Monday, December 17, 2007

Technology used to create web pages

I think people are talking about technologies without looking into the using them and even more importantly, having learners access the pages created by them.

Presently HTML is king when it comes to a language used to create web pages. I think HTML's limitations has some advantages, specifically making it accessible to more people. for example this is a an easy way for SMEs to create web courses. Actually,
HTML by itself isn't used by SME's, but authoring tools that
create it does. Then, there are other tools like Google Docs,
FrontPage, & DreamWeaver that can easily edit what the authoring tool has
created (if there is a need for that).

XML is being used more often for display of content, and that is
fine. Likewise, it is being used more frequently for data storage, and
that too is good. However, simultaneous use of an XML document for both
display and for data storage seems extremely difficult and what would
be the point of doing this?

XSLT (style sheets and display for XML) is and will be very useful for what it is intended to do. I don't know what tools actually do the XSLT translation?
For example, browsers can display HTML and XML, and different parsers
can read/parse XML. If you use an XML parser, you still have to program
the interpretation and use of the parsed data, and that is fine because
now there is a more accessible file storage format. With XSLT, what
creates the output file?

Friday, December 14, 2007

LMS update - what they can can & can't do

Based on my last post I got some questions on that status of LMS's. Some of my information is a little dated, but here is what I can bring up-to-date.

1. Aspen: SumTotal has several LMS offerings including "Docent", "Voyager", and "Aspen". Of these, I have only seen that "Aspen" has the interactions and objectives support.
2. IBM: Yes, they keep changing names. The last time I saw things with their system was about 2 years ago. I think what I saw later became Lotus Learning Space, but this is conjecture.
3. Oracle: The Oracle iLearning, which is now "OLM" would be the most complete offering to look at from Oracle. I only recommend Oracle for companies that already have other Oracle financial systems already installed. As with any LMS, the cost of the LMS is only a small fraction of the total cost of ownership. The bulk of the cost is the integration of the LMS's database with any existing personnel system. If Oracle is already available at the organization, then adding their LMS is cost-effective. My experience with the other systems you mentioned that are now owned by Oracle is that their data support for the eLearning standards were minimal at best.

I have seen many other LMSs (including other "big" ones) that were purposely omitted from my list because they do not support for interactions and objectives. In some cases, delivering interactions even caused the LMSs to throw exceptions. Unfortunately this occurred on some well known brand-name LMSs.

As you observed, the "big" LMS vendors seem to have less to offer than some of the smaller ones. My experience has been that some of the smaller ones have much better technical support, better technical implementation, and better customer support. Among the big ones, Oracle impressed me, but that was at least 3 years ago. For those of you who would like to know my qualifications to make these statements, I work on the LMS integration for ReadyGo. We have an authoring tool that creates AICC or SCORM conformant packages. In the course of doing this, we have integrated with dozens and dozens of LMSs. What we have seen is that each LMS has their own interpretation of the specifications, especially with AICC. With SCORM, there are fewer interpretations, but there are still behaviors and limitations imposed in the LMS that can affect the learner experience. We have made ReadyGo open enough that it is possible for us to create "LMS-packs". These are analogous to printer drivers. When you go to generate (print) your course, you can choose what LMS or specification you want it to work with like you would choose a printer. This allows the course to report as much information as possible to the LMS without causing the LMS to interfere with the learning experience. For example, some LMSs are set up so that once the student completes the course, they are not allowed to take it again. When customers don't like the one-time-only use of courses, we can set up the LMS-pack so that the course never reports a completion status. Then, the LMS doesn't block the learner from re-using the content. The learning level of the student could be passed, for example, through the score. So that is why I feel that I can provide my opinions on LMSs.

My greatest frustration has been that most LMSs and, as a result, most authoring tools have gone for the minimum necessary to be able to put "SCORM Conformant" on their sales brochures. You can see this when the authoring tool only offers one "AICC" and/or one "SCORM" output option. Course developers have then had to rearrange their courses or just forget about tracking anything more than course completion. This has crippled the true capability of SCORM and AICC, and has resulted in "junk food" courseware as the norm.


Lots of LMS's and authoring tools limit their system by their limited reading of the SCORM specification and their limited implementation of SCO's. There is nothing in the SCORM specification that says that a SCO must be a single page of content. In fact, I believe it is a bad
design decision to make each page a SCO, but this appears to be the
group-think way of doing things. Here is my rationale: When you go from
SCO to SCO, the LMS has to close the previous SCO and then Launch the
next one. In some cases this means that the user must return to the
table of SCOs and manually launch the next one. That completely breaks
the learning flow.

Imagine that every time you want to see a new page of Google results you
have to go back to the start page, and re-input your search. That would
chunk your learning experience down into little pieces. And, we haven't
even addressed the delays most LMSs have in closing one SCO and
launching the next. I have seen delays as big as 20 seconds when going
from unit to unit, and this is on a DSL line.

If you want to see a tool that uses chapters or entire courses as a SCO,
you can take a look at the ReadyGo Web Course Builder.

One of the reasons that the "group-think" has gone with page=SCO is that
the majority of LMSs don't capture or report interactions and
objectives. This means that you can't get granularity of reporting
unless you have granularity of content. That is to find out if someone
answered a specific question correctly, you must use question=SCO.
Technically, it is easier to implement a single question on a web page,
but instructionally, I wonder if that isn't worse. In school, do
teachers hand out a single question, then pick it up, grade it, tell you
your grade, and then hand out the next one? I usually learned the
answers to one question by understanding another question on the test.
Since the objective should be training/teaching rather than measuring,
wouldn't it be better to use the pedagogical methods that have been
refined over centuries? Yes, we have new technologies available, and
these afford new opportunities, but it doesn't mean that we should
jettison older methods just because some 20-somethings with their cool
new iPods walking around with headphones believe they are the first to
ever do this, and thus anything the above-30s do must be trashed and
dis'-ed. (Does that stand for disregard, disparage, disagree, or all of
the above?)

Another tragedy of the minimalist LMS approach is that it becomes
impossible to use the LMS to carry out surveys and assessments. The
good news is that there are LMSs (Avilar, Oracle, MeridianKSI, IBM,
Aspen) that do capture and report interactions and objectives. So now,
there is hope that your LMS can be used to evaluate your course both
from the point of view of figuring out if the instructor is giving bad
questions and for the student to let you know their thoughts.

So, if you let a SCO be more than a single page, you can have the
summary as part of the SCO, and have the navigation/reusability, without
making compromises.

Monday, December 10, 2007

My rebuttle from an online article

In the 13-February-2007 issue of OnlineLearning News and Reviews, Dr.
Patti Shank discussed how the "best" authoring tools for certain
organizations may be tools that convert Microsoft Office content to
web-viewable format. I believe that this approach is part of the
natural inclination to avoid/delay change, because change requires
learning new ideas. It is like early cars that used reins to steer.
MS-Office is an excellent suite of tools to create/manage data specific
to those tools. Word is good for creating letters and documents to be
printed. PowerPoint is good for creating talking points that help a
live presenter. Excel is excellent for storing/manipulating tabular
data. However, we don't use Excel to give presentations, or PPT to
create printed documents, although we certainly could.

Likewise, using PPT and Word to create on-line documentation is a misuse
of these products. Most people have these tools, but yet, there are
almost no web sites that are built using these tools. So why should
users settle for web courses that are built using these tools? The
answer appears to be that course authors find it convenient to do so
(irrespective of the learners' needs.)

I actually see that in the future we will be using web courses as the
backup documentation when we give live presentations. There are
tremendous advantages to doing this. A properly structured web course
gives you links on each main content page to additional resources.
(PPT does not.) So, in the middle of the presentation, if an audience
member asks you a question, you can go to one of these links where you
may have a simulation, step-by-step layout, or in-depth article about
the question they have asked. Of course, you wouldn't read the in-depth
article, but it may have a table with statistics to back up your points.

In terms of visual display (the main reason people use PPT), through the
use of style sheets you can make this content format nicely for multiple
purposes. Ahead of time, you can create several style sheets that work
best in different delivery settings (e.g. laptop computer, PDA, visually
disabled). The end-users can click on different links to the same
content. Each link can use a different style sheet. So, users
following along on their own computer can choose their own content
layout. This is especially useful if you are giving the live
presentation to remote sites.

After the session, the complete web course is still available for the
audience to review. If they didn't understand one of your talking
points, they can look at the links to more data, and find how you've
explained the points in more depth. By giving multiple presentations of
the same material, you can then improve the chances that they'll be able
to catch your concepts. Further, the author can give tracked test
questions, and can find out directly if the learners are understanding
the concepts.

Even more than this, users can come back to your content and they don't
have to go through the entire presentation in order to get to the pages
of interest. This means that the content can be used as reference
material as they are trying to accomplish a task ("just-in-time"
learning). Course authors have been concentrating or the term "re-use"
only from the authoring side of things. We really should be considering
re-use from the learner side, as the purpose of the training is to
facilitate the end-user rather than the course author.

New mechanisms like XML are just the underlying technologies that in the
end will allow the kind of re-use and shifts that are mentioned above.
So rather than hanging on to our old tools trying to find convoluted
ways to use them in the changing environment, content developers would
be well-served to start thinking about how the new capabilities for
content delivery will modify how they do their tasks, and what tools are
really "best" for them, and more importantly, their audience.

Thursday, December 6, 2007

Building an effective learning community

Here's a sampling of the ingredients one needs to build an effective learning community.
  • A good story. You need to have the basic ingredient of a community, which is a common interest shared by many people. Start with a good set of content (courses, discussions, resources) so that people have a foundation upon which they can build and participate.

  • Self-interest. There has to be a benefit for the participants, and that benefit has to be self-evident. Why should people cooperate and spend time working on the "community" when it distracts them from doing the 50 million other tasks that are assigned to them?

  • Critical mass. The community will only be self-regenerating once you have enough people actively engaged. Getting to this point is terribly difficult, and maintaining the activity can be difficult. One trick Moser has found to be helpful is to cause controversy. "This gets people more interested. The hard part is finding controversial subjects that are not insulting. For example? Discuss why you think one approach to solving the problem is better than another."

  • Ease of use. The slightest hurdle for people to participate either as contributors or as data receivers will turn people off from the experience. Stick to "best of breed" Web practices (e.g., easy navigation, fast download, staying away from heavy multimedia, using an effective search engine, and avoiding clutter). "You may need librarians and moderators to ensure that the site doesn't get cluttered with 15 versions of everyone's documents and that the discussions don't get hijacked by a few individuals."

  • Culture. If there isn't a culture of sharing or of informal learning in your organization, you won't be able to create one just by using a social network software site. "People have to feel that there is no downside to participating in the activity. Management has to support it. (Good luck!) And management has to participate, also."

  • Luck. This is by far the most important ingredient, says Moser. "You can have all the ingredients, but there is that intangible thing about luck that probably comprises 60 percent of the deciding factor as to whether the site will mushroom, die, or hobble along. But just because it hobbles along doesn't mean you should give up. Luck can come along at any time." A new person may join the organization and spice up the discussions, for example, prompting people to be drawn in. "Why is YouTube more popular than Yahoo! videos or Google videos? Mostly luck. There were many other video sites that were comparable and available at the same time. YouTube got lucky."

Monday, December 3, 2007


If your goal is to train engineers, my recommendation, as an engineer, is to stay away from the "more interactive" (also known as "more entertaining") approach that uses video, audio, and hot-zone mouse-overs. Contrary to what has been advertised, video and audio are actually very passive forms of training. We have seen that if you include an audio in a course (as an adjunct to slides), the learner often abandons the visual portion (since they are receiving audio) and drift to check their e-mails or other tasks they need to do. The result of this is that their focus on the learning is reduced, and eventually lost. Only 10% of material is learned through audio. Keep this in mind.

Engineers have been trained to be able to learn from visual content (primarily math and physics with lots of equations). We (engineers) are accustomed to both quick learning (by looking up a formula) and in-depth learning (understanding how the formula was derived). A video can provide an example of the concept, but will generally not provide fundamental engineering concepts.

True interactivity involves changing behaviors based on the student's response. The simple form of this is to provide the learner with multiple expositions of the same material accessed through links on a page. These expositions could include a step-by-step table, a link to a journal article, a formula derivation, a practical example of the implications of the formula, a quiz, a tracked test. The more complex form of interactivity (which is much more expensive) is a true-to-life simulation, such as Flight Simulator. People should really question what is interactive about a hot-zone mouse-over. If you are training people how to move a mouse, this is a good simulation; however, if the purpose is to teach some other concept, it becomes "eye-candy" (and can be very distracting). Every graphic, audio, video, hot-zone, or other non-text element should be carefully scrutinized and justified. Does it convey new information? Does the student's action relate to the content, or is their action simply a display control?

My recommendations: A good "book" layout will be most effective at training engineers. An instructionally sound "book" approach is more than a series of linked slides with a table of contents on the sidebar. Each "page" of content should include links to sub-pages with different expositions of the same material, since each learner will gain differently from each presentation. The material should be easily accessible, meaning that at any time when the employee is doing their job, they can use the "course" as reference material to look up the procedure or formula. If the course is designed as a one-time, linear set of content, you lose this possibility. If you can put a search engine on your site, and the content is properly searchable, you will create a re-usable resource that the engineers will quickly adopt as part of their "library" of knowledge.