The Three Kinds Of Reuse

A CONFERENCE ROOM FULL OF PEOPLE

R: Welcome back, all. I think we're in for a treat tonight. All the talk of reuse recently has led me to believe that it's finally time to define what me mean when we say the word "reuse."

C: This sounds like the kind of exercise academics just love.

D: Sounds like murder by brutal boredom.

O: Not so fast... remember how productive the conversation we had about the definition of a "learning object" was?

(Lots of laughter)

R: Let's begin with something concrete. V - when your organization "reuses" a learning object, what do they do?

V: Well, I mean... there's not a lot to say. During the instructional design process, certain blocks of language, images, and sounds may be called for. Before beginning production of any of these materials, the development team looks through our catalog to see if any materials like those specified already exist. If they do, we just reuse those rather than creating new materials.

S: You use metadata to facilitate the search, I assume?

V: We do for images, audio, and video. But our experience has been that full-text search is much more useful than metadata-based search for materials that are text-based.

S: Really?

R: Let's come back round to the topic, please. Any other thoughts?

C: I suppose we're trying to something similar. I mean, my manager's goal is to eventually completely automate the reuse process - have it driven by an intelligent system - but at its root its the same. You find an existing object that meets your need and drop it into the instructional sequence.

O: Actually, I think that you're talking about two very different things. I believe V said he tries to find some material "like" the stuff called for in the design spec.

C: Yea. So what's so different?

R: I think the difference is what happens next. Let's hear about the next step in detail.

D: What detail? You find the object that meets your needs and drop it into the sequence of other objects...

R: V? What do you do next?

V: Well, more often than not, we end up tinkering with the thing to make it fit into the course better.

S: What do you mean by "fit better?"

V: Well, if we develop some training for the Army, and then build the same training for the Air Force, it would be nice to have people in the pictures wearing blue uniforms instead of green, right?

C: (To self, worriedly) Oh boy...

R: What's wrong, C?

C: My manager's not going to be happy to hear about this. He's got it in his head that you can develop learning objects that a computer can just reuse.

D: There's no reason you couldn't. You just have do to one of a couple of things: either leave out the little details like pictures that are going to come back and cause you problems later, or have some way of marking up the things that are going to need to change and changing them on the fly.

C: How do you mean?

O: They're not little details - they're context! I believe we agreed that these little details are what make the content really speak to people, instead of sounding like an encyclopedia.

D: Anyway, I can imagine setting variables in the text, and then dynamically inserting values and media based on who was getting the content. For example, using a tag like &amp;lt;!--armed_forces_branch_name--&amp;gt; and then inserting Army or Navy on the fly, depending on who was accessing the content. You could do things like &amp;lt;!--armed_forces_branch_name--&amp;gt;_uniform.jpg, too.

C: But that will only work as long as we reuse for the military, right? What happens when we're doing something for GM or Proctor and Gamble?

O: Were we supposed to be differentiating between MNCs and the military?

R: Come back, O!

D: Then just use something like &amp;lt;!--customer_name--&amp;gt;. It would obviously take some careful thinking, but I think you could do something like this if you put your brain to it.

R: Actually, this is very similar to the approach that Gibbons advocated in The Nature and Origin of Instructional Objects. He calls it the "semantic string" approach.

D: Huh. There really is nothing new under the sun, is there?

(Laughter)

C: Good try, though.

O: I want to come back to my point though, if I may, that there's something very different about dropping an object into a sequence and having to "tinker" with it first. (I can't believe I'm saying this, but) The economics of it are certainly different.

R: There are some people who refer to this difference as "reusing" versus "repurposing." I would highly recommend Derek Morrison's Innovations in the Reuse of Electronic Learning Materials - Drivers and Challenges. He does some very interesting talking about reuse and repurposing.

O: It's a lot like Lessig's talking about rip, mix, burn.

R: Yes, Derek uses that phrase, too. It's a meme that seems to be making its way around the internet.

V: Yes, Google knows of about 51,000 occurences of the phrase.

C: Does this mean that the idea of learning objects is finally becoming more mainstream?

O: No, it just means that the marketing people at Apple are great.

S: Also, you really should keep in mind that most of the people in the standards world are thinking about reuse, not repurposing. They're much more interested in the type of automated reuse C was talking about.

O: Maybe that should tell you something about why the learning objects idea hasn't been catching on...

D: Oh, the idea is catching on, it's just that no one has been able to do anything useful with it yet.

(More laughter)

R: Ok, ok. I actually think there is something a little more interesting in the idea of repurposing than just the economics of it. Obviously it's more important to open objects before reusing them...

V: (Cutting in) Which should lead to a cost/benefit analysis of how much it costs to tweak these little things versus how much the learning is improved.

C: And an analysis of the oportunity costs of not being able to scale like you want to if you have to touch each object each time it gets reused.

R: (Continuing) We started to go there a little with C's talk of the army and air force, and D's reinvention of semantic strings. O said that these "little details" are really the context we spoke of earlier, and that context is responsible for making the content speak to the learner. Does there come a point at which it costs less and / or is more effective to just make something new than to re-engineer an existing piece of content?

O: Steve Carson, from the MIT OCW, recently said, "I don't think the 'localization' process will turn out to be as simple as swapping out a few images and culturally inappropriate metaphors any more than translating a novel is a process of looking up words in a dictionary. Each stands as its own authentic creation." You're talking problems adapting stuff from one branch of the US military to another. Steve's talking about adpating MIT course materials for use by students in Africa.

S: Are you arguing that it would be easier for you to create material - from scratch - for use in Africa than it would be to repurpose existing material?

C: I think the point is that it isn't possible for him to create materials appropriate to that local context. He isn't African, doesn't know the cultures, doesn't know the languages, etc.

R: Of course he wouldn't be the one doing the re-contextualizing. Someone from something like African Virtual University would. But, unlike what has been said so far today about switching pictures from Army to Air Force, Steve sees instructional materials as being artistic wholes.

S: It's almost as if the idea of a lesson comprised of pre-existing parts is Frankenstein-ish.

O: Well, it is, isn't it? If you don't bother to adjust things so that they fit together in a pleasing way.

R: I want to stop for a higher level comment momentarily. Even though we've identified two activities that we're calling "reusing" and "repurposing," we seem to be talking across two kinds of repurposing that we haven't clearly articulated yet. One is adjusting a learning object so that it will "fit" with the other objects in its aggregation. A second kind is adjusting a learning object so that it will "fit" with the learner who will be using it.

C: Shouldn't they have names? You know, so we can be clear about which one we're refering to?

R: (Grinning) You're obviously an academic at heart, C.

D: How about "repurposing for alignment across objects" and "repurposing for alignment with learners"?

V: But what kinds of things do you change in each case? Are they different things? Seems like it would be the same stuff you change, you're just trying to align to different targets.

R: For example?

V: Well, if all the objects in an aggregation but one are written at a sixth grade level, and the last one is written for undergradutes, then you're looking to repurpose for alignment across objects by changing the level of writing. If you turn around next week, and try to reuse the aggregation with a group of undergraduates, you may change the writing level across all the objects again, but this time to better align the content with your learners. In both cases, you're changing the academic level of the material.

O: So, if I could disagree with R for a minute, they aren't two kinds of repurposing, they're two reasons for repurposing. And it may be that you do exactly the same kinds of things in both cases.

C: But if most of the objects had one kind of interface or one type of navigation, you would update the others for "alignment with objects." But would you ever change the navigation for "alignment with learners"?

V: You would if, for example, you felt like navigation-style was something that significantly impacted learning. For example, if instead of a strict sequence through materials you wanted to give another group of learners a more exploratory interface into the collection.

D: I was actually at a conference a few weeks back with Steve Carson, and overheard him say he thought there were three types of repurposing that might have to be done to materials like learning objects: cultural, academic, and technical. He didn't say this, but I think calling it ACT is a nice acronym.

S: Definitely better than calling it CAT would have been.

R: Well this conversation is evolving quite the little list of ways materials can be reused or repurposed. There is one more major piece of the reuse conversation waiting for us, but I want to stop here for tonight and carry the conversation on next time. Let me try to capture what we've said so far tonight about reuse and repurposing:



V: That captures it pretty well, R.

R: Thank you. I know I'm not an artist, but sometimes you just want to see a diagram, you know?

C: So are there no summary questions tonight, then?

R: Oh, no. There are plenty of questions. Fire away...

V: What is the value of personalizing / repurposing learning objects relative to the direct costs of adapting each object and the opportunity cost of being able to scale up to huge number of learners if you don't have to touch each object?

R: Ok.

C: Let's not just talk about costs to the developer. What are the impacts on learning if one doesn't repurpose for alignment with other objects? Like the extraneous cognitive load created by dealing with different navigation schemes?

S: Whoa. That was actually a very intelligent quesiton, C. At least it sounded intelligent.

R: Agreed! I suppose one could ask the same question about materials not tailored to specific learners, but that isn't really a learning objects question. I expect there's a body of research available on this question if anyone wants to dig around for it.

O: How about a question about the types of adaptation that should be done? Do academic, cultural, and technical considerations really cover the bases?

D: And, what I thought was the most interesting question of the night - is there ever a time when it's just cheaper or faster to develop your own materials, rather than repurposing an existing object?

R: Alright. Any others? So far we've got:


 * 1) What is the value of personalizing / repurposing learning objects relative to  the direct costs of adapting each object and the opportunity cost of being able to scale up to huge number of learners if you don't have to touch each object?
 * 2) What are the impacts on learning if one doesn't repurpose for alignment with other objects?
 * 3) Are there types of repurposing other than academic, cultural, and technical adaptation that we should consider?
 * 4) Is there ever a time when it's just cheaper or faster to develop your own materials, rather than repurposing an existing object

R: Ok. If not, then I'll see you next week! 5656456222