(Note: I haven't thought about this very hard).
In technical terms...
Current metadata approaches assume a 1:1 relationship between a 'learning resource' and an (implicit) 'educational activity' (because the activity itself isn't modelled explicitly).
It is therefore possible to take an existing DC metadata record that uses some learning-specific properties and split out those properties into a model that does explicitly separate the 'learning resource' from the 'educational activity'.
There would be no ambiguity in doing this.
So I don't think we have a huge technical problem - though the domain of some of the DC-declared properties might have to be changed.
Our problem will be in getting widespread agreement/understanding of the more complex model - particularly from any developers whose mindset is that (current LOM is good enough).
Andy
--
Andy Powell
Research Programme Director
Eduserv
t: 01225 474319
m: 07989 476710
twitter: @andypowe11
blog: efoundations.typepad.com
www.eduserv.org.uk
> -----Original Message-----
> From: Stuart Sutton [mailto:[log in to unmask]]
> Sent: 11 February 2010 14:19
> To: Andy Powell; [log in to unmask]
> Subject: RE: Functional requirements and the model
>
> Andy, the modeling is now clear to me when combined with your message
> of a few minutes ago around the "have been used" language [1]. So,
> what do you see as the consequences of such a modeling on legacy data
> (and there is a lot of it)? All of what might be previously considered
> DCAM conformant data would be coming from that place you speak of where
> the "educational activity" is implicit.
>
> Stuart
>
> [1] http://tinyurl.com/y9xaftp
>
>
> > -----Original Message-----
> > From: DCMI Education Community [mailto:[log in to unmask]]
> On
> > Behalf Of Andy Powell
> > Sent: Thursday, February 11, 2010 5:38 AM
> > To: [log in to unmask]
> > Subject: Re: Functional requirements and the model
> >
> > In modelling terms... or rather, in terms of how I am modelling the
> > world in my head :-) a 'learning resource' has an "intended learning
> > time" only because the resource creator had a particular 'educational
> > activity' in mind when they created it. I.e. the 'intended learning
> > time' is always a property of the 'educational activity', not of the
> > 'learning resource' itself. The problem is that the association
> > between the 'learning resource' and the 'educational activity' is
> often
> > (nearly always?) implicit (but it is, nonetheless, a real
> association).
> >
> > A 'learning resource' with no associated 'educational activity'
> (either
> > implicit or explicit) is just a 'resource' (and should be described
> as
> > such - i.e. without the use of any learning-specific properties).
> >
> > All learning-specific properties are actually properties of the
> > 'educational activity', not of the resource itself (even where that
> > 'educational activity' exists only in the head of the resource
> > creator).
> >
> > For me, the issue at hand is...
> >
> > Do we want to explicitly model the association between a 'learning
> > resource' and its associated 'educational activity/ies' OR do we want
> > to leave that association implicit (as it is with current metadata
> > approaches).
> >
> > The advantage of explicitly modelling it is that the model can then
> > cope unambiguously with situations where a 'learning resource' is
> taken
> > away from the 'educational activity' that the original creator had in
> > mind and used in the context of a completely different 'educational
> > activity' (with different target audiences, levels of difficulty,
> > learning time, etc.).
> >
> > The disadvantage of explicitly modelling it adds significant
> complexity
> > to the model - which may well end up confusing the hell out of just
> > about everyone!
> >
> > Note: I've created complex models before, SWAP springs to mind :-),
> and
> > gone on to see them used by almost no-one... so I'm under no
> illusions
> > that this is not a real issue.
> >
> > Andy
> >
> > --
> > Andy Powell
> > Research Programme Director
> > Eduserv
> > t: 01225 474319
> > m: 07989 476710
> > twitter: @andypowe11
> > blog: efoundations.typepad.com
> >
> > www.eduserv.org.uk
> >
> >
> >
> > > -----Original Message-----
> > > From: Stuart Sutton [mailto:[log in to unmask]]
> > > Sent: 10 February 2010 22:32
> > > To: Andy Powell; [log in to unmask]
> > > Subject: RE: Functional requirements and the model
> > >
> > > Andy, I'm not sure I would agree. The semantics for "difficulty"
> in
> > > LOM state: "How hard it is to work with or through this learning
> > object
> > > for the typical intended target audience" [1] which is quite
> > different
> > > from "How hard it WAS to work with this learning object for the
> > ACTUAL
> > > target audience." If you look at the definitions of nearly all of
> the
> > > DCEd properties (of which I have a more than passing familiarity)
> and
> > > the LOM [1], they are framed in terms of design and intention--
> things
> > > designed for intended use ("intended or useful", "described
> resource
> > is
> > > intended", "intended to take place", "typical intended user",
> > "typical
> > > intended target audience" "approximate or typical time").
> > >
> > > We keep throwing out typicalLearningTime as not applying to things
> > like
> > > lesson plans and descriptions of designed activities etc. but
> rather
> > > being appropriate to apply to an activity instance (some actual
> > event).
> > > I'd note that notions like typicalLearningTime and typlicalAgeRange
> > are
> > > quite different from actualLearningTime and actualAgeRange that
> would
> > > adhere to an activity instance where there is no longer the
> 'typical'
> > > but rather the 'actual'.
> > >
> > > It does not seem to me that we need new properties to talk about
> all
> > of
> > > these millions of resources I noted--not given the semantics of our
> > > existing properties. If anything, we need new properties for
> > > describing an activity instance. That's the new kid on the block.
> > >
> > > Stuart
> > >
> > > [1]
> > >
> >
> http://dublincore.org/educationwiki/Existing_20DCMI_20Education_20Prope
> > > rties
> > >
> > > > -----Original Message-----
> > > > From: DCMI Education Community [mailto:DC-
> [log in to unmask]]
> > > On
> > > > Behalf Of Andy Powell
> > > > Sent: Wednesday, February 10, 2010 1:48 PM
> > > > To: [log in to unmask]
> > > > Subject: Re: Functional requirements and the model
> > > >
> > > > Agreed... maybe.
> > > >
> > > > But we come back to the central problem... namely that very few
> so-
> > > > called learning objects have an inherent 'difficulty level' and
> > even
> > > > those that do can be used in different ways which means we have
> to
> > > > associate properties like 'difficulty' with an educationalUsage
> > > rather
> > > > than with the Resource itself.
> > > >
> > > > I suppose we could define properties like 'intendedDifficulty'
> with
> > > > definitions like 'the intended level of difficulty, as envisaged
> by
> > > the
> > > > creator of the learning resource' but it seems to me that would
> be
> > a
> > > > significantly less useful property than something like
> 'difficulty'
> > > ??
> > > >
> > > > Andy
> > > > --
> > > > Andy Powell
> > > > Research Programme Director
> > > > Eduserv
> > > >
> > > > t: 01225 474319
> > > > m: 07989 476710
> > > > twitter: @andypowe11
> > > > blog: efoundations.typepad.com
> > > >
> > > > www.eduserv.org.uk
> > > > ________________________________________
> > > > From: Stuart Sutton [[log in to unmask]]
> > > > Sent: 10 February 2010 17:47
> > > > To: Andy Powell; [log in to unmask]
> > > > Subject: RE: Functional requirements and the model
> > > >
> > > > Andy, here we hit upon rough shoals because limiting use of these
> > > > education properties (DC and LOM) to resources "that HAVE BEEN
> > USED
> > > as
> > > > part of educational activities" eliminates 99% of all the
> resources
> > > of
> > > > interest to the community to which these properties have been
> > > applied-
> > > > millions of resource descriptions.
> > > >
> > > > Stuart
> > > >
> > > > From: DCMI Education Community [mailto:DC-
> [log in to unmask]]
> > > On
> > > > Behalf Of Andy Powell
> > > > Sent: Wednesday, February 10, 2010 8:33 AM
> > > > To: [log in to unmask]
> > > > Subject: Functional requirements and the model
> > > >
> > > > I was just taking a quick look at the functional requirements
> > > > (http://dublincore.org/educationwiki/Functional_20Requirements).
> > > >
> > > > As written, I think we have a problem with these in terms of the
> > > model
> > > > (http://dublincore.org/educationwiki/Model).
> > > >
> > > > We currently say things like:
> > > >
> > > > Support the discovery of learning resources and activities
> targeted
> > > at
> > > > particular levels of difficulty.
> > > >
> > > > What I think we should be saying (in terms of the model) is:
> > > >
> > > > Support the discovery of learning resources that have been used
> as
> > > part
> > > > of educational activities targeted at particular levels of
> > > difficulty.
> > > >
> > > > Andy
> > > >
> > > > --
> > > > Andy Powell
> > > > Research Programme Director
> > > > Eduserv
> > > > t: 01225 474319
> > > > m: 07989 476710
> > > > twitter: @andypowe11
> > > > blog: efoundations.typepad.com
> > > >
> > > > www.eduserv.org.uk
|