I’ll offer some alternatives here…

SCORM 2004. That’s right, this is one of the areas where SCORM 2004 is profoundly better than SCORM 1.2. It includes a full text question description and a descriptive identifier for answers. This means that the you can start to do some really effective reporting on questions and the distribution of answers. It’s a dramatic improvement. (Now, if SCORM 2004 looks like too much of a hassle for this benefit, you should check out the SCORM Cloud, which already has that problem solved.)

IMS QTI. IMS QTI is a specification from imsglobal (who contributed greatly to SCORM) for question test interoperability. It is not a silver bullet, but, it is on the right path… allowing for some of the banking, etc, that you seek. Further, QTI is starting to gain some traction. It’s a part of IMS’s broader spec IMS Common Cartridge, and I think it’s possible it may find its way toward SCORM sometime in the future. No promises there, but it’s something to consider. (At this point, we don’t have support for QTI in our products…)

So, options, but nothing perfect. SCORM walks a fine line with regard to specifying a lot or a little. And there’s no doubt that SCORM has specific issues… Feedback like this helps us and, more broadly, the community, effectively evolve to the standard to serve more effectively.

You can access the kind of the question, the correct response, the student response and the score, but not THE ACTUAL QUESTION. This puts a big limit on the role that the LMS can perform.

For example, it can’t create a reusable question bank of all the SCORM questions, you can’t browse the questions from the LMS side of things or perform a high-level analysis over the set of all questions asked by all the SCOs, etc.

Maybe I’m missing something, but it seems like a very basic weakness with the SCORM spec. It would take very little to send the actual text of the question along with all the other data that is sent from the SCO.

Cheers,
Graham

Most of the course packages I currently develop do not take advantage of Sequencing. For maximum compatibility across LMSes, I purposely code to features common to SCORM 1.2, SCORM 2004, and AICC-HACP. I’ve developed a translation layer that detects the LMS environment and uses the appropriate protocol to communicate with the LMS.

As a content developer, I find that SCORM does what I need, and does it well. I’ve encountered a lot of issues like those that Antonia mentions. SCORM leaves a lot open to interpretation, and within those gaps is where I find I need to tweak. Over time, I’ve encapsulated these tweaks into “configuration options” that enable me to deal with quirks of various LMSes, but hardly a week goes by that I don’t encounter some interesting new twist. SCORM could stand improvement. The approach is more than a decade old — an antique in software years. But SCORM was definitely a huge leap forward. I think the folks who developed it did a great job of bringing some semblance of order to the chaos.

I come from an LCMS background and need to work with various LMS. I find that the interpretation of the specification varies widely. Is this a problem with the vendors or is this a problem that the spec just does not define what should be done in an unambiguous way.

Week 1.
LMS does not support cmi.interaction data model elements

Week 2.
LMS requires cmi.exit to be specifically set to “” if the SCO is to be resumed

Week 3.
LMS allows a learner to re-enter a passed course and, if they browse the assessment, they get a failed assigned.

Week 4.
LMS sets the lesson status to completed when the SCO’s browser window is closed and not the correct status as when the SCO’s ‘exit’ button is clicked.
…

My personal opinion on sequencing is that its utility is largely linked to the reuse of SCOs… which is clearly an issue of merit. When there are great, reusable SCOs, that value probably merits the headache. But the lack of reusable SCOs makes head hurt.

The obfuscation via Get and SetValue is a valid thought as well. In some sense, the API is more complicated than the 8 methods. If you drill into our SCORM Driver documentation, you’ll see that we set aside the methods available in the spec and use “English-sounding function names” to great effect.

So, thanks for the real/valid complaints… you can pick on my little brother anytime.

First off, apologies are due to Dave, and I offer them here and now. He and I had a brief twitter exchange (more). I failed to include mention of his responses, as I read them to be in line with his initial comments. It is my mistake, though, to include only the first post based on my interpretation. Dave certainly doesn’t deserve credit as “Opening Villain”… the conversation alone should have earned him a far more ambiguous role… Opening villain probably should have gone to @frdsrcks, who was completely unambiguous.

Two meaty questions here in my book:

Is the balance of “time required” to “level of functionality” appropriate in SCORM?
Is it poor form to blog about/quote a previously public interaction (in this case, twitter)?

Dave and I clearly differ on both questions. I’m certainly interested to hear your collective thoughts on both issues.

Lastly, I hope to hear from more people like Dave. One of our core philosophies is that the best idea wins. I hope this post brings not only those folks who agree with me (likely followers of the blog) but those who disagree strongly (like Dave). And if Dave is interested in continuing this conversation here publicly, getting into the details of SCORM, etc, great. If not, I hope I haven’t offended him too much.

Since we do one large sco (in this particular deployment because 2004 is not supported) that has many lessons and tests we never set the status to failed. We keep it incomplete and give them the ability to retake the scored sections to get a passing score then set the course to complete. This avoids them having to go through all the content.

Both parts are following the standard but different expectations of course design conflicted. What I find often is the client has an expectation of perfect scorm support in their LMS’s and i have to walk them through the fact it’s more of a compatibility thing and less than a standards thing.

Saba had an issue with exit value and how that affected the sessions count it displayed. Another LMS displayed the score and the client always expected us to display some kind of score-they looked at it more as a progress value than a score. I’ve had similar experiences with all the lms’s I’ve come across.