OLPC:News: Difference between revisions

From OLPC
Jump to navigation Jump to search
Line 7: Line 7:
=Laptop News 2008-01-05=
=Laptop News 2008-01-05=


1. Intel: John Markoff’s article in Saturday’s '''New York Times''' provides an accurate description of the events of the past 48 hours regarding Intel (See http://www.nytimes.com/2008/01/05/technology/05laptop.html). We made a sincere effort of rapprochement, but it was clear from even the way that Intel terminated the relationship—with an “inadvertent leak”—that their was no reciprocal sincerity. We made great strides before Intel joined us and we will continue to make great strides now that they have left the OLPC association.
1. Intel: John Markoff’s article in Saturday’s ''New York Times'' provides an accurate description of the events of the past 48 hours regarding Intel (See http://www.nytimes.com/2008/01/05/technology/05laptop.html). We made a sincere effort of rapprochement, but it was clear from even the way that Intel terminated the relationship—with an “inadvertent leak”—that their was no reciprocal sincerity. We made great strides before Intel joined us and we will continue to make great strides now that they have left the OLPC association.


See also two recent '''Wall Street Journal''' articles (access to the full 2nd article requires a subscription until two weeks have elapsed from date of publishing): http://online.wsj.com/article/SB119948070480568405.html and http://online.wsj.com/article/SB119940537839566305.html.
See also two recent ''Wall Street Journal'' articles (access to the full 2nd article requires a subscription until two weeks have elapsed from date of publishing): http://online.wsj.com/article/SB119948070480568405.html and http://online.wsj.com/article/SB119940537839566305.html.


2. Lagos: Ayo Kusamotu filed a preliminary objection to the Nigerian keyboard lawsuit (See [[Lancor]]). The details of the case have been discussed extensively on Groklaw (See http://www.groklaw.net/article.php?story=20071226210020415).
2. Lagos: Ayo Kusamotu filed a preliminary objection to the Nigerian keyboard lawsuit (See [[Lancor]]). The details of the case have been discussed extensively on Groklaw (See http://www.groklaw.net/article.php?story=20071226210020415).

Revision as of 17:03, 6 January 2008

  This page is monitored by the OLPC team.
   HowTo [ID# 92927]  +/-  

You can subscribe to the OLPC community-news mailing list by visiting the laptop.org mailman site.

Laptop News 2008-01-05

1. Intel: John Markoff’s article in Saturday’s New York Times provides an accurate description of the events of the past 48 hours regarding Intel (See http://www.nytimes.com/2008/01/05/technology/05laptop.html). We made a sincere effort of rapprochement, but it was clear from even the way that Intel terminated the relationship—with an “inadvertent leak”—that their was no reciprocal sincerity. We made great strides before Intel joined us and we will continue to make great strides now that they have left the OLPC association.

See also two recent Wall Street Journal articles (access to the full 2nd article requires a subscription until two weeks have elapsed from date of publishing): http://online.wsj.com/article/SB119948070480568405.html and http://online.wsj.com/article/SB119940537839566305.html.

2. Lagos: Ayo Kusamotu filed a preliminary objection to the Nigerian keyboard lawsuit (See Lancor). The details of the case have been discussed extensively on Groklaw (See http://www.groklaw.net/article.php?story=20071226210020415).

3. School Server: A long delayed update to the School Server software to fix problems with large file transfers due to a now antiquated libertas driver has finally happened. Pick up Build 141 (XS_Installing_Software#OLPC_XS_141), request an Active Antenna build through our developer program, and turn some old junker PC near you into a school server! This was delayed by the holidays and a QA process which turned out to be as necessary as it was difficult. Several problems and unwanted interactions had crept into the build, but these were mostly fixed in this release. (John Watlington finally gave up getting an early version of the school server software to work properly with the Active Antenna he had for testing the upgrade process. After upgrading to Build 141 (stable) it worked fine.)

We now have a jabber server running on the schoolserver in Cambridge (See Ejabberd Configuration), and are starting to test against it. We have seen the “register” button work! This should reach a school server build by the beginning of next week.

4. Open firmware (OFW): Mitch Bradley made a number of improvements this week. It now reports the OFW version when in a failure condition to simplify field support; the “remove all power” error message has been made clearer; and the status of the game buttons is indicated when pressed.

5. Embedded controller (EC): Richard Smith reports that several EC issues are pending. The one that has received attention the SCI Mask Corruption bug (Ticket #5467). While chasing this bug, Richard found several small but critical typos in the handling of some of the commands he had added since November. The net result of these typos is that under some conditions, a value passed in as data would be run as a command and some commands would not get run at all. Unfortunately, fixing those typos does not fix #5467. Its cause goes much deeper into the EC protocol handling. The next couple of days should shake out what the problem is and get it fixed for good.

Battery problems: A growing number of reports of short battery life are coming in. People are starting to run olpc-logbat bat and Richard has been looking at the resulting data. Based on the data he's seen so far, he conjects that either (1) there are some “funky” batteries in the field; or (2) the EC is failing to charge the battery up to full capacity, yet it is marking it as full. Most of the data gathered so far has been discharge info. Richard will be responding to many of the trouble tickets requesting several cycles of charge/discharge while running olpc-logbat to flush out whats going on.

The report of “shutdown yet no red LED” is the result of the capacity never going below 15% but the battery voltage dropped to the critical cutoff point, followed by the EC dropping system power. An enhancement Richard will make to the EC code is to also do something with the status LED to indicted that a critical voltage shutdown is looming so there's some warning your laptop is about to shutdown.

6. Schedules: For the next few weeks we will be focused on stabilizing Update.1 (based on Joyride) through testing, documentation, and limited number of bug fixes. We recently found two more critical bugs that will need an unscheduled software release (USR): touchpad/mouse jumpiness and data loss if you fill up the memory. We have created a procedure for these USRs; we are using this process to get these fixes out sooner than the next scheduled release (See Operating system release procedures).

7. Test: Chih-Yu Chao is helping out with both test and support this week. She has gotten through the 1-Hour Smoke Test on a recent Joyride build, which revealed a few regression bugs from the Ship.2 (650) release. Next we need to create and document some test cases for the new features in Update.1, and some testing with the school server.

8. Support: Adam Holt has been working days, nights, and weekends to grow the volunteer support group (now up to 40 people), who are answering emails, hanging out on forums and IRC to help people. A core crew of about 15 volunteers drives the process forward with increasingly sophisticated answers. Others contribute part-time working from the Support FAQ (Support FAQ) and “RTFM” template answers as they get up to speed. We have almost hit 1000 email help requests in the database! Katie Belisle and Casey Ratliff are working on a next-generation documentation ideas for our Support FAQ.

Adam is also coordinating “4PM Sunday” (Eastern Standard Time) conference calls for the entire support-volunteers team. Last week’s call was extremely successful due to the contributions of the OLPC developer community (special thanks to Bernie Innocenti and Arjun Sarwal) (See Support meetings). Anyone you wants to join, email me well in advance at “holt AT laptop DOT org” and be sure to include your phone number! All volunteers worldwide will be considered, after a very brief phone call. (See http://lists.laptop.org/listinfo/support-gang if you are interested in volunteering.)

Adam, Kim Quirk, and Greg Babbin are now able to provide RMAs, which will help off-load the donor support 800-number and email. Kudos to Greg’s genuine heroics. Our Asterisk-based VoIP virtual call center has been briefly delayed. Matthew O’Gorman and Joe Phigan continue to work hard on this, scripting prompts for rudimentary integration with http://rt.laptop.org, and we should have our volunteer-training shortly.

9. Localization: Sayamindu Dasgupta helped start Pashto and Bulgarian translation teams and resolve an issue with Pootle that caused it to reject new user registrations; it was being caused by a invalid username in the Pootle database.

A long-standing blocker bug (Ticket #1525) regarding the invalidation of the fontconfig cache was finally fixed. Font cache generation in the XO should be more robust now, even in the face of clock failures.

10. Software potpourri: Tomeu Vizoso proposed fixes for a number of bugs that have highest priority: previews are not deleted when their matching datastore objects are removed (Ticket #5707); deleting a large file from a USB stick copies it into NAND (often filling NAND) (Ticket #5719); Sugar shell consuming vast amounts of memory (Ticket #5532); “resuming” a large file from USB copies it into NAND (filling NAND) (Ticket #5744); and objects accumulate on the clipboard, impacting system performance (Ticket #5760).

Chris Ball fixed many power manager bugs. We now perform power management regardless of whether you're on an external power source, we remember the user's previous brightness setting when we dim the screen during suspend, and open hardware manager (OHM) no longer exits when X does.

Chris Ball wrote “slideshow” over Christmas, which is a Pippy example that queries the datastore for camera images and then shows them full-screen in a slideshow. He can't decide whether it should be a Pippy example (since it demonstrates performing datastore searches in Python) or a separate activity.

Dennis Gilmore spent most of the week troubleshooting issues, working around an issue today causing build failures and mostly trying to put the pieces together to make things better.

Michael Stone and Dennis spent some time working out why iputils fell out of our builds. Michael also worked with Bernie and Tomeu on address a problem with olpc-update in regards to persistent activity directories (Ticket #5033), with Ben Schwartz on problem with stream sockets (Ticket #5818), and with Eben Eliason on the beginnings of a security user interface.

Ivan Krstić is exploring a more secure way of isolating Browse for Update.1; it might be trivial.

11. Presence: Robert McQueen finished an out-of-band data (OOB) implementation (he added IP detection code) and wrote tests for it. That means OOB bytestream is now working with Gabble. The next step will be to define and implement bytestream renegotiation and fallback.

Dafydd Harries made updated packages for Presence Service and Avahi, although Koji cannot build the former for some reason. He also debugged problems registering laptops with school servers (Ticket #5834); it turns out that the ejabberd RPM doesn't generate an X.509 certificate. Dafydd also spent time trying to coax OpenFire into working. It works ok as long as your account is not in the shared roster group, but authenticating becomes unreliable as soon as you are a member. The web interface becomes unreliable from time to time too, necessitating restarting the server. It seems that, like with ejabberd, we are using it in a way it is not designed to handle. Our scalability improvements should solve this for Update.2, but it is not clear yet what the best approach is for the Update.1 time frame.

Morgan Collett fixed the scrolling bug (Ticket #2351) in Chat for Update.1 thanks to a patch from Marco Pesenti Gritti. (C. Scott Ananian’s view source changes for Chat are in Update.1, but will require a newer Pippy.) Morgan is testing a fix for Presence Service #5368 where the buddies in an activity weren't reliably clustering around the shared activity icon.

12. Activities: Muriel Godoi progressing on his port of Food Force for the XO (See Food Force 2). Progress includes artwork (builds and villagers); next, the game model need more work to get a playable game. The code is in his public_git folder (https://dev.laptop.org/git?p=users/murielgodoi/foodforce2;a=summary). Please contact Muriel if you’d like to help.

Arjun Sarwal reports progress on the Measure activity: he is rethinking certain aspects of the code design of the activity that would make it more easily expandable and scalable.

13. OLPC Pakistan: Dr. Habib Khan reports progress amidst chaos. Urdu localization is 100% complete. They have had a useful discussion with an Afghan graduate student of International Islamic University (IIU) who is keenly interested in translating OLPC bundles into Dari and Pashto. They are also mobilizing volunteers from the Pakistan Institute of Engineering & Applied Sciences, (PIEAS). A training package for Afghan teachers is nearing completion, including software, hardware, and activities tutorials. They’ve also launched an effort to convert into e-books all the text books written on curriculum of the Federal Ministry of Education, Islamabad. Beta versions for Grades one through four are ready. The subjects are English, Social Studies, Science, and Urdu.

14. Cow power: Arjun have completed documentation of the project (See Cow Power). The page details the current design and the proposed mechanical design. He is hoping to get feedback from the community on the proposed mechanical design before moving forward in the implementation of the changes.

15. Community: The OLPC Austria team reports progress on OpenWRT. It boots an XO (currently with minimal driver support) in 15–20 seconds. John Crispin and others want to look into porting Sugar to OpenWRT if there is community interest.

Pascal Martin of Linterweb, an open source software company based in France that has worked on desktop and wiki search tools, has offered their support and development time to help with the search component for the Journal. Tomeu Vizoso spent some time explaining to Fabien Coulon from their team what has been done to date in the datastore.

Jesper Taxbøl is helping organize this year's Nordic Game Jam; he is angling to run it on XOs and lead off with an introduction to PyGame. They are looking for 10 laptops for their 100 participants to use February 1–3. This is quite a popular jam and produces some pretty polished games each year.

Many people are asking for ways to contact the creators of bundles and activities. Please add your name and some sort of contact info to projects you have worked on, on their own wiki pages, and on the Activities page.

More News

Laptop News is archived here.

You can subscribe to the OLPC community-news mailing list by visiting the laptop.org mailman site.

Press requests: please send email to press@racepointgroup.com

Milestones

Latest milestones:

Nov. 2007 Mass Production has started.
July. 2007 One Laptop per Child Announces Final Beta Version of its Revolutionary XO Laptop.
Apr. 2007 First pre-B3 machines built.
Mar. 2007 First mesh network deployment.
Feb. 2007 B2-test machines become available and are shipped to developers and the launch countries.
Jan. 2007 Rwanda announced its participation in the project.

All milestones can be found here.


Press

You can subscribe to the OLPC community-news mailing list by visiting the laptop.org mailman site.

  This page is monitored by the OLPC team.
   HowTo [ID# 92927]  +/-  

You can subscribe to the OLPC community-news mailing list by visiting the laptop.org mailman site.

Laptop News 2008-01-05

1. Intel: John Markoff’s article in Saturday’s New York Times provides an accurate description of the events of the past 48 hours regarding Intel (See http://www.nytimes.com/2008/01/05/technology/05laptop.html). We made a sincere effort of rapprochement, but it was clear from even the way that Intel terminated the relationship—with an “inadvertent leak”—that their was no reciprocal sincerity. We made great strides before Intel joined us and we will continue to make great strides now that they have left the OLPC association.

See also two recent Wall Street Journal articles (access to the full 2nd article requires a subscription until two weeks have elapsed from date of publishing): http://online.wsj.com/article/SB119948070480568405.html and http://online.wsj.com/article/SB119940537839566305.html.

2. Lagos: Ayo Kusamotu filed a preliminary objection to the Nigerian keyboard lawsuit (See Lancor). The details of the case have been discussed extensively on Groklaw (See http://www.groklaw.net/article.php?story=20071226210020415).

3. School Server: A long delayed update to the School Server software to fix problems with large file transfers due to a now antiquated libertas driver has finally happened. Pick up Build 141 (XS_Installing_Software#OLPC_XS_141), request an Active Antenna build through our developer program, and turn some old junker PC near you into a school server! This was delayed by the holidays and a QA process which turned out to be as necessary as it was difficult. Several problems and unwanted interactions had crept into the build, but these were mostly fixed in this release. (John Watlington finally gave up getting an early version of the school server software to work properly with the Active Antenna he had for testing the upgrade process. After upgrading to Build 141 (stable) it worked fine.)

We now have a jabber server running on the schoolserver in Cambridge (See Ejabberd Configuration), and are starting to test against it. We have seen the “register” button work! This should reach a school server build by the beginning of next week.

4. Open firmware (OFW): Mitch Bradley made a number of improvements this week. It now reports the OFW version when in a failure condition to simplify field support; the “remove all power” error message has been made clearer; and the status of the game buttons is indicated when pressed.

5. Embedded controller (EC): Richard Smith reports that several EC issues are pending. The one that has received attention the SCI Mask Corruption bug (Ticket #5467). While chasing this bug, Richard found several small but critical typos in the handling of some of the commands he had added since November. The net result of these typos is that under some conditions, a value passed in as data would be run as a command and some commands would not get run at all. Unfortunately, fixing those typos does not fix #5467. Its cause goes much deeper into the EC protocol handling. The next couple of days should shake out what the problem is and get it fixed for good.

Battery problems: A growing number of reports of short battery life are coming in. People are starting to run olpc-logbat bat and Richard has been looking at the resulting data. Based on the data he's seen so far, he conjects that either (1) there are some “funky” batteries in the field; or (2) the EC is failing to charge the battery up to full capacity, yet it is marking it as full. Most of the data gathered so far has been discharge info. Richard will be responding to many of the trouble tickets requesting several cycles of charge/discharge while running olpc-logbat to flush out whats going on.

The report of “shutdown yet no red LED” is the result of the capacity never going below 15% but the battery voltage dropped to the critical cutoff point, followed by the EC dropping system power. An enhancement Richard will make to the EC code is to also do something with the status LED to indicted that a critical voltage shutdown is looming so there's some warning your laptop is about to shutdown.

6. Schedules: For the next few weeks we will be focused on stabilizing Update.1 (based on Joyride) through testing, documentation, and limited number of bug fixes. We recently found two more critical bugs that will need an unscheduled software release (USR): touchpad/mouse jumpiness and data loss if you fill up the memory. We have created a procedure for these USRs; we are using this process to get these fixes out sooner than the next scheduled release (See Operating system release procedures).

7. Test: Chih-Yu Chao is helping out with both test and support this week. She has gotten through the 1-Hour Smoke Test on a recent Joyride build, which revealed a few regression bugs from the Ship.2 (650) release. Next we need to create and document some test cases for the new features in Update.1, and some testing with the school server.

8. Support: Adam Holt has been working days, nights, and weekends to grow the volunteer support group (now up to 40 people), who are answering emails, hanging out on forums and IRC to help people. A core crew of about 15 volunteers drives the process forward with increasingly sophisticated answers. Others contribute part-time working from the Support FAQ (Support FAQ) and “RTFM” template answers as they get up to speed. We have almost hit 1000 email help requests in the database! Katie Belisle and Casey Ratliff are working on a next-generation documentation ideas for our Support FAQ.

Adam is also coordinating “4PM Sunday” (Eastern Standard Time) conference calls for the entire support-volunteers team. Last week’s call was extremely successful due to the contributions of the OLPC developer community (special thanks to Bernie Innocenti and Arjun Sarwal) (See Support meetings). Anyone you wants to join, email me well in advance at “holt AT laptop DOT org” and be sure to include your phone number! All volunteers worldwide will be considered, after a very brief phone call. (See http://lists.laptop.org/listinfo/support-gang if you are interested in volunteering.)

Adam, Kim Quirk, and Greg Babbin are now able to provide RMAs, which will help off-load the donor support 800-number and email. Kudos to Greg’s genuine heroics. Our Asterisk-based VoIP virtual call center has been briefly delayed. Matthew O’Gorman and Joe Phigan continue to work hard on this, scripting prompts for rudimentary integration with http://rt.laptop.org, and we should have our volunteer-training shortly.

9. Localization: Sayamindu Dasgupta helped start Pashto and Bulgarian translation teams and resolve an issue with Pootle that caused it to reject new user registrations; it was being caused by a invalid username in the Pootle database.

A long-standing blocker bug (Ticket #1525) regarding the invalidation of the fontconfig cache was finally fixed. Font cache generation in the XO should be more robust now, even in the face of clock failures.

10. Software potpourri: Tomeu Vizoso proposed fixes for a number of bugs that have highest priority: previews are not deleted when their matching datastore objects are removed (Ticket #5707); deleting a large file from a USB stick copies it into NAND (often filling NAND) (Ticket #5719); Sugar shell consuming vast amounts of memory (Ticket #5532); “resuming” a large file from USB copies it into NAND (filling NAND) (Ticket #5744); and objects accumulate on the clipboard, impacting system performance (Ticket #5760).

Chris Ball fixed many power manager bugs. We now perform power management regardless of whether you're on an external power source, we remember the user's previous brightness setting when we dim the screen during suspend, and open hardware manager (OHM) no longer exits when X does.

Chris Ball wrote “slideshow” over Christmas, which is a Pippy example that queries the datastore for camera images and then shows them full-screen in a slideshow. He can't decide whether it should be a Pippy example (since it demonstrates performing datastore searches in Python) or a separate activity.

Dennis Gilmore spent most of the week troubleshooting issues, working around an issue today causing build failures and mostly trying to put the pieces together to make things better.

Michael Stone and Dennis spent some time working out why iputils fell out of our builds. Michael also worked with Bernie and Tomeu on address a problem with olpc-update in regards to persistent activity directories (Ticket #5033), with Ben Schwartz on problem with stream sockets (Ticket #5818), and with Eben Eliason on the beginnings of a security user interface.

Ivan Krstić is exploring a more secure way of isolating Browse for Update.1; it might be trivial.

11. Presence: Robert McQueen finished an out-of-band data (OOB) implementation (he added IP detection code) and wrote tests for it. That means OOB bytestream is now working with Gabble. The next step will be to define and implement bytestream renegotiation and fallback.

Dafydd Harries made updated packages for Presence Service and Avahi, although Koji cannot build the former for some reason. He also debugged problems registering laptops with school servers (Ticket #5834); it turns out that the ejabberd RPM doesn't generate an X.509 certificate. Dafydd also spent time trying to coax OpenFire into working. It works ok as long as your account is not in the shared roster group, but authenticating becomes unreliable as soon as you are a member. The web interface becomes unreliable from time to time too, necessitating restarting the server. It seems that, like with ejabberd, we are using it in a way it is not designed to handle. Our scalability improvements should solve this for Update.2, but it is not clear yet what the best approach is for the Update.1 time frame.

Morgan Collett fixed the scrolling bug (Ticket #2351) in Chat for Update.1 thanks to a patch from Marco Pesenti Gritti. (C. Scott Ananian’s view source changes for Chat are in Update.1, but will require a newer Pippy.) Morgan is testing a fix for Presence Service #5368 where the buddies in an activity weren't reliably clustering around the shared activity icon.

12. Activities: Muriel Godoi progressing on his port of Food Force for the XO (See Food Force 2). Progress includes artwork (builds and villagers); next, the game model need more work to get a playable game. The code is in his public_git folder (https://dev.laptop.org/git?p=users/murielgodoi/foodforce2;a=summary). Please contact Muriel if you’d like to help.

Arjun Sarwal reports progress on the Measure activity: he is rethinking certain aspects of the code design of the activity that would make it more easily expandable and scalable.

13. OLPC Pakistan: Dr. Habib Khan reports progress amidst chaos. Urdu localization is 100% complete. They have had a useful discussion with an Afghan graduate student of International Islamic University (IIU) who is keenly interested in translating OLPC bundles into Dari and Pashto. They are also mobilizing volunteers from the Pakistan Institute of Engineering & Applied Sciences, (PIEAS). A training package for Afghan teachers is nearing completion, including software, hardware, and activities tutorials. They’ve also launched an effort to convert into e-books all the text books written on curriculum of the Federal Ministry of Education, Islamabad. Beta versions for Grades one through four are ready. The subjects are English, Social Studies, Science, and Urdu.

14. Cow power: Arjun have completed documentation of the project (See Cow Power). The page details the current design and the proposed mechanical design. He is hoping to get feedback from the community on the proposed mechanical design before moving forward in the implementation of the changes.

15. Community: The OLPC Austria team reports progress on OpenWRT. It boots an XO (currently with minimal driver support) in 15–20 seconds. John Crispin and others want to look into porting Sugar to OpenWRT if there is community interest.

Pascal Martin of Linterweb, an open source software company based in France that has worked on desktop and wiki search tools, has offered their support and development time to help with the search component for the Journal. Tomeu Vizoso spent some time explaining to Fabien Coulon from their team what has been done to date in the datastore.

Jesper Taxbøl is helping organize this year's Nordic Game Jam; he is angling to run it on XOs and lead off with an introduction to PyGame. They are looking for 10 laptops for their 100 participants to use February 1–3. This is quite a popular jam and produces some pretty polished games each year.

Many people are asking for ways to contact the creators of bundles and activities. Please add your name and some sort of contact info to projects you have worked on, on their own wiki pages, and on the Activities page.

More News

Laptop News is archived here.

You can subscribe to the OLPC community-news mailing list by visiting the laptop.org mailman site.

Press requests: please send email to press@racepointgroup.com

Milestones

Latest milestones:

Nov. 2007 Mass Production has started.
July. 2007 One Laptop per Child Announces Final Beta Version of its Revolutionary XO Laptop.
Apr. 2007 First pre-B3 machines built.
Mar. 2007 First mesh network deployment.
Feb. 2007 B2-test machines become available and are shipped to developers and the launch countries.
Jan. 2007 Rwanda announced its participation in the project.

All milestones can be found here.


Press

You can subscribe to the OLPC community-news mailing list by visiting the laptop.org mailman site. Template loop detected: Press More articles can be found here.

Video

Miscellaneous videos of the laptop can be found here.

More articles can be found here.

Video

Miscellaneous videos of the laptop can be found here.