Flash For Learning

At work we’re upgrading from an older, sunset-ed version of our LMS to the latest service pack of the more recent version. We’ve been going through the upgrade for Vendor X for the past eight weeks, and up until last Thursday it was like a dream — nary a glitch to be found. We started the move on our Production server to make the switch last Thursday night, and that is when our tune took a sour note.

We’ve been experiencing a host of content issues — not SCORM communication issues, mind you. Content. Like, the first time you launch a course, everything works just right. If you close a course and it bookmarks your progress, logging back in (depending on the content) you get the interface, but you don’t get the actual content. In some courses, this breaks the content outright and there’s no way of getting it back.

At first, we were (I was) thinking it was some weird kind of security issue where it was treating Flash content recently run as a security risk and locking it out on subsequent sessions — but this was quickly dispelled, since multiple users could log in after the initial blockage and see all the content the first time they would experience it in their own login. Then, we thought that the scope of it was contained inside our firewall, so we started to divert traffic to our external content server — but that, too, started showing the same issues. So while it’s not ruled out as a factor, our network environment isn’t the only factor involved.

Currently we have a theory that there’s a security setting that’s scrubbing data sent to the LMS by content, escaping characters and such. And it’s doing that on content that is, on its own, escaping characters (like what Articulate does). And that may be underneath whatever Tomcat or IIS is doing to filter data for malicious strings. So, perhaps with all this filtering of the data, something breaks down when the feed comes back. We can trace the string coming back on consecutive data transactions with SCORM content and that’s definitely going on — pipes (|) are being sent the first time. Then, when they’re coming back, they’re URL-encoded. Then it looks like the URL encoding is changing the Ampersand (&) into something else… so it’s being URL-encoded in one place and then that URL encoding is being treated as a straight-up string by something else.

Which for you non-technical types — is un-good.

Everyone is working very diligently to figure this out, but I predict quite a few help desk calls today (I don’t get them, but I know they come). And with members of the team having invested so much time and effort, it’s a drag to see the snags come in so late in process.

I’m pretty confident that everything will be fixed — it’s only a matter of time, as are all things digital. But it’s a drag, and that is no doubt. It’s not a matter of finger-pointing — it’s not Vendor X’s slip-up or anyone’s negligence — this is just a point of pain that comes with enterprise software, especially since enterprise software generally has to be configured specially for an enterprise’s unique needs.

I know I’m not the first (nor will I be the last) to witness it, but I feel horrible for other members of the team who have put in so much time — only to find out they’re going to put in a lot more time

I recognize for many of you, SCORM 2004 Sequencing and Navigation is impossible. My guess is that for most of you, it’s a jumbled problem that combines the complexities in understanding how to create a manifest that correctly interprets the intent of your instructional flow — and the problem of how an LMS actually interprets your sequencing instructions.

I’m pretty lucky to be able to narrow down the problem, because I spent many long days and nights side-by-side with Angelo at ADL, who’s the Godfather of Sequencing and Navigation (though I’m sure he’s loathe to go down in Wikipedia history for that effort in lieu of his Level 70 Shaman in WoW). In other words, my problem isn’t generally writing an effective manifest with Sequencing in it. All I want to do for my company right now is create a three-item tree for Pretest, Content and Posttest.

My problem is our LMS. I won’t name it, but it rhymes with “BaBa.” It’s a certified product that we’re upgrading to. It’s certified to SCORM 2004, 2nd Edition. I import content from ADL to run in its content administration. I can see from the debugging window that its handling the data model well (enough). But in between it redrawing the Table of Contents every time I navigate from one SCO to another and (what I think is happening with) the interpretation of what the rule is on the active node in the activity tree — I’m not being presented with any content as a result of passing or failing an objective. It’s recording the SCO itself as being passed or failed in the resulting transcript. But it’s not flowing on.

And that sucks. Bad. It makes me miserable. I have felt and feel for people dealing with these issues in our armed services and in all branches of government, where the decision is made to go with Vendor X for an LMS but the person who has to make the content work in the LMS has no say in the selection.

We talked at great length about this at the SCORM Technical Working Group meeting last month. It will be discussed even more. There was talk at the time about the various vendors actually getting together with each other and figuring out how to share the interpretation of what a manifest is instructing a system to do. If you care about this… which is key to “interoperability” that many people are looking for… make it known to whoever your LMS Vendor is via their customer service or sales contacts with you and yours… and let them know that the next time you renew or upgrade, a factor in the decision to spend money with them is going to be how well they’re playing with others.

They won’t come to the same table by themselves, but they will if it means holding onto and growing the investment in YOU as a paying customer.

That’s my rant for the day.

EDIT: That’s not really fair of me to say at all. There were a handful of representatives from LMS Vendors large and small at the TWG. Several of them were very willing to “come together” and commit to working with a shared set of content packages as test cases to tweak their implementations to behave the same way. I won’t out them either at this point (I don’t want to be counter-productive). My main point is that if they’re going to do this, as customers we can hold a great amount of sway by encouraging this activity — either with the carrot or with the stick.

You could hear a pin drop this entire Learning 2007 conference if you were trying to have conversations about LCMSs. In other words, no one was talking about it other than me. Lance Dublin, whom I admire a lot, led a session on implementation and lumped the LCMS and LMS in the same bucket, but everyone was talking LMSs. LCMSs? Nada, zilch, zip, nein!

So, by a nifty mistake, I ended up in a Learning Consortium meeting instead of the 10 Years of SCORM meeting I had planned to show up to. I sat with Judy Brown and Rovy Brannon from the ADL Academic Co-Lab at University of Wisconsin, and at 8:30 in the morning on the last day of the conference, it became a very intimate open discussion with Elliott Masie, instead of the guided activity to collect themes for the next year of the Consortium.

I addressed the elephant in the room and asked Elliott why no one was talking LCMSs this year when last year we were beaten over the head with it. His reply was honest and maybe a bit surprising.

Elliott told me that an LCMS purchase right now has about a 30-month lifespan, because the real shifts in the tools learning organizations will use will be in powerful Talent Management systems and powerful Content Systems (not solely learning content or authoring tools), because there is a shift more and more towards immediate knowledge, which means that the traditional e-learning as we do it will be supplemented more and more with our broad spectrum of documents (excel, word, powerpoint stuff within what we currently call a document management system). Elliott suggested this is because the search and retrieval features of document management will significantly improve.

Now, Elliott did tell me that it’s not a bad idea to turn on an LCMS because of the workflow benefits we can gain and the way in which it, as a tool, will support standard quality of content… but his caveat was that we should know going in that in three years or so it will probably be outmoded because of the leaps in both technology and the necessary shifts in talent management that will make the top-down management of content too time-taking and too laborious to do. In other words, “rapid” will get “rapid-er.”

So as “learning content” changes in its form, its authors are going to be spread throughout the organization. I think an LCMS is useful for the reasons Masie described, but in the planning for my organization, I need to organize the change management issues related to shifting our Instructional Designers into learning content producers and then, eventually, learning content specialists consulting with the rest of the organization which will do the authoring. That’s a very distinct set of change issues that is related but not necessarily coupled with change management dealing with the LCMS.

When I think about this distinction, I think the LCMS isn’t such a big deal for my organization to handle… and the change around the roles, communication chains and workflow in our future are going to be much more difficult if we’re not very clear about the change we want to create and understand the impacts on all of us and the clients inside the company we serve…

Projects Underway:

• LETSI – Beyond SCORM
  • Natural Evolution
  • From ADL
  • Adoption since 1997 has been enormous
  • Things were not designed back then to scale the way they need to now
  • Truly Global stewardship of SCORM
  • Taking over the responsibilities from ADL to be able to scale that support up
  • Charter, Governance, Infrastructure in the works now
  • Big things to emerge in 2008
  • Masie Consortium has given $10K to support LETSI, representing 250+ commercial organizations
  • Tom King is Masie’s delegate to LETSI.
• 3D Learning Project
  • Masie has paid to develop custom Second Life objects (islands, rooms, etc)
  • The “Compuserve” of virtual worlds
  • Not ready for corporate life, but there is a lot of innovation
  • Only Consortium members can come on to the Masie Island
  • Assets can be reused for free by any of the Consortium members
  • Interests in using Second Life as a narrative-based learning environment
  • Common thread — many of the people in organizations whose job is to research nextGen stuff has to do it from home, because nextGen stuff (Second Life, Facebook, Twitter) is blocked from corporate firewall.
• Voice of the Learner Survey
  • Looking for 10,000+ participation
  • Quick beta in November
  • Will aggregate member data for the member itself (separately) — for free.
  • The goal is that this happens every year.
  • Question this should answer: How is Learner Preference changing?
• Mobile Learning Project (iPhone+)
  • Judy Brown and Tom King
• Coaching & Mentoring
  • Stretch Assignments (Critical element in development)
  • Little scalability
• Performance Support
  • Interface issues
• Visual Front End of LMS