Content formats for the XS and XO

[Martin Langhoff]

SJ gave me a heads up recently, and quite timely, about the need to
discuss what formats we can support on the XS and the XO.

My high level goals are that

1 - Local teams are able to include enormous amounts of gorgeous
content coming from untold number of sources on the XS and XO. The
teams handling deployments get to pick and choose content from many
sources repositories repositories.

2 - The content is in such a format that we can serve it from the XS
to XOs (and other clients) and we can also provide it as a bundle (or
similar) so you can download it to the XO and take it with you.

3 - What we do is interoperable with the rest of the world as it
exists. (ie: use established standards)

In general, this means that we need content to be in the simplest,
most standard formats possible. Project Gutemberg clearly gets this.
ASCII, HTML, PNG, SVG and DublinCore metadata are my friends.
Super-simple package formats based on them (think IMS-CP) make me
happy.

The XS is going to grow a "browse-and-download library content" leg
sometime soon. It will be a facility that knows how to unpack and
index the IMS-CP and SCORM metadata (which is all DublinCore stuff),
and it'll be transparent to users how to browse, search and use any
such content placed on the XS.

It's about the content. No special client software, no special server
software. This levels the ground for all the projects -- so you don't
need to worry that your outrageously fantastic content won't be used
because the player ain't sexy. Local teams don't have to worry about
the burden of a particular content requiring special software.
Teachers don't have to worry about the confusion of a different
content player.

My plan is to make that XS feature so that it can also automagically
make a bundle for the XO to download -- after all, happiness is a warm
bundle (  http://blog.laptop.org/2009/01/13/happiness-is-a-warm-bundle/
) :-)

This also means that XS and XO can leverage the large existing pools
of IMS-CP and SCORM content. And that any content you prepare for
OLPC, you are also preparing for the world...

That's my plan at least :-) -- no plan ever survives contact with reality.

cheers,




m
--
 [hidden email]
 [hidden email] -- School Server Architect
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff
_______________________________________________
Library mailing list
[hidden email]
http://lists.laptop.org/listinfo/library

[Seth Woodworth-2]

I would add speex as a codec for human voice content.  I can pack about an hour worth of audiobook into ~7mb using speex, and have it fairly nice.  Gstreamer on sugar plays .spx files just fine.  They are an ogg varient and developed by xiph.

Ideally we could use the 1500+ hours of Human Read audiobook that Librivox has produced and stored on archive.org in some of our deployments.

--Seth

On Sun, Feb 1, 2009 at 10:21 PM, Martin Langhoff <[hidden email]> wrote:

SJ gave me a heads up recently, and quite timely, about the need to
discuss what formats we can support on the XS and the XO.

My high level goals are that

1 - Local teams are able to include enormous amounts of gorgeous
content coming from untold number of sources on the XS and XO. The
teams handling deployments get to pick and choose content from many
sources repositories repositories.

2 - The content is in such a format that we can serve it from the XS
to XOs (and other clients) and we can also provide it as a bundle (or
similar) so you can download it to the XO and take it with you.

3 - What we do is interoperable with the rest of the world as it
exists. (ie: use established standards)

In general, this means that we need content to be in the simplest,
most standard formats possible. Project Gutemberg clearly gets this.
ASCII, HTML, PNG, SVG and DublinCore metadata are my friends.
Super-simple package formats based on them (think IMS-CP) make me
happy.

The XS is going to grow a "browse-and-download library content" leg
sometime soon. It will be a facility that knows how to unpack and
index the IMS-CP and SCORM metadata (which is all DublinCore stuff),
and it'll be transparent to users how to browse, search and use any
such content placed on the XS.

It's about the content. No special client software, no special server
software. This levels the ground for all the projects -- so you don't
need to worry that your outrageously fantastic content won't be used
because the player ain't sexy. Local teams don't have to worry about
the burden of a particular content requiring special software.
Teachers don't have to worry about the confusion of a different
content player.

My plan is to make that XS feature so that it can also automagically
make a bundle for the XO to download -- after all, happiness is a warm
bundle (  http://blog.laptop.org/2009/01/13/happiness-is-a-warm-bundle/
) :-)

This also means that XS and XO can leverage the large existing pools
of IMS-CP and SCORM content. And that any content you prepare for
OLPC, you are also preparing for the world...

That's my plan at least :-) -- no plan ever survives contact with reality.

cheers,




m
--
 [hidden email]
 [hidden email] -- School Server Architect
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff
_______________________________________________
Library mailing list
[hidden email]
http://lists.laptop.org/listinfo/library

_______________________________________________
Library mailing list
[hidden email]
http://lists.laptop.org/listinfo/library

[Martin Langhoff]

On Mon, Feb 2, 2009 at 4:34 PM, Seth Woodworth <[hidden email]> wrote:
> I would add speex as a codec for human voice content.  I can pack about an
> hour worth of audiobook into ~7mb using speex, and have it fairly nice.
> Gstreamer on sugar plays .spx files just fine.  They are an ogg varient and
> developed by xiph.

Sounds like a good idea -

 - if Browse.xo gets served a speex file with the right mimetype, does
it do the right thing?
 - do browsers in other desktop/laptop platforms (conventional linux,
Windows, OSX) do the right thing?

> Ideally we could use the 1500+ hours of Human Read audiobook that Librivox
> has produced and stored on archive.org in some of our deployments.

Great project -- didn't know it existed. Tried it with Alice in
Wonderland, in part to see what my ubuntu laptop would do with the
speex files, but it lists mp3 and ogg only...? Is speex a codec that
gets wraped in ogg?

cheers,


m
--
 [hidden email]
 [hidden email] -- School Server Architect
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff
_______________________________________________
Library mailing list
[hidden email]
http://lists.laptop.org/listinfo/library

[Seth Woodworth-2]

Sounds like a good idea -

 - if Browse.xo gets served a speex file with the right mimetype, does
it do the right thing?

Yes.  It didn't as of 656, but it did as of 711 and does in 767
 

 - do browsers in other desktop/laptop platforms (conventional linux,
Windows, OSX) do the right thing?

*nix yes,
OSX sometimes,
Windows... not in IE and I don't think so in FF 3.1 maybe
 

> Ideally we could use the 1500+ hours of Human Read audiobook that Librivox
> has produced and stored on archive.org in some of our deployments.

Great project -- didn't know it existed. Tried it with Alice in
Wonderland, in part to see what my ubuntu laptop would do with the
speex files, but it lists mp3 and ogg only...? Is speex a codec that
gets wraped in ogg?

They're not yet encoded as speex @ archive.org.  I've spoken to sj and the team of Internet Archiver's that he's been working with and it seems possible that they may re-encode their archives to provide speex as an option.

Your ubuntu machine should already have a speex file on it.  If you still have the Desktop-Examples package installed Ubuntu comes with a collection of flyers and OOo documents, an Ubuntu-Sax.ogg file and a Librivox Aseop's Fables recording in Speex.  

~/Examples/fables_01_01_aesop.spx
/usr/share/example-content/fables_01_01_aesop.spx

_______________________________________________
Library mailing list
[hidden email]
http://lists.laptop.org/listinfo/library

[Martin Langhoff]

On Mon, Feb 2, 2009 at 6:31 PM, Seth Woodworth <[hidden email]> wrote:
>>  - if Browse.xo gets served a speex file with the right mimetype, does
>> it do the right thing?
>
> Yes.  It didn't as of 656, but it did as of 711 and does in 767

Cool.

>>  - do browsers in other desktop/laptop platforms (conventional linux,
>> Windows, OSX) do the right thing?
>
> *nix yes,
> OSX sometimes,
> Windows... not in IE and I don't think so in FF 3.1 maybe

Is Ogg better in terms of existing support? It's been around for
longer, and we hit vid and audio in one go. Librivox is all
Ogg'ified...

> They're not yet encoded as speex @ archive.org.

Hmmm... Speex has been in existence for a while, but it has the
strange achievement of being used less than Ogg ;-)

So clearly the XO supports it. I would still recommend Ogg if the
content producer can support only one format.

> Your ubuntu machine should already have a speex file on it.

Indeed - my machine plays the sample files from speex.org directly from ffox.

So an updated formats list:

 - ASCII
 - HTML/XHMTL/CSS
 - PNG
 - SVG
 - Ogg
 - Speex
 - PDF (somewhat discouraged - editable content is nicer)
 - Flash (discouraged unless you have the resources to ensure it is
Gnash compatible)
 - Editable content - RTF as it is the default for Write.xo . Or HTML.

and should be wrapped in an "IMS-CP" formatted package, which is a
fancy way of saying "a zipfile with a manifest and some DublinCore
metadata".

I have done some work with IMS-CP, and the format is trivial enough...
but the standards documents are horrible. Used to have a cheatsheet on
it -- but can't find it.

The XS will definitely support IMS-CP. No promise to support some of
the fancy and convoluted things it lets you define -- our platform is
mainly interested in

 - metadata to organise, browse and find the content
 - a pointer to the 'index' file if your CP has a big forest of html pages

in other words, we keep it simple, simon.

cheers,


m
--
 [hidden email]
 [hidden email] -- School Server Architect
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff
_______________________________________________
Library mailing list
[hidden email]
http://lists.laptop.org/listinfo/library

[S Page-2]

Martin Langhoff wrote:

> The XS is going to grow a "browse-and-download library content" leg
> sometime soon. It will be a facility that knows how to unpack and
> index the IMS-CP and SCORM metadata (which is all DublinCore stuff),
> and it'll be transparent to users how to browse, search and use any
> such content placed on the XS.
> ...
> My plan is to make that XS feature so that it can also automagically
> make a bundle for the XO to download -- after all, happiness is a warm
> bundle (  http://blog.laptop.org/2009/01/13/happiness-is-a-warm-bundle/
> ) :-)

Sounds great.  .xol bundles and the auto-expanding Library home page are
great but Sugar-specific, so having your server create an appropriate
bundle is worthwhile.  I guess if the server just noticed a request from
a non-Sugar client, it could return a different mime type and possibly
extension, and that would be enough for a non-Sugar client to get an
plain ZIP file.

I just stumbled across http://sugarlabs.org/go/Activities/InfoSlicer 
(from its YouTube video!), which turns web pages into .xol bundles, the
code is
<http://git.sugarlabs.org/projects/infoslicer/repos/mainline/blobs/master/Processing/Package_Creator.py>

> This also means that XS and XO can leverage the large existing pools
> of IMS-CP and SCORM content.

"SCORM", that takes me back to working at a company where the learning
division kept saying "It's a billion dollar business!" (before exiting
that business).  Does SCORM let me work through a course offline or
online, a) taking a quiz to verify I've learned each "learning object"
and b) keeping track of my progress?  It seems those functions are what
a school would want to tack on to existing web-based curricula.  Back in
the day I never got a straight answer whether there was an easy way to
do this on our web site, even though we were selling the software.
Judging from http://www.adlnet.gov/scorm/ , it remains opaque and confusing.

> That's my plan at least :-) -- no plan ever survives contact with reality.

Good luck with your efforts!  http://wiki.laptop.org/go/XS_Roadmap 
doesn't match this rallying cry.

--
=S Page
_______________________________________________
Library mailing list
[hidden email]
http://lists.laptop.org/listinfo/library