Jump to: navigation, search

Dict | Journal | Repos | Server | Awards | Books | People · Notes +/-

Repository notes.

stories | math | science | reading | language

A few hypothetical numbers quickly gathered about the repository:
Available books* 1357
Number of pages 149330
Number of available modules 5665
Number of authors  3254
Number of registered contributors  42
Total size of distributed files  75.25 TB
Number of distributed files	99911
Finished BitTorrent downloads	555666
Data transfered with BitTorrent  2.01 PB
Number of languages**  17
Number of reviews   28023

* of at least 17 pages
** with at least 4 pages


What does the repository look like?

The OLPC repository [OR] will have three main tiers:

  • Tier 1 : Material that ships preinstalled on every laptop. Within each country, 100-400MB of material will be selected for installation at the factory on the laptops. Children and teachers can change this when they receive the machines; but this material will have the widest reach, and will influence the out of the box experience for new users. Material in this tier -- software and texts -- will all be available under free content licenses.
  • Tier 2 : Material that ships on every school server. This material should be reasonably free.
  • Tier 3 : A national tier within each country, with all materials that are available for free to that country's laptop users.
  • Tier 4 : A global tier combining the shared free materials from all national tiers.

What is useful/usable content?


native speakers


  • Museums (slow changing, l to xl items)
  • Libraries (med changing, s to l)
  • News collections (fast changing, xs to s)
  • Publisher archives (texts, s to m)
  • Stock archives (individual items, xs to s)

  • GR, &c

=== games ---

  • mac emulator, nes emulator, also c64, other
  • ps :
  • sokoban (gen)
  • set (c)
  • checkers [1-2], chess [2], go [2]

By country

Static content

  • Blackbox 1: demo pack
  • Blackbox 2: country-specific reference pack
  • Blackbox 3: manuals, bootstrapping (interface, OS hacking; accessorizing)
  • Texts (with links to national standards/requirements as needed)
  • Modules (indexed by subject, author, source-pool, date, metadata-size, various ratings)

Extant static repositories

Creative content


  • Kibbitzing
  • Metadata editing : rating, other
Tools planned : unknown; general chat, not stored as metadata
Tools needed : filesharing & metadata viewer, w/ md-editor


  • Text editing
  • Image editing
    Special: for photo manipulation
  • Sound editing
    Special: for human voice, sounds from nature, instrumental music
  • Animation
  • Video
    Special: for video shot by hand, in natural light, of people and natural and manmade objects
    Special: manipulating metadata of video and audio without changing large raw datafiles
  • Overview/summary tools for the above (thumbnails, previews, fingerprints)
  • See also #Publishing
Tools planned :
  • chat editor. chat-graphics. some default previewing?
  • text editor. previewing? equations, other embedding?
  • image editor. history? sharing? thumbnails?
  • csound. special filters/editors?
  • helix-based vid editor? no special anim editor, no metadata-based splicing tools
Tools needed : high-level previews for many different views, audio/video thumbnailing, photo/audio/video filters for standard uses


  • Server- or shared-versioning-based text/image/media editing
  • History: one app to provide history to many other apps (journal on top of the above), or patches to each authoring/viewing tool
  • Metadata: trust/experience/shared-use networks
Tools planned: wiki server option; similar option? for editors of other media; wiki-like version history for all journaled content
Tools needed: metadata tools - social networks, programmatic metric analysis, simple authorship & tagging, rating/blessing by groups (as above)


overlay with authoring and collaboration tools; also dark content below

  • Sharing/publishing tool for the above (general journal sharing extended to all filetypes)
    View : Publishing history, use tracking
  • Licensing tool for publishing/viewing the above (off by default)
Tools planned: publishing flags as part of each editor?
Tools needed:
  • Content managment : shared privs for publishing process (consider a private 3-person effort which can be published only by consensus, or by one of them in particular)

Dynamic content

Transactions: through what facility?


  • Continuous inputs : time, location [network and geo], sensors [external and laptop-internal]
  • Tools for queries at a distance: of sensors, databases, other machines, people
  • Simulations


  • Server-bsaed (see also 'collaborative authoring', above)
  • Passive feedback (over the network; aggregation of default-shared properties)
  • Asynchronous feedback
  • Semi-synch feedback (chat, simultaneous editing)
  • Synch feedback (tug of war, network mapping, 3D sound/photo creation, joint simulations)

Dark content


Note network/sharing/privacy dependencies.

Capacity of author to change privacy settings.

6-mo, 2-yr, 10-yr backups.

Compression, prioritization.
'On'line, offline
Feedback/tracking : where are backups going?

Default and DIY backup solutions.

Right to vanish? Who gets to authorize deletion?


Network/sharing/privacy dependencies.

Default use, capacity of author to change/delete revisions.
Right to vanish? For shared content? For shared and collab-edited works? Who authorizes?

6-mo, 2-yr, 10-yr revisions.

Compression, reduction over time
Distinguishing content and metadata in revisions

curr notes

1) free storage/streaming/transcoding, only one format. --> streaming flash and one download format...

  • we need to add ogg transcoding process for the world

2) audio/video capture from the browser --> talk to viditalk? find a way to make this happen over the long term.

3) Uploading archives of any form

  • IMS may require membership to be able to use their membership... common cartridge too? weird.
  • Curr doesn't limit what forms are allowed; review process can help formalize these collections. implications for curation.

4) Making static bundles of a collection -- at the 'print' point in the process.

  • Post the result to
  • Generate a permalink to the resulting bundle;
    SCORM learning objects: use the same manifest structure
  • Or work with Curr to generate a "publish to .xol" module inside
  • This would be the first render-to-package option; might want some logic to avoid things that won't render well on the laptops.

5) two tracks: 1. find a way for users to uplaod a package; so it can be downloaded in the right way. 2. work with an intern to develop a plugin and rendering extension for curriki.

6) figure out Java stories.


  • basic metadata, type, &c. database-backed website
  • community comments and ratings
  • webform for submission by templates
  • akamai/planetlab replication notes
  • simple localization/ontology hooks : track linkage b/t objects
    special-case versioning, summaries, and language translation
  • upload options: mtowtdi
    upload details: directory
    write email to this address; attach doc and put title in subject?
    conversion of one data format to many others.
  • connect: red-forgers.