Hello there,

Hope you all are doing fine.

For the past few weeks, we've been working on the SCORM implementation for OpenACS and there has been some interesting progress and Carl has encouraged me to present to you some of the steps we have taken.

SCORM can be a bit tricky, but just to make sure you fully understand what we do and why it is like that and not like this, have a look at this small document I put together: "SCORM 1.3 for Random Strikers". This document simplifies the specs and puts it in a really colloquial language (you've been warned). If you read this small doc, the rest it should make a bit more sense. if you still have questions, please don't hesitate to ask me.

As I mentioned in my last posting, we have divided the project in to three parts:

* Import/export functionality:  importing IMS and SCORM compliant courses and content aggregations.

Progress: we have finished the import functionality. We can import IMS and SCORM Content Aggregation compliant courses as well as all their million of fields of metadata. So SCORM and IMS packages can be extracted and their manifest files (XML) parsed and imported. We have used tDOM since some of these XML documents can be quite large... and tDOM parses like a beauty. However, we couldn't use XPath which would have been really cool (and easy) due to a problem with namespaces and different interpretations from various content authoring tools on how to create these XML files. Although, the way we have done it, it is the safest way. It took some time but it works really well. Export, supposedly should be the inverse case... (so I hope We've been talking with Caroline Meeks about exporting dotLRN content to IMS Content Packages so they can be imported later into OCW. However it seems that specifications haven't been yet finalized yet. We hope we can help Caroline here since we are going to have an export functionality for SCORM and IMS in any case.

* Managing courses within OpenACS

Progress: the first part is a sound design of how and where we are going to store the learning objects and their massive amount of metadata. We currently are considering the content repository (CR) to store: content aggregations (courses), SCO/As (learning objects), assets (cluster of files or one file), and files. Each one of them would be a content type and would work in the following way:

A SCORM root folder (SCORM Repository) where all the content aggregations would live. A content aggregation would be a content type that serves as a container for SCO/As. By the same token, SCO/As are containers for assets, and assets contains files. So I think this would look like this

SCORM Repository (SCORM root folder)
   |
   |--- [Content Aggregation]
             |
             |----- [SCO/As]
                         |
                         |---> [Assets]
                                    |
                                    |---> [files]

So in this way we have content types defined as:

Content aggregation
SCO/As that have an associated parent (content aggregation)
Assets that have an associated parent (SCO/As)
Files that have an associated parent (files)

All these content types share the same sc_metadata_table (SCORM metadata tables) that contain all their metadata attributes. Files are the only physical entity (the others are containers) and we will store them on the file system somewhere under the Content Aggregation ID folder (/mounting point/SCORM-repository/@cont_aggre_id/folder/etc..) Then we will have to work on the index.vuh to see how we are going to source these files, but that is going to be a different issue that is being discussed here.

We still are trying to figure out a couple of things that have to do with the reusability of learning objects in different courses. You see, a learning object (SCO/A)  might have to be shared with another courses (content aggregations)... so an SCO/A might belong or (be linked?) to two or more different content aggregations. Example:

SCORM Repository (SCORM root folder)
    |
    |--- [Content Aggregation: DB design]
            |
            |----- [SCO/As]
                        |-> [Chapter 1: Using the "Select" statement]
                                |
                                |---> [Assets]
                                            |
                                            |---> [select.html, select.jpg]
                                            |---> [from.html, from.jpg, where.jpg]

SCORM Repository (SCORM root folder)  
    |
    |--- [Content Aggregation: OpenACS for random strikers]
            |
            |----- [SCO/As]
                        |-> [Chapter 1: Intro to OpenACS]
                        |-> [Chapter 2: What is SQL]
                        |-> [Chapter 2: Using the "Select" statement] (LINK or ?)
                                    |---> [Assets]
                                                |
                                                |---> [select.html, select.jpg] (link?)
                                                |---> [from.html, from.jpg, where.jpg] (link?)

So now we are studying how we are going to link the SCO/As to several content aggregations without causing many problems or minimizing scalability issues.

If you want to have a look how the courses or content aggregations are packaged here are some good examples.

• Finalize export functionality
• Integration tool of SCORM course into dotLRN classes and communities
• Managing courses tool
• Runtime environment (including delivering content based on the learner -user- preferable locale, of course if the course support his/her locale -Tils idea)
• IMS Simple Sequencing Engine (hopefully from Polyxena)

I have to thank my Indian-Austrian fellow Tils , and my Mexican/German friend Denis for their help on this. Please feedback and comments are more than welcome.

Thank you,

Ernie

It's good to see that you're going public with your plans for and knowledge about SCORM, Ernie. These are important (and complex) matters that we'll want to get right - with a little help from our friends; the TAB, the UAB, and the community at large. And they're exciting matters, too. Standardized handling and sharing of content is really cool. Looking forward to helping you out in any way we can, dude.

What specifically is it we all want to see here? A new SCORM package? Or a complete rewrite of File Storage so that it becomes a full-fledged front end to CR, handling the "SCORM Repository" (storing, importing, and exporting) plus other content? This issue now appears to be part of the (not quite so radical) revamp of File Storage that Carl has called for. Don has suggested that File Storage may have to be rewritten from scratch. What gives?

The upcoming OpenACS4.x/dotLRN Curriculum package will eventually fit into this picture by providing sequencing - the engine plus the setting up of the tree of activities with their particular metadata (what could be called "course management" if that term wasn't already used for the file storage of SCORM content). The UI for this latter sequence (or curriculum or course) design part will benefit from the valuable input of über course designer Michael Feldstein, we hope.

"Welp, that about does her, wraps her all up."

This is fabulous. We are following your work with keen interest. After the dotLRN v1 release, I hope we will be able to focus more on your work, curriculum and polyxena's simple sequencing engine.

Al

It may be of interest to those who have followed this thread ...

That's a really good idea. Originally I was thinking it the other way around... hehee

Anyway, if you can point me to (or send me) a Content Packaging PIF file from one of these MIT course, I can start mapping it out. As part of the SCORM implmentation, I have gotten IMS content packaging and metadata to work quite acceptable now, so that should be cool.

Ernie

could you send me the package to import IMS CP?, I would like to install it in OpenACS 5.2 to make some tests.

Thank you

Dario (University Of Valencia)

great to hear that...looking forward to test it. Will you be able to come to Heidelberg?

Greetings,
Nima

> Will you be able to come to Heidelberg?

I'll give it a shot, but I can't promise it. I'm working in a way to finance it... more later.

Ernie

First of all, thanks for your time!

Testing your Random package, it works well
but when I try to view the files (extracted of
your example of course) I get the next error.

Any pointer?

Best regards,
Agustin

************************************************
Database operation "0or1row" failed (exception NSDB, "Query was not a statement returning rows.")

ERROR:  -20000: Invalid item ID: <NULL>
CONTEXT:  PL/pgSQL function "content_template__get_path" line 7 at return

SQL:

        select
          content_item__get_live_revision(content_item__get_template('3282525', 'public')) as template_id,
          content_template__get_path(content_item__get_template('3282525', 'public'),'-200') as template_url
        from
          dual



    while executing
"ns_pg_bind 0or1row nsdb0 {

        select
          content_item__get_live_revision(content_item__get_template(:item_id, :context)) as template_id,
.......
************************************************

I found a tool called zms which is a scorm compliant environment for zope. It is open source and might be useful.

Greetings,
Nima

Let me see if I can address all your comments here.

Importing webCT packages to .LRN

I have done some testing importing webCT 4.1.x IMS packages into .LRN.  However, there's a lot of tweaking required as their export fuction has proven, at best, fairly poor. For start webCT 4.1.x doesn't allow you to export full courses (unlike their previous version 3.x!). Let that aside, you can only export two types of content: quizzes and content modules.  Moreover, in several ocassions I have had packages missing files (although they were described on the organization of the package -imsmanifest.xml file- as existing!

Unfortunately, it does not export users information, neither course settings, which is quite a limitation.

At any rate, if most of your courses in webCT have a lot of content modules, you'll be able to export those and then -hopefully- import them seamlessly into .LRN using the LORS libraries.

Atutor

I haven't had much exposure to Atutor, although I have read it has a strong support for IMS specifications and good authoring tools as well. It export content in according to SCORM 1.2 specs (and it seems that they adhiere scrictly to them -which is good), so you can import those easily with the LORs libraries. Although, take into account that only exports content. It won't export forums,

Moodle

I'm not so sure about its compliance to standards. Here (http://en.wikipedia.org/wiki/Moodle) it says that it complies with IMS QTI.

I have tested Moodle before and I have to say I was impressed in many ways with it. First, it is a debian package, so you can install it on a fly. Second, it is a system that revolves around the student and his/her experience: therefore it is quite intuitive. For example, their discussion forums have a picture of all people posting messages there. I'm not a learning or community guru here, but that gave me the impression of a nice community feeling.

Actually, I would encorage people in the community to have a look at it, as it is always good to see other perspectives and approaches.

Random and LORS package

Random, as I mentioned before, it's just an example on how to use LORS libraries. I would strongly suggest you not to use if for any production purposes without previously tayloring. They can both be found here:

Random:
http://mlweg.pc.ee.usyd.edu.au:8080/files/view/20040309_random.tgz

LORS:
http://mlweg.pc.ee.usyd.edu.au:8080/files/view/lors_v01.tgz

I hope that helps,

Ernie

Thank you for your comments. I got yours lors/random package and I'm going to install them. I'm reading about Adam Ullman RTE implementation now.

I'm going to install atutor to see how lors works on importing content exported from it.

Moodle is still a little confuse to me about SCORM implementation. At moodle.org I saw that version 1.2.1 offers support to import quizzes in webCT format but they say nothing about SCORM (see: http://moodle.org/doc/?frame=features.html). There's a company at moodle.com that offers enhances moodle with  support for SCORM packaged learning object at charged (about US$15000).

Anyway I will continue reading about SCORM and dotlrn implementations. Maybe me and people in the company I work could help you and community in the brief future (at leat at user case level).

Orzenil