Mobile
|
Size: 6535
Comment: started fuller comments
|
Size: 11596
Comment:
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 12: | Line 12: |
| * Very good discussion, talked about concentrating on mid devices, and leaving issues with netbooks to the desktop. | Conversation focused on extending kourou, with identification of specific requirements for implementation in Jaunty. These included finger scrolling, background image support, and various miscellaneous bugfixes. |
| Line 15: | Line 15: |
| Line 23: | Line 22: |
| * Dropped | After some review of the requirements, level of effort, and upstream focus, participants agreed that no special effort would be put into updating and coordinating hildon packaging for Jaunty: packages will generally track Debian for merge/sync as normal. |
| Line 33: | Line 32: |
| * The ship seed discussion, settled on about 6 items to be added to the ship seed for mobile. | https://blueprints.launchpad.net/ubuntu/+spec/ship-seed-for-mobile-images Discussion included review of requirements, and identification of specific packages to be included. Further input from other developers was sought to ensure the results would be useful for the widest range of sensible environments. With high probablility, this will mirror the ship seeds used for the Desktop images. |
| Line 36: | Line 37: |
https://blueprints.launchpad.net/ubuntu-mobile/+spec/moblin-compliance |
|
| Line 42: | Line 46: |
| * Expecting greater divergence in future between performance-optimised and power-optimised x86 decendents. lpia remains useful for experimentation, and as example for later divergence. * Real divergence requires porting autotools, etc. |
https://blueprints.launchpad.net/ubuntu/+spec/lpia-versus-i386 Intel offered insight into the benefits of lpia, and future plans for the platform. As the performance-optimised and power-optimised processor trees develop, there is greater divergence expected, with a strong possibility that the two branches may not remain binary compatible in the future. While lpia has little clear benefit today, it remains useful as a platform for experimentation with various optimisations available. Once the processor families diverge sufficiently, lpia will need to be defined as an architecture for autotools, etc. before which time it remains an experimental x86 branch. |
| Line 46: | Line 51: |
| * Plenty of clever brainstorming coming up during the application switcher session * We need application singletons for *all* applications; this can be done in the launcher. * Plans to show running apps and switch to them: new Hildon home plugin + new Kourou! |
https://blueprints.launchpad.net/ubuntu/+spec/mid-application-switcher Wide discussion on use cases and requirements for an application switcher, with many good ideas presented for solutions. Consensus developed on the replacement of the Home plugin with a more advanced tool to select the desired application. For best results, such a tool should be coordinated with adjustments to kourou to better handle single-launch of applications. |
| Line 51: | Line 57: |
| * Everyday except the first day, did a quick review of prior days specs and results. | These sessions tended to include short reviews of meetings attended the prior day in other tracks, implications for expected future discussions on the Mobile track, and continued review of application selection. There were no special items of note on any days, and so the description will not be repeated chronologically. |
| Line 54: | Line 61: |
| * We are continuing to make USB images and adding a question on Brainstorm to understand if anyone really needs CD iso's for MID UMPC * Discussing building ISO images for Ubuntu MID and UMPC has been deferred due a lack of data -- a brainstorm idea to vote on coming shortly! |
https://blueprints.launchpad.net/ubuntu-mobile/+spec/mobile-iso-images Conversation focused on benefits: while the implementation is fairly trivial, and there are tools that allow one to convert an ISO to a USB key for booting, it was expected that very few users would be able to directly use a CD to install a system. It was determined to open an idea on brainstorm to understand the support behind the use cases for CD images (distribution at events, ease of sharing and archiving, etc.) is sufficient to outweigh that the majority of end-users would need to use special tools to prepare installation media. |
| Line 58: | Line 67: |
| * Drop mobile-dev : no longer needed * Move seeds to ubuntu-mobile-dev bzr control * rename "mobile" seed to something else: likely netbook or similar |
https://blueprints.launchpad.net/ubuntu/+spec/mobile-team-seed-management Fairly quick review of the state of the seeds, and how to manage them. Nobody present used the "mobile-dev" set of packages, so that shall be dropped. Seed management will likely move from ubuntu-core-dev to ubuntu-mobile-dev, pending further insight into the results from ArchiveReorganisation. The "mobile" seed, task, and metapackage will be renamed, as part of the effort to subsume the Netbook Remix into Ubuntu itself. |
| Line 63: | Line 73: |
| * UNR (Ubuntu Netbook Remix) based on Clutter 0.6, - Intrepid has 0.8. Its not based on the newest stuff * UNR needs a new name. (maybe) * Name remained Ubuntu-Netbook-Remix * OEM Has UNR image, Mobile has UMPC that contains most of the UNR packages, but a bit out of date. * Might not be able to avoid netbook name due to existing brand establishing * Was reason for name remaining UNR * UNR is using moblin-image-creator, but that change is for another day. * The translations for UNR are hosted in Launchpad, but not currently in langpacks * UNR has ~43 modified packages, to make apps run at 1024x600, change strings to be smaller, and to make buttons larger and easier to use. * UNR seed uses additional packages such as desktop-switcher, and diverges from the desktop seed. * The goal of the UNR seed is to be usable on netbooks on screen, and makes sense on netbooks (i.e. no gimp) + simplicity * Should GNOME customizations be kept in UMPC? * UMPC images, current virtual keyboard is cellwriter, but disabled because it was buggy in intrepid. * UNR launcher needs more work to be finger friendly * Clutter actually needs 3D, which VMware doesn't support, nor does KVM |
https://blueprints.launchpad.net/ubuntu/+spec/unr-handling-jaunty Conversation focused on defining the requirements related to integrating UNR into Jaunty. UNR packages will be integrated into Jaunty, and Mobile team will produce a UNR image for Jaunty. The UMPC image produced for Intrepid will be dropped going forward as a result. Outstanding issues, tasks, and splitting of work between Mobile and OEM teams was discussed as well and is documented in the specification. |
| Line 81: | Line 79: |
| * ARM images - netboot/live images/alternative images * ARM boot - uboot/redboot/others?? * ARM USB boot >>>Kernal-less-img * ARM netboot - quick installer * ARM - Netboot/Alternate images small UI, Live Images a larger UI * ARM Live images - ubuntu desktop, porting packages * ARM - ubuntu-mid, ubuntu server, UNR, zubuntu * ARM build - ubuntu,kubuntu, xubuntu, server, mid, umpc =9 cds * ARM - Mid moves to future?? Support = commited testers is an issue |
|
| Line 91: | Line 80: |
| https://blueprints.launchpad.net/ubuntu/+spec/selection-of-arm-images Initial consensus on netbook, live, and alternate images. Some discussion about various bootloaders (uboot, redboot, etc.), and how to handle variance in how ARM hardware boots. Consideration of live images being provided without kernels, so they may be reused for multiple subarchitectures, as kernels take a fair amount of space. Given the wide number of flavours, and initial smaller set will be selected, with flavour-specific teams to focus on other images if desired. Firm decision on Ubuntu Desktop (live), Ubuntu Server (alternate), and a universal netboot image. Kubuntu, Xubuntu, MID, etc. to be deferred for renewed investigation by flavour-specific teams. |
|
| Line 93: | Line 85: |
| * Mobile devices - ARM N770, N800, N810, Sharp Zaurus 3200, Beagleboard, NSLU2, Atom Toshiba N80, Compal Jax10, Sharp D4. Other: Samsung Q1 Ultra | |
| Line 95: | Line 86: |
| https://blueprints.launchpad.net/ubuntu/+spec/mobile-hardware-support Very little discussion on outstanding issues, or specific hardware that needed support. Participants demonstrated multiple devices on which they had or would like to install Ubuntu. Specific devices demonstrated included: Nokia n770, Nokia n800, Nokia n810, Sharp Zaurus 3200, Sharp D4, Beagleboard, NSLU2, Toshiba N80, Compal Jax10, Samsung Q1 Ultra. |
|
| Line 97: | Line 91: |
| * Mobile Spec Cleanup : http://ur1.ca/05re : whitewashing history * Mobile Team just discussed roughly 80 specs in 30 minutes to drastically clean up our spec list |
https://blueprints.launchpad.net/ubuntu/+spec/mobile-spec-cleanup Lightning review of 75 specifications in 30 minutes, with many determined obsolete or deferred for review at the next UDS. Most specifications were previously completed, but launchpad had not been updated to reflect the status accurately. |
| Line 101: | Line 97: |
| * In the bootmenu session it was discovered that the jaunty images already have all we need. | https://blueprints.launchpad.net/ubuntu-mobile/+spec/gfxboot-like-menu-for-mobile-images After short discussion it was realised that most of the goals for this effort had already been accomplished by the change in the image building method, and that some others were rendered obsolete by the adsorbtion of Netbook Remix into Ubuntu itself. It was decided not to proceed with a separate effort. |
| Line 105: | Line 104: |
| https://blueprints.launchpad.net/ubuntu/+spec/mobile-setup-wizard The expected leader was not present, so after a late start, most of the conversation was related to questions about oem-config, and how it might be used, and whether it would meet the requirements. It was decided to include oem-setup and test it throughly with Jaunty, with possible UI adjustments if required for lower resolution environments. |
|
| Line 107: | Line 110: |
| https://blueprints.launchpad.net/ubuntu/+spec/recovery-partition Many OEMs ship with machines that have an onboard recovery partition so recovery is possible by tapping a single key and then restoring from that image. As Ubuntu is now being shipped by OEMs, we should have similar functionality incorporated into ubiquity. Casper should be modified to have a restore/recover mode, and update-manager should be able to upgrade the rescue partition. |
|
| Line 108: | Line 115: |
| * Focus on screen area rather than specific libraries or framework * Significant UI changes planned for ubiquity in Jaunty, many of which will meet MID use cases. |
https://blueprints.launchpad.net/ubuntu/+spec/ubiquity-hildon Ubiquity has significant UI adjustments planned for the Jaunty cycle. As most of the MID-specific requirements were more closely related to issues with low resolution than with issues with hildonisation, conversation focused on a review of the current UI to discuss which elements did not fit in 480 vertical pixels. The results are expected to feed into the general ubiquity redesign, with no specific work being done on a separate front-end. |
| Line 113: | Line 122: |
| https://blueprints.launchpad.net/ubuntu-mobile/+spec/poulsbo-packaging Discussion of how to repackage the Intel binary driver dump for the Poulsbo graphics cards to allow usage in Intrepid with minimal user effort, including the ability to cleanly uninstall the driver set. Provided data shall be repackaged in a manner consistent with Intrepid, and placed in a PPA for user access. |
|
| Line 114: | Line 127: |
https://blueprints.launchpad.net/ubuntu-mobile/+spec/offline-installer Good discussion on the requirements, UI, and features of a tool to generate customised bootable images for a variety of situations, including cross-architecture preparation of images. Primary unresolved issue was handling of maintainer scripts for installed packages in a sane manner when performing cross-architecture images. |
|
| Line 117: | Line 134: |
| Discussion of issues with Malone of interest to participants in the Mobile Track. Some items of note included being able to identify all bugs against all packages included in a specific flavour (or unique to a specific flavour), clean movement of bugs between projects (e.g. ubuntu-mobile) and distributions (e.g. ubuntu), and discussion of general usability issues. |
|
| Line 119: | Line 138: |
| https://blueprints.launchpad.net/ubuntu/+spec/general-resolution-for-touchscreen-handling |
|
| Line 120: | Line 141: |
https://blueprints.launchpad.net/ubuntu/+spec/squashfs-initrds |
Mobile
Provide an easy-to-use xrandr interface for MID
https://launchpad.net/ubuntu/+spec/mid-screen-rotation
Conversation focused on the need for a UI widget to allow for general xrandr services for MID. Alberto Milone is working on a python interface for xrandr, which is probably a sensible basis for this tool. Devices with accellerometers or dedicated hardware screen change buttons may benefit from support for hotkey activation of some services.
Discuss the launcher used for Jaunty on Ubuntu MID
https://blueprints.launchpad.net/ubuntu-mobile/+spec/mid-jaunty-launcher
Conversation focused on extending kourou, with identification of specific requirements for implementation in Jaunty. These included finger scrolling, background image support, and various miscellaneous bugfixes.
Applications used in Ubuntu MID
https://blueprints.launchpad.net/ubuntu/+spec/mobile-applications
General review of the core application set for MID usage, and a selection of applications to meet that need. Discussion did not complete within the session, and is expected to continue throughout the week at the Mobile roundables. This is likely to be a repeating review, to ensure that the selected applications are the best fit for the use cases of those available within Ubuntu.
Packaging of Hildon in Jaunty
After some review of the requirements, level of effort, and upstream focus, participants agreed that no special effort would be put into updating and coordinating hildon packaging for Jaunty: packages will generally track Debian for merge/sync as normal.
Display Manager for MID
https://blueprints.launchpad.net/ubuntu/+spec/mid-display-manager
Early in the discussion it was determined that supporting a separate display manager would be a fairly high overhead. Solutions considered included looking at how the new gdm might fit, although it has an increasing amount of overhead to support Desktop use cases, and to modify startx to work properly, as startx is a common use case in other contexts as well. Conclusion was to look at enabling proper support for consolekit and policykit in startx itself.
Define what goes into the ship seed on the mobile images
https://blueprints.launchpad.net/ubuntu/+spec/ship-seed-for-mobile-images
Discussion included review of requirements, and identification of specific packages to be included. Further input from other developers was sought to ensure the results would be useful for the widest range of sensible environments. With high probablility, this will mirror the ship seeds used for the Desktop images.
Moblin Overview and Compliance
https://blueprints.launchpad.net/ubuntu-mobile/+spec/moblin-compliance
- Intel provided information as to what Moblin 2.0 Compliance is.
- Superset of Linux Standard Base (LSB) 4.0 Specification
- Will provide additional tests to show compliance.
Weighting the benefits of lpia versus i386
https://blueprints.launchpad.net/ubuntu/+spec/lpia-versus-i386
Intel offered insight into the benefits of lpia, and future plans for the platform. As the performance-optimised and power-optimised processor trees develop, there is greater divergence expected, with a strong possibility that the two branches may not remain binary compatible in the future. While lpia has little clear benefit today, it remains useful as a platform for experimentation with various optimisations available. Once the processor families diverge sufficiently, lpia will need to be defined as an architecture for autotools, etc. before which time it remains an experimental x86 branch.
Application switcher for MID
https://blueprints.launchpad.net/ubuntu/+spec/mid-application-switcher
Wide discussion on use cases and requirements for an application switcher, with many good ideas presented for solutions. Consensus developed on the replacement of the Home plugin with a more advanced tool to select the desired application. For best results, such a tool should be coordinated with adjustments to kourou to better handle single-launch of applications.
Mobile Roundtable
These sessions tended to include short reviews of meetings attended the prior day in other tracks, implications for expected future discussions on the Mobile track, and continued review of application selection. There were no special items of note on any days, and so the description will not be repeated chronologically.
Create ISO images for Ubuntu Mobile and Ubuntu MID
https://blueprints.launchpad.net/ubuntu-mobile/+spec/mobile-iso-images
Conversation focused on benefits: while the implementation is fairly trivial, and there are tools that allow one to convert an ISO to a USB key for booting, it was expected that very few users would be able to directly use a CD to install a system. It was determined to open an idea on brainstorm to understand the support behind the use cases for CD images (distribution at events, ease of sharing and archiving, etc.) is sufficient to outweigh that the majority of end-users would need to use special tools to prepare installation media.
Discuss the mobile seed, and related topics
https://blueprints.launchpad.net/ubuntu/+spec/mobile-team-seed-management
Fairly quick review of the state of the seeds, and how to manage them. Nobody present used the "mobile-dev" set of packages, so that shall be dropped. Seed management will likely move from ubuntu-core-dev to ubuntu-mobile-dev, pending further insight into the results from ArchiveReorganisation. The "mobile" seed, task, and metapackage will be renamed, as part of the effort to subsume the Netbook Remix into Ubuntu itself.
Discussion of UNR issues for Jaunty
https://blueprints.launchpad.net/ubuntu/+spec/unr-handling-jaunty
Conversation focused on defining the requirements related to integrating UNR into Jaunty. UNR packages will be integrated into Jaunty, and Mobile team will produce a UNR image for Jaunty. The UMPC image produced for Intrepid will be dropped going forward as a result. Outstanding issues, tasks, and splitting of work between Mobile and OEM teams was discussed as well and is documented in the specification.
Find the right set of images to build for the ARM architecture
https://blueprints.launchpad.net/ubuntu/+spec/selection-of-arm-images
Initial consensus on netbook, live, and alternate images. Some discussion about various bootloaders (uboot, redboot, etc.), and how to handle variance in how ARM hardware boots. Consideration of live images being provided without kernels, so they may be reused for multiple subarchitectures, as kernels take a fair amount of space. Given the wide number of flavours, and initial smaller set will be selected, with flavour-specific teams to focus on other images if desired. Firm decision on Ubuntu Desktop (live), Ubuntu Server (alternate), and a universal netboot image. Kubuntu, Xubuntu, MID, etc. to be deferred for renewed investigation by flavour-specific teams.
Roundtable review of available hardware and device support
https://blueprints.launchpad.net/ubuntu/+spec/mobile-hardware-support
Very little discussion on outstanding issues, or specific hardware that needed support. Participants demonstrated multiple devices on which they had or would like to install Ubuntu. Specific devices demonstrated included: Nokia n770, Nokia n800, Nokia n810, Sharp Zaurus 3200, Sharp D4, Beagleboard, NSLU2, Toshiba N80, Compal Jax10, Samsung Q1 Ultra.
Review and clean up mobile-related specs in launchpad
https://blueprints.launchpad.net/ubuntu/+spec/mobile-spec-cleanup
Lightning review of 75 specifications in 30 minutes, with many determined obsolete or deferred for review at the next UDS. Most specifications were previously completed, but launchpad had not been updated to reflect the status accurately.
Have a selection menu in the mobile images, similar to the CD bootmenu
https://blueprints.launchpad.net/ubuntu-mobile/+spec/gfxboot-like-menu-for-mobile-images
After short discussion it was realised that most of the goals for this effort had already been accomplished by the change in the image building method, and that some others were rendered obsolete by the adsorbtion of Netbook Remix into Ubuntu itself. It was decided not to proceed with a separate effort.
Ubuntu Mobile Setup Wizard
https://blueprints.launchpad.net/ubuntu/+spec/mobile-setup-wizard
The expected leader was not present, so after a late start, most of the conversation was related to questions about oem-config, and how it might be used, and whether it would meet the requirements. It was decided to include oem-setup and test it throughly with Jaunty, with possible UI adjustments if required for lower resolution environments.
Recovery partition
https://blueprints.launchpad.net/ubuntu/+spec/recovery-partition
Many OEMs ship with machines that have an onboard recovery partition so recovery is possible by tapping a single key and then restoring from that image. As Ubuntu is now being shipped by OEMs, we should have similar functionality incorporated into ubiquity. Casper should be modified to have a restore/recover mode, and update-manager should be able to upgrade the rescue partition.
Create a hildonised ubiquity front-end
https://blueprints.launchpad.net/ubuntu/+spec/ubiquity-hildon
Ubiquity has significant UI adjustments planned for the Jaunty cycle. As most of the MID-specific requirements were more closely related to issues with low resolution than with issues with hildonisation, conversation focused on a review of the current UI to discuss which elements did not fit in 480 vertical pixels. The results are expected to feed into the general ubiquity redesign, with no specific work being done on a separate front-end.
Packaging of the Poulsbo drivers
https://blueprints.launchpad.net/ubuntu-mobile/+spec/poulsbo-packaging
Discussion of how to repackage the Intel binary driver dump for the Poulsbo graphics cards to allow usage in Intrepid with minimal user effort, including the ability to cleanly uninstall the driver set. Provided data shall be repackaged in a manner consistent with Intrepid, and placed in a PPA for user access.
Provide an offline tool that generates an installed image for physical installation.
https://blueprints.launchpad.net/ubuntu-mobile/+spec/offline-installer
Good discussion on the requirements, UI, and features of a tool to generate customised bootable images for a variety of situations, including cross-architecture preparation of images. Primary unresolved issue was handling of maintainer scripts for installed packages in a sane manner when performing cross-architecture images.
Mobile Bugs
Discussion of issues with Malone of interest to participants in the Mobile Track. Some items of note included being able to identify all bugs against all packages included in a specific flavour (or unique to a specific flavour), clean movement of bugs between projects (e.g. ubuntu-mobile) and distributions (e.g. ubuntu), and discussion of general usability issues.
Review the touch screen situation with regard to calibration tools and driver situation
https://blueprints.launchpad.net/ubuntu/+spec/general-resolution-for-touchscreen-handling
Initrds in squashfs format instead of a gzipped-cpio stream
https://blueprints.launchpad.net/ubuntu/+spec/squashfs-initrds
UDSJaunty/Report/Mobile (last edited 2009-01-17 15:20:41 by serris)