XOCamp 2

The XOcamp 2 miniconference in Has location city::Cambridge Has location country::USA, is scheduled for Start date::January 12, 2009 to End date::January 15, 2009, with a presentation from a Quebec trial on January 16.

The two days of heavy public discussions will be Monday and Tuesday, January 12-13.

FUDcon will be Jan 9-11 See: http://fedoraproject.org/wiki/FUDCon/FUDConF11. XOCamp will be the week following.

The purpose of the XOCamp is to review status of code, specifications and audience for the 9.1.0 release; review activity development; and frame our long-term development effort.

Presentations will be broadcast on XOCamp's Justin.tv channel.

Agenda

All meetings held in the Boardroom at One Laptop per Child, 10th floor, 1 Cambridge Center, Cambridge, MA

A conference line will be open for remote attendees.
Within the United States : 866-213-2185
Outside the United States: 1-609-454-9914
access code: 8069698#

Live streaming from http://www.justin.tv/xocamp

Monday January 12, 2009

Time	Main conference room	Foyer and hacking space
0845	Coffee and registration
0900	Introductions and Agenda Review.
0915	Requirements review and deployment status (Greg's notes)	Open Space whiteboard
1000	Future development of XOOS : 8.2.1, 9.1, F11 and beyond.
1100
1200	Lunch, FUDCon review, and demonstrations
1300	Deployment Notes from Nepal by Bryan
1400		Activity hacking & showcase
1500	Break
1515	Synchronizing Sugar and Fedora : features, testing, releases and related toolchains (workshop)
1530-1615	OLPC QA, now with 100% Community Goodness (Mel Chua, Ed McNierney)
1700	Review & closing comments
1900	Drinks @ TS
evening	Open hacking

Tuesday January 13, 2009

Time

0845 - Coffee & Donuts

0900 - School Server Roadmap and 0.6 Design Review by Martin, potential followup in the afternoon

1015 - Break

1030 - Activation/lease/signing/image customization - timeline and testers.

1100 -

1200 - Lunch

1300 - Fedora Compatibility - Rebase on Fedora 10, F11, or CentOS? Running Fedora applications and window managers. NAND vs. SD card and Erikg's latest F10 build

1430 - Moodle and Content

1600 - Review & wrapup : next steps

1700 - Planning for future events (FOSDEM, &c)

1830 - Dinner (meet downstairs @ 1CC)

Wednesday January 14, 2009

Time

0900 - [morning free]

1000 - OLPC Startup presentation by Bryan Berry

1130 - Ed Cherlin on Digital Books

1200 - [lunch elsewhere] Afternoon : future roadmap discussions

1500 - Power Management led by Chris

1600 - Open discussions on localization, Alternate file systems, beyond JFFS, performance and Improving memory management, led by Scott?

1600 -

1645 - Review & close

Thursday January 15, 2009

Time

0900 - Open session / new activity demonstrations

1000 -

1100 - Karma - Framework to make it stupid-easy to create Activities - Bryan B

1200 - Lunch

1300 - John Gilmore talk & ideas

1500 - Possibly Asynchronous collaboration and Synchronous Collaboration or open session

1600

1645 - Review and close

Friday January 16, 2009

Time

0900 - Open session

1000 - Presentation on 1 - 1 Laptop deployment (using Macs) in Quebec by Ron Canuel & 3 others. See http://www.etsb.qc.ca/en/EnhancedLearningStrategy/default.shtm

1100 -

1200 - Lunch and close

Attendees

Sign up here if you plan to attend.

1. Andrés Ambrois
2. Andrius Kulikauskas
3. Bernie Innocenti
4. Bobby Powers (80% likely)
5. Brian Jordan
6. Bryan Berry
7. Caroline Meeks
8. Stefan Unterhauser
9. Joe Feinstein
10. Jonas Smedegaard
11. Luis Michelena 06:23, 20 December 2008 (UTC)
12. Martin Langhoff
13. Mchua 23:31, 19 December 2008 (UTC)
14. Mokurai 19:51, 19 December 2008 (UTC)
15. rjhatl 13:54, 19 December 2008 (UTC)
16. Ryan Kabir
17. --Sj talk
18. Tony Anderson
19. David Bauer
20. Marco Pesenti Gritti
21. Walter Bender (in and out)
22. Jim Gettys
23. Paul Fox
24. Polychronis Ypodimatopoulos
25. Daniel Bennett as available.
26. Henry Edward Hardy as available.
27. Jameson Quinn
28. Anil Daswani
29. Lucia Lovison-Golob

Technical Track

This section gives details on each talk.

Who is the user and what do they want?

Overview of top concerns from deployment leads, students and teachers. An aggregated look at the top items requested. An in depth look at the motivation of a few specific requests. Time permitting, can include a discussion of how to turn user feedback in to actionable requirements. Can include a detailed "use case" explaining the technical environment and social context of a target school.

See presentation File:XO-Camp-Usersv4.ppt

Deployment is Hard! Notes from Nepal's Deployments

Summary: Bryan Berry will talk about Nepal's deployments with help from Tony Anderson. Will talk about tough stuff like teacher training, developing local support infrastructure, managing volunteers, and particular technical needs.

• Time - 90 minutes + 30 minutes of discussion
• Bryan Berry and Tony Anderson will lead the discussion
• Powerpoint presentation will accompany the talk

See also Bryan's blog - http://blog.olenepal.org/

Sugar Synch Up

• Presentation on features and important bug fixes of Sugar 0.84.
• Discussion of Trac usage and bug fixing.
• Agreement on quality metrics. How many open bugs of which type will we accept and still release.
• Review of schedule and relevant meetings.

Sugar Roadmap

XO Software Release 9.1.0

School Server

Summary - Status and plans.

• Quick overview of the status of the XS
• Areas where can relieve pressure from the XO
  • Narrative and content publishing/ sharing
  • Upgrade and other mgmt tools
• Focus areas to improve XO-XS interaction for 9.1 release cycle
  • Service announcement (DNS? mDNS? Zeroconf? DNS masking?)
  • "XS within sight" flag on XO
  • Lease management protocol - (to be continued in theft deterrence)
  • Browse.xo automagic auth
  • ejabberd, presence service / roster mgmt
  • "Thicker" wikislices
• Interesting goals for 9.2 or later
  • Backup as extension of Journal (a la Time Machine)
  • "Universal" book reader
  • Alternative mesh strategies - Cerebro
  • Sneakernet - wwwoffle
  • MikMik and the XS

Fedora Compatibility

Feature roadmap/Rebase on Fedora 10
Test and development work to base 9.1.0 on Fedora 10.
List of packages that to get in to Fedora

Feature roadmap/Run_Fedora_applications_on_XO Activity compatibility. Making it easy to run Fedora Linux apps on the XO.
Picking a desktop manager.

Activation/lease/signing/image customization

Feature roadmap/Activation_lease_security
Deployment controlled activation lease security

Feature roadmap/Image_customization
Updating faster/better for 9.1

Feature roadmap/Image_signing_key_delegation
How to allow deployments to sign their own images.

Feature roadmap/Faster_imaging

Feature roadmap/Activation_via_wireless

Power management

Feature roadmap/Improved_battery_life
Continuing to evolve software to wring all potential power savings out of the hardware. Timed wakeup, allowing screen blanking to recover power and allowing more aggressive suspending. Measuring power usage to avoid regressions. Testing for power against real world usage, pushing the remaining suspend-related bugs to completion.

Feature roadmap/No_power_regressions

Feature roadmap/Shutdown_menu
The power button is currently underutilized. It should allow either shutdown or suspend of the laptop, with the help of an on-screen dialog or menu.

Localization

Some key features from Software features in the area Localization

i18n and l10n: 9.1 and beyond
The areas I would like to cover include:

• Input methods: Our current input system (XKB) is not enough for Chinese/Japanese/Korean, and even existing customers like Ethiopia have needs which are met with considerable difficulty with XKB. For 9.1, I propose that we switch to SCIM[1], which should take care of most (if not all) of our current requirements.
• Multiple, non English fallback languages: Many of our users will probably be multilingual. An Aymara speaker would like to see the UI strings falling back to Spanish translations in absence of Aymara instead of the normal, default English. We have the existing infrastructure in our core i18n infrastructure, implementing this is simply a matter of making changes in Sugar (and maybe Rainbow).
• Language packs: The current system of language packs is not very reliable (it overwrites the original translations in the system, installations cannot be easily undone, no versioning, etc). I want to switch to a RPM based language pack system for 9.1, which will be easier to deploy (especially if the customization key mechanism gets support for RPM). This will require support for multiple locale directories, and I'll talk about that as well (Ubuntu uses a similar approach - I plan to reuse their patches if possible).
• User modifiable translations: There has been a number of ideas on this - initially (during my last visit to 1CC), we were thinking of a "Translate Activity" which would let children add/modify translations, and optionally let them share the translations over the mesh. SJ and Scott have recently suggested a wiki-like editable UI, which would, definitely be much cooler ;-). I want to brainstorm on this, and try to figure out the amount of time we need to do this, the advantages, risks, tradeoffs involved etc.

Presenter: Sayamindu

Filesystems

Topics to be covered include: Why replace JFFS2, a summary of the existing alternatives, some preliminary performance data, an overview of all the SW stack changes required to implement a new filesystem. We will also discuss the alternative option of sticking to JFFS2 and trying to fix the known issues. See also: http://wiki.laptop.org/go/Feature_roadmap/Replace_JFFS_file_system_with_better_one

Performance

System-level performance tweaks for UI responsiveness.

See also: http://wiki.laptop.org/go/Category:Performance

Dealing with Low Memory conditions

Thread started by Deepak on low memory at: http://lists.laptop.org/pipermail/devel/2008-October/020543.html

See also: http://wiki.laptop.org/go/Feature_roadmap/Memory_pressure

John Gnu Block

• Replacing Sugar Totally - whether and how to climb out of the dead-end without further investment - gnu

• Consider eliminating DRM
  

Now that Fedora runs out of the box on the XO, it's become more obvious that shipping jailed laptops is bad policy. It's been done via a misguided "security = good" mindset and a lazy "fewer SKUs is easier" approach. - gnu

• Eliminating Mesh, keeping 802.11

Mesh barely works. We should expand our hardware choices and avoid locking our software into a failed effort, by moving our software and documentation to straight 802.11 ad-hoc and 802.11 access point configurations. We are almost there now, just need to clean up a few loose ends. - gnu

Asynchronous Collaboration

File or Journal Object Sharing -- Ability to transfer information for distribution or asynchronous collaboration

Collaboration

-- Synchronous and asynchronous collaboration, definition and requirements.

An idea for achieving greater scale for synchronous collaboration by using centralized management and groups.

An explanation of how asynchornous collaboration can be used. A discussion of the lessons learned from the EduBlog and Amadis experiences and consideration of the applicability of Moodle.

Also submitting homework and student - teacher interaction. See thread started by Yama on Devel: http://lists.laptop.org/pipermail/sugar/2008-October/009472.htmland thread started by Mikus: http://lists.laptop.org/pipermail/devel/2008-October/020436.html

Gregorio 19:27, 23 October 2008 (UTC)

eBook Reader

Three parts below need to be combined. Gregorio 21:45, 12 November 2008 (UTC)

Martin is very keen on this topic. We need something really compelling, standard and compatible for OLPC and all/most other viable platforms to win. Currently, every content project creates their own viewer (often webbased). We want to be able to show the content from a thousand sources, through a single viewer. We need something so easy for content creators to embrace, and so obviously good that they embrace it.

What does an electronic textbook look like?

Since the 1960s there have been experiments in high-powered educational software, but not a lot of textbook development that integrates this software into the text, and very little classroom experience. This session will look at the available materials, the types of software and content available, and the implications for future curricula.

What do we know? What examples do we have and what do they show us?

What opportunities can we see? How do we make this happen? What questions should we ask next?

Examples:

• Edison Talking Typewriter to teach reading and writing to pre-school children
• Ken Iverson's textbooks, Arithmetic, Algebra, and Calculus
• Smalltalk and the Dynabook concept
• Matlab, Mathematica, and other powerful software
• Notebook and workspace formats and capabilities
• Teaching programming to children: TurteArt (sic), Smalltalk, Logo, APL, others

UPDATE: Alan Kay, Doug Engelbart, FLOSS Manuals, Open Learning Exchange, Creative Commons ccLearn, The Tech Museum of Innovation, and others have agreed in principle to join together to create a new kind of textbook. I'll invite them all to XOCamp. See Creating textbooks at Sugar Labs for more.

Submitted by Ed Cherlin--Mokurai 02:13, 29 October 2008 (UTC), updated 19:47, 16 December 2008 (UTC)

Andrius Kulikauskas (ms@ms.lt), Minciu Sodas, working on open source materials for learning mathematics, and sharing ideas from the Includer, and connecting with independent thinkers in Africa and around the world.

Synchronous Collaboration

Network principles -- how do we find others and collaborate with them?

--User:CScott (I can give a talk on this topic)

• How much time is needed?
• Who can lead the discussion?
• What preparation and materials should be created (e.g. presentations, code samples, other)?

Time Delayed internet --- offline caching and browsing, library integration?

--User:CScott (I can make a brief proposal; I'd like to hear others present)

• How much time is needed?
• Who can lead the discussion?
• What preparation and materials should be created (e.g. presentations, code samples, other)?

Gadget -- improving XMPP server based presence

Note: This was already presented at SugarCamp.

• How much time is needed?
• Who can lead the discussion?
• What preparation and materials should be created (e.g. presentations, code samples, other)?

Karma - Framework to make it stupid-easy to create Activities

Bryan Berry will talk about Karma, a framework that aims to make it stupid easy to create activities for the XO using Flash technologies

• Time - 60 - 90 minutes including discussion
• Bryan Berry will lead the discussion
• Powerpoint presentation will accompany the talk, along with demonstration

Pyjamas-desktop : Framework to make it easy to create FF-based activities

The Man In Black will talk about Pyjamas-desktop, a framework for writing AJAX apps in Python and running them locally, and what is needed to extend this into a framework for activity creation.

• Time - 60 minutes? discussion tied into Karma discussion

Community Track

Community

Why it matters

In my view it is the whole bottom line, la raison d'être of the whole project. No matter the XO and us are agents in developing scientists, artists, lawyers, whatever (who anyway are wont to migrate North ASAP), if there is not a process of community getting built among people who are not familiar with the culture and processes of collaboration, we are for naught, except maybe in making things worse. And of course, it is not at all just about sharing an activity, which is a software/hardware issue, it's the whole concept of building things together, sharing, being open to give and take, losing the "no te metás" (don't get involved) mentality. Yamaplos 17:33, 2 December 2008 (UTC)

• Discuss specific strategies so that all the rest of the software / hardware effort is not wasted
• Specific fixes to software and hardware

• How much time is needed? a couple lifetimes? OK, gimme an hour at least, maybe in such a way that it can overflow into dinner or lunch
• Who can lead the discussion? Yama

I'm leaving this section in case anyone wants to follow up. I was told there is no funding at all for volunteers besides some fundraising happening on the side (link?), so I will not be going. Actually the very fact that a whole track is dedicated to community but there is no formal funding set aside to bring in community people proves community issues are many, not taken seriously, and it's not only a bottom-up problem... _{Yamaplos 16:31, 18 December 2008 (UTC)}

• What preparation and materials should be created (e.g. presentations, code samples, other)? I have volunteered EquipoSur_XOCamp2 to help

Distributing OLPC

Summary: Open discussion to address the following questions.

1. How can OLPC encourage worldwide decentralized community development models for pedagogy and the content and software to support it?
   1. And then let the whole worldwide community benefit? _{Yamaplos 19:39, 13 November 2008 (UTC)}
   2. Having the pedagogy community to present about their experience and work and challenges and triumphs would help... _{Yamaplos 16:37, 18 December 2008 (UTC)}
2. How can communication between/among OLPC-the-organization and grassroots pedagogical, content, and software developers be improved?
   1. Define, clarify channels. Who do you contact for a small deployment? How do you distribute not just green bricks but knowledge and training? Who should you contact in 1CC for what kind of issue? _{Yamaplos 19:39, 13 November 2008 (UTC)}
3. Whether and how we want the volunteer community to help us with 9.1
   1. If OLPC-the-organizations says that we support the open-source mentality and that we encourage volunteers, we should articulate exactly what kind of volunteer support we want and need, the kind of relationship the organization would like to have with the community, and the kind of support (including support and resources that are conditional) we can provide them with. Mchua 16:54, 29 October 2008 (UTC)
   2. I personally think that it is especially important to articulate the kind of volunteer projects we won't support and do not have resources for, so that community members know what they have to do completely independently of OLPC. Knowing the answer is a no is better, for planning and getting-things-done purposes, than remaining in a nebulous "maybe?" state. Mchua 16:54, 29 October 2008 (UTC)

totally second both points _Yamaplos

• How much time is needed - 1 hour
• Who can lead the discussion - Rafael, (Mel available for backup)
• What preparation and materials should be created (e.g. presentations, code samples, other)?

I am working on a 10-minute "canned" presentation to be used for getting people non-familiar with the program to get enough to jump in - Far from ready, but hope it will be Yamaplos 17:23, 17 December 2008 (UTC)

Status - On track to include. Next steps, ToC for presentation.

Why is this under "software engineering"? No doubt there are software-related issues we need to address, but this is primarily a wetware issue. The tools are there, agreed they are not perfect, yet the fact they are not being used has to do with how we reach (or miss reaching) people, not with hardware or software. Yamaplos 17:57, 2 December 2008 (UTC)

QA

Co-moderated by Mel Chua (OLPC QA Community) and Ed McNierney (OLPC VP of Software). All are welcome; the format's a bit unusual, as this started as "Mel and Ed should have a meeting to talk about QA going forward," ran into "But there's no time, and the community should be involved anyway," and rapidly turned into "well, why don't we just invite everyone else and have it at XOCamp?"

In a nutshell, OLPC has a QA staff of 0 as of Jan 9, 2009, but testing still needs to be done.

Proposal

3 areas of focus for the community, with OLPC-the-org support - big question: What does OLPC need to do now so that it doesn't have to do QA ever in the future if it doesn't want to?

1. Integration testing (minor, pushing out to other groups - over time this will diminish as we become more distribution/software agnostic) - "Fedora! You guys need to do your testing!"
2. System testing / testbeds (major, turning into minor)
3. Acceptance testing at deployments (minor, turning into major, but being done by not-OLPC-the-organization) - hopefully eventually deployments will be sending software images to OLPC!

History

• Community capabilities and activities so far
  • Friends in Testing (Brian Pepple - previously Michael Stone and Joe Feinstein)
  • G1G1 Activity Testing and local test groups (Tabitha, Alastair, the Welly Testers, et al)
  • Test case/results reporting infrastructure, with plenty of Semantic Mediawiki Magic (Skierpage, Carl, etc.)
  • Other things; cool tools? Great insights! (Gary, Ben, Various People Who Are All Named Chris)

General concerns

• Concerns about organization/communication (Mel: I think these concerns are right now too generic for us to address, but we should certainly keep them in mind.)
• Will local organization of test groups scale?

Current OLPC testing efforts

• Things that we know will need software QA: what's the gameplan?
  • 8.2.1 interim release (testing managed by Mel, Mel to present the current status/plan) Gameplan:
    • 17 bugs, see http://dev.laptop.org/report/38
    • Immediate next step: 5 in the stage where we can test them (test in build)
    • Immediate next next step: check with the committer of the patch (deepak, etc.) and see if there are areas of potential regression to test around
    • If There Is Time: regression risk analysis for "add to release" tickets (2), but this is lower-hanging fruit.
    • Question: How does this interact with triage?
    • Question: What criteria do you want for us to set for QA signoffs? (Who can sign off, etc?)

For 8.2.1, Ed is the signoff person.

• • 9.1.0 release (testing managed by... migration to Fedora!)
• Acceptance testing at deployments: Something we haven't done a lot of yet
• What are the other things that will need QA in the near future? the not so near future?
• What kind of tools and info access are needed to carry out that testing?
• What resources can OLPC provide community testers with?
• How can we encourage local testing groups like the one in Welly to form up?
• And more?

Tools

• semantic mediawiki test cases and test reporting forms. This is... functional, but sort of awkward. Potential to explore alternative solutions - not sure if we have manpower to allocate this yet, but if this is a priority for you, I'll make it so. (Alternative: spreadsheets, which Mel is not particularly psyched about, but sure.)
• skierpage, carl, gary, cjl have been particularly helpful in this area, thank you!

People

• Do you have minimum requirements for people contributing to the 8.2.1 (and future) test efforts?
• I don't.
• Big focus of community test: training/teaching newcomers. A lot of testing seems to be learned by people on the fly anyway, let's see if we can step that up.

equipment

• Do volunteers have access to large enough testbeds to do, say, 50 XOs + XS testing? (Interim solution: visit 1cc and use their testbeds, but this is suboptimal to say the least.)
• I am assuming that we will operate out of the testbeds at 1cc for 8.2.1 by default, and that I'll come in / send in local volunteers to use this testbed. Is this ok?
• I count 118 "test" XOs around 1cc. Who is responsible for them now?
• Request: that we get 500 XOs allocated to community testing, in the form of 5 Change The World deployment shipments, one per month for the first 5 months of 2009 (Jan-May). Mchua will take personal responsibility for making sure (or finding someone who you'll be comfortable with making sure) these XOs will be allocated, tracked, and results from the groups receiving them made useful to OLPC in a way that OLPC (Ed, most likely) will be happy with.
  • Usage:
    • 400 of these XOs should seed 8 large-scale dedicated system testbeds (50 XOs per testbed) at engineering colleges or other institutions (in particular, with RF equipment). Minimum of one per continent with a deployment, and one per type of environment (urban, rural, etc.)
    • The remaining 100 XOs can be distributed in smaller (10 XO, perhaps) beds to groups without this equipment but who wish to specialize in a particular type of testing that relies on XOs being present - power, for example.
    • We are close to having groups formally lined up to take the first large-scale testbed on: UWMadison, and possibly SFSU. Oceania would be an easy second spot, possibly followed by a deployment support community in Peru, Uruguay, or Paraguay - note that Madison is solid, but the rest of them are bigger guesses as to where I think we'd start.)
  • Cost to OLPC: If $21,900 is enough to fund shipping to a single Change The World deployment, then $21,900x5 = $109,500 is enough for 5, distributed across 5 months,
• Special super extra bonus: community/travel/shipping funds would be fantastic; even better would be permission to ask for funds on behalf of ourselves (for instance, say Gary wanted to spend a week working at the testbed nearest to him, right before a release - how could this trip be funded?)

coordination with upstream testing

• Fedora (Mel does not know very much about Fedora QA, or if this handoff has been managed yet, or if the responsibility needs to be picked up.) Reference: http://fedoraproject.org/wiki/QA
• Sugar Labs: should we be moving Activity (and other) testing upstream to Sugar Labs (note Sugar Labs does not test; they have a bug squad, see http://sugarlabs.org/go/BugSquad.)

Community infrastructure collab

Work out how to coordinate efforts with Fedora's Community Infrastructure group. Streamlining discussion, testing, and more. How do we build a lasting infrastructure so that local groups can carry out their own education projects / help empower projects in their communities? How can we share knowledge and inspiration across different types and styles of education efforts using Sugar and XO? What is the role for a strong research community in building infrastructure (including funding and visibility)? Specifics:

• community processes that have worked and that have not, effective arms of local organizations (in PR, in event organization, in design, in school or uni outreach)
• what engagement programs are most important? OLPC has contrib programs but doesn't send swag kits. Have the XOs for Fedora groups been useful? Would more be? Is sticker famine fatal? And more.
• how should we follow up with ongoing project maintainers? surveys, emails, pairing up, mentoring.
• how should we make best use of Summer of Code and other regular interns? Division of labor and effort; what is upstream and what is not.
• Roadshows : how to show off the best of both Fedora and OLPC with a compact kit and demo materials. Notes from the field -- from Diane, Mel, and more.

--Sj talk 19:24, 5 January 2009 (UTC)

Follow up owner - SJ

Time needed: 90 minutes

Discussion leaders - SJ, Holt, GDK?

Moodle and Content

Above all we need free (CC or equivalent) educational content aimed at primary school students. This content should be deliverable by Moodle (or what is a Moodle for?). It needs a host (school.sugarlabs.org or ...). We need to energize the community to contribute. Possible contributors (university faculty and students, deployments, G1G1 owners, ....). Possible sources (existing Moodle courses repurposed, currwiki, connexions, Open Learning, WGBH).

Time needed: 60 minutes

Discussion leaders - Tony Anderson and interested volunteers

Moodle and Content [discuss building host for open educational resources deliverable via the schoolserver and Moodle - Tony Anderson]

List of Presentations for Download

• File:NepalDeploymentnotes.odp Deployment Notes for Nepal - Bryan Berry
• File:Startupv1.odp How to startup an OLPC deployment in your country - Bryan Berry
• File:Karmanotforhackers.odp Karma is not for Hackers - Bryan Berry
• Diagram of OLPC roadmap - Marco/Bernie's talk (drawn by SJ, digitized by Brian)