[Server-devel] XSCE Sprint

[David Farning]

I apologize to Tony, George and any others confused by this thread. I
am still learning how to balance my various roles and responsibility
to the project and ecosystem.

In the back ground, Tony and I have been having a separate thread
searching for areas of mutual value between XSCE and the deployments
he supports. I hope that he and I can continue that discussion and
achieve consensus among the various stake holders.

In the mean time, we are have this public XSCE thread searching for
way to get started. In this case a good implementation strategy seems
to be picking high value services and/or content and making it
available on XSCE. In this way, Tony and other deployments can pick
and chose between the various available content and services to build
a stack which meets their needs.

The things George was asking about about are three strategic pieces
which fill holes in the current XSCE  project:
1. epath library system: George is talking about the server side
service which distributes the content to students. XSCE is currently
working with Pathagar and Internet In a Box. We would like to see what
synergy we can achieve by working together.
2. english language content: Our ability to grow XSCE as a project
depends on our ability to 'show value' to students, teachers, and
deployments. Off line learning content is the single biggest way to
show that value.
3. "schools": Once we get beyond content, the single most requested
feature is the ability to keep track of the relationships between
students, teachers, classrooms, and schools. The  learning curve for
tools like moodle and schooltool are a barrier to their adoption for
many teacher, schools, and deployments. The simple django web app you
use looks like a nice step toward meeting classroom management
requests without become overwhelming.



On Fri, Jul 12, 2013 at 1:24 AM, Tony Anderson <tony at olenepal.org> wrote:
> Hi,
>
> I think this should be done in the overall context of XSCE as proposed by
> David Farning. I think of what I am doing as a system and not as isolated
> pieces. The ds-backup is independent because it only addresses backup and
> restore of the Journal. However, this is going to become more a system
> element as deployments turn to the shared model. It may become moot, if the
> community abandons dependence on Sugar.
>
> Tony
>
>
> On 07/12/2013 03:39 AM, George Hunt wrote:
>
> Hi Tony,
>
> When you sent me your ds-backup script to migrate student datastore to the
> server based upon the "favorite" star in the journal, I downloaded the olpc
> repo, and added your version as a branch, and uploaded it to
> https://github.com/georgejhunt/ds-backup/blob/ds_on_xs/client/ds-backup.py.
> This is a branch which I called "ds_on_xs", but which could just as easily
> be called "tony's ds-backup".
>
> If you are interested, I'd like to create a repo at github for any of the
> following: (can't do everything at once):
>
> epath library system,
> english language content,
> schools, a django application to keep track of students and teachers
>
> And then we can all have access to it and make changes to separate branches,
> and contribute to one another's code.  If you'd like, you can have your own
> github account (they're free), or I can give you shared
> access to the repo that we create together at the github.com/georgejhunt
> account.
>
> George
>
>
> On Wed, Jul 10, 2013 at 8:32 AM, Tony Anderson <tony at olenepal.org> wrote:
>>
>> Hi,
>>
>>
>> On 07/10/2013 02:07 PM, David Farning wrote:
>>>>
>>>> Pathagar is based on Django. The digital library on the school servers
>>>> in
>>>> >Rwanda and Lesotho is based on the same technology but supports any
>>>> > item
>>>> >with a recognized mime-type. The issue is how to organize the contents
>>>> > so
>>>> >that it can be easily accessed.
>>>
>>> Are these open source projects? Can you send links to project code so
>>> we can learn from the approach or include it directly
>>> in XS? The plugin structure enables us to run multiple libraries.
>>>
>> Django is open source. I have sent you copies of the scripts with install
>> Django. Django is organized by applications - it provides a framework to
>> build an application.
>> The basic application is called schoolsite (this is sort of a master
>> application that handles the interface to Apache and to the other
>> applications). The library is handled by the 'library' application.
>>
>> Essentially the library content is organized into collections. A
>> collection is a set of media files (library items), a folder of thumbnails
>> (e.g. the first page of a pdf), and a json file (books.json). The json file
>> provides title, author, path to the item, and mime-type, and path to the
>> thumbnail. A script in the library application loads the collection (i.e.
>> puts the books.json information in the database). The library is accessed by
>> urls (e.g. http://schoolserver/library/ for the home page). Clicking on a
>> category in the home pages goes to a topic page. A button on the topic page
>> goes to a list of items (show 9 per screen). A click on the item, downloads
>> it to the XO and installs it (activity) or puts it in the Journal (pdf,
>> etc.).
>>
>> Logically, the library includes the Wikipedia (Wiki4Schools) although that
>> is not a Django application. It also includes Wiktionary which is based on
>> Mediawiki (and currently not working because of the switch to PostgreSQL
>> from MySQL). The Django content consists of the Sugar Activities from ASLO,
>> the English pdfs from E-Pustakalaya, and a large collection of Old-time
>> Radio and Classical music (Musopen) audio files.
>>
>> I am not sure about the comment about plugins. The current model is to
>> install XS-0.7 to obtain a running server (with the two configuration
>> scripts which should be eliminated), a deployment-specific xs-custom script
>> (which installs the usbmount script, for example). The content is installed
>> from a hard drive using the usbmount script from the booted server.
>>
>> My sense is that the deployment really needs to determine what content it
>> wants on the server particularly since the available content is approaching
>> a terabyte. We need software (api, application) to enable this to be done,
>> but the process will need content specialists more than software engineers.
>>
>> I can supply you with scripts and some rudimentary documentation at the
>> Django level and a sketchy index of what is in the content, if that helps.
>> It is very similar to Pathagar except that Pathagar seems to intermediate
>> between the internet and the server where this application mediates between
>> the servers and the XO clients.
>>
>> Tony
>>
>
>



-- 
David Farning
Activity Central: http://www.activitycentral.com