under NDA so only few people may contribute, after OM suspended the kernel development we cannot say if it will be really improved in the future. |
I really hope that the volounters maintaining the kernel will grow in number and find the time to fix it, only with a rock solid kernel we may think to use the device seriously. | I really hope that the volounters maintaining the kernel will grow in number and find the time to fix it, only with a rock solid kernel we may think to use the device seriously. | ||
| Line 60: | Line 60: | ||
Peoples are not working on them for a simple reason, they are working on "basic" software, as it's not complete. | Peoples are not working on them for a simple reason, they are working on "basic" software, as it's not complete. | ||
=What I'm doing= | =What I'm doing= | ||
| − | As I think the better way to solve | + | As I think the better way to solve problems is to try fix them yourself instead of boring peoples, in the spare time I'm coding some apps I feel missing on the freerunner. The problem is that I have few time, documentation in the free software is poor, and above all I'm an old C/C++ coder with some QT extension, and this toolkit seems to be a forbidden topic in the freerunner world, the consequence is that there are few people that may collaborate with me, the above of them prefers to spend time on qtmoko. |
Revision as of 03:43, 28 June 2009
Hi
I'm Nicola, born on 1973 in Paduli a little country of southern Italy.
I'm graduated in Informatic science, and since 1997 I'm working as system manager for the University Federico II of Naples. When I arrived I found about 40 servers, equipped with some proprietary Unix (and not Unix) OS, time after time I migrated all them to Linux, today about 400 servers are free, and 99% of them have community distro.
My mission is Linux everywhere, I use it on all my desktop, servers, wireless AP, hp ipaq, and finally on my phone too. While not struggling over all my devices I play accordion, flute and piano in an ethnic popular music band, or take photos around the world.
On this page you may see what I'm thinking about the freerunner and what I'm doing for it!.
Contents |
Distros
2007.12
I do not know it, when my freerunner arrived on september 2008 it was already deprecated so I flashed ASU after a couple of minutes. I did an error, I lost the knowledge of some important features and ideas I was able to see only some month ago with Hackable:1, and that are inspiring me now.
Qtopia
Qtopia was the first distro I installed on the SD, it was FAST and complete, but cannot run X11 applications.
ASU
A mix between E, QT and GTK, the idea was to provide a subset of the mature qtopia stack on the X server together with all other X11 apps. So I swapped and put Qtopia on the flash and ASU on the SD card to have more space to play with it.
FSO
"All development should go to FSO, ASU is deprecated" was the rule, and as I was excited to write my apps for my phone with simple dbus calls so I deleted Qtopia, put ASU on the flash and FSO on the SD and resumed my OE knowledge. There is not much to say on FSO images, it's not a really distro, but a way to carry and use the FSO frameworkd.
SHR
In the mean while, SHR was born to recycle the 2007.12 code porting it to FSO, it was growing and now it's quite all based on E* software. It's a very promising distro, but a lot of users are annoyed because it frequently breaks. The release of a stable milestone was announced several months ago, but we are still waiting, developers sometime freeze the unstable branch to a testing snapshot tree calling user to test and report feedbacks, and after a bit they delete it and freezes again. This happened a lot of time while peoples are asking for a stable distro. SHR application are based on python, so you actually have python apps, interacting with a python ophonekitd, interacting with a python frameworkd. The second problem is that flash/sd is slow, and phone functionalities are splitted between several apps, the load time of each of them is high, the result is a very slow device. Please to not blame me! I respect the effort of SHR developers, but think that some decision must be revisited above all when we may see in the past better solutions!
OM2009
OM2009 was initially the new offical OM distro, as OM stopped developing actually it's based on volounter work. It has two strong points over SHR, the first is that it uses more legacy software to try effectively a stabilization for release, an example is the kernel as the last have serious problems on wifi, the second is that the main application, Paroli, is a semi monolitic python app, so the python load time is limited only on startup, and the big part of functionality are ready to use because they are in ram or is loaded easyly with plugins.
Gentoo
My preferred distro on laptop and servers, but it's too slow on the device, the lack of precompiled binaries and freerunner customization make it usable only after some weeks of emerging/twicking.
Hackable:1
A very interesting distro, it uses hardware directly, stable underliyng system (debian lenny), and revisited and improved 2007.12 code. No python in core parts, so very it's fast, it has some important features and ideas that I think it'absourd we losed in ASU/SHR/OM2009, for example the possibility to launch a stay on top window by holding the aux menu to kill applications and switch fullscreen windows! Try actually to close a fullscreen application on distro that use aux to lock and power to suspend! Luckily SHR has restored the use of power button to launch a dialog that make you kill applications or poweroff the system. The only problem of hackable is that it uses hardware and not FSO, so many applications FSO based does not work, developers has to drive the hardware without a central mediator. For example if two applications power on the BT at startup and power off it at exit, you cannot close the first and use the second, while the FSO resource system handles that for you. And the last, it uses an old 2.6.24 kernel, but actually it seems more stable than recents!
Debian SID
My actually preferred distro, has quite recent kernel and fso, tons of precompiled ready to use software, the best distro to try and test software and to develop new ones! One only problem, it's not targeted for freerunner (of course) so you have to customize it, and it upgrades slowly, but this may be not a defect :)
Neovento
Not tested yet
A stable device?
After one year I may say my freerunner is *not* a stable platform, I encountered all the problem you may read on the list archives:
- SD partition corruption
- WSOD
- NSOD
- no suspend/resume
- interferences with 8GB micro sdhc with glamo errors on GSM
- Bt does not works after resume on 2.6.28
- wifi really unstable on 2.6.29
- a lot of kernel ops, device freezing
So the real problem is that the kernel is not rock solid. Some chip specification are under NDA so only few people may contribute, after OM suspended the kernel development we cannot say if it will be really improved in the future.
I really hope that the volounters maintaining the kernel will grow in number and find the time to fix it, only with a rock solid kernel we may think to use the device seriously.
A fast device?
The freerunner is slow:
- python abused everywhere
- xglamo is accelerated but cpu has to busy wait after requesting accelerated ops
- software functionality are splitted in too many apps
Software status
Some basic tasks are impossible to do with graphical UI, for example connect to WPA enterprise networks, transfer obex file with bluetooth. Some phone functionalities are missing, see for example MMS. Peoples are not working on them for a simple reason, they are working on "basic" software, as it's not complete.
What I'm doing
As I think the better way to solve problems is to try fix them yourself instead of boring peoples, in the spare time I'm coding some apps I feel missing on the freerunner. The problem is that I have few time, documentation in the free software is poor, and above all I'm an old C/C++ coder with some QT extension, and this toolkit seems to be a forbidden topic in the freerunner world, the consequence is that there are few people that may collaborate with me, the above of them prefers to spend time on qtmoko.
