Audience proposal for LRMI for Schema.org inclusion

[Greg Grossmeier]

Hello LRMI Technical Working Group members and the wider LRMI community,

As you may know, Schema.org has had a few proposed extensions that
include a term that is similar in purpose as our intendedEndUserRole.

One member of the Schema.org team created this page to describe one
possible way forward:
http://www.w3.org/wiki/WebSchemas/Audience


Below is my proposal for changes to LRMI so that it can be included in
Schema.org cleanly with the new Audience framework.

1) Create Thing > Intangible > Audience > EducationalAudience

2) EducationalAudience type will have a property 'educationalRole'
(expected type: schema.org/Text)

3) remove 'intendedEndUserRole' from LRMI proposal and use
EducationalAudience 'educationalRole' instead


I have updated the w3 wiki page for our proposal:
http://www.w3.org/wiki/WebSchemas/LearningResources

The diff:
http://www.w3.org/wiki/index.php?title=WebSchemas%2FLearningResources&diff=60033&oldid=59830

See that page for the updated example that includes this Audience
markup.

One question for you: Should I include a small list of recommended
values for this term, eg: student, teacher, parent?


I know this is different than our latest version, but I hope that this
change is minimal enough to still be acceptable to our TWG and
community, but also sufficient to deal with the "audience issue" (my
term) that Schema.org is facing.


If there are no substantial blockers with this proposal, I will make an
official update to the LRMI specification on lrmi.net tomorrow,
hopefully with an equally exciting announcement alongside that.

Thanks!

Greg

--
Greg Grossmeier
Education Technology & Policy Coordinator
twitter: @g_gerg / identi.ca: @greg / skype: greg.grossmeier

[Peter Pinch]

> One question for you: Should I include a small list of recommended
>values for this term, eg: student, teacher, parent?

I think proposed values would be tremendously helpful (among other things,
it might avoid another discussion about grade level).

Since it's very likely that real-world content will be intended for both
students and teachers, it would also be helpful if there could be a
recommendation for that situation as well, e.g. "Repeat the field with
different values" or "use a combined value."

- Peter

-----------
Peter Pinch
Production Manager, MIT OpenCourseWare
pdp...@mit.edu
http://ocw.mit.edu

[Greg Grossmeier]

<quote name="Peter Pinch" date="2012-07-19" time="20:21:24 +0000">

> > One question for you: Should I include a small list of recommended
> >values for this term, eg: student, teacher, parent?
>
> I think proposed values would be tremendously helpful (among other things,
> it might avoid another discussion about grade level).

Good point.

> Since it's very likely that real-world content will be intended for both
> students and teachers, it would also be helpful if there could be a
> recommendation for that situation as well, e.g. "Repeat the field with
> different values" or "use a combined value."

Another good suggestion to make it clear on the LRMI spec site that all
terms are repeatable.

[Joshua Marks]

Greg,

There is some complexity here between the EducaitonalAlignment for leveling
framework (e.g. grade levels) and an Audience at some level. Some on the
discussion thread have expressed concern that educational level is not
adequately expressed as an EducaitonalAlignment. As Brandt touched on, there
are terms to describe the intended audience's attributes and related terms
to describe the level of and intended use/users of the materials. Are these
always equivalent?

For example a work of literature written for 3rd grade level in the US might
be intended for an older student with remedial reading level (Sometimes
called "low reading level high interest publications"). So Role might be not
exactly what is needed if level if to be part of the audience definition.

I will also note that the W3C proposal seems to muddy the waters with the
concept of a parent of a child, which describes relatedness of audiences.
Similarly we might have Teacher of a child. I am not sure how the applies or
can be leveraged.

In general, I think leveraging the Audience intangible is a good idea, but I
think we need to be _more_ thoughtful about the learning context extensions
of this Audience tag.

Joshua Marks
CTO
Curriki: The Global Education and Learning Community
jma...@curriki.org
www.curriki.org

I welcome you to become a member of the Curriki community, to follow us
on Twitter and to say hello on our blog, Facebook and LinkedIn communities.

-----Original Message-----
From: lrmi...@googlegroups.com [mailto:lrmi...@googlegroups.com] On Behalf
Of Greg Grossmeier
Sent: Thursday, July 19, 2012 12:46 PM
To: LRMI TWG; LRMI
Subject: Audience proposal for LRMI for Schema.org inclusion

[Zoe]

I agree with Joshua about EducationalAlignment having unaddressed
complexities.

Here at the BBC we're about to launch a completely new product for all
our learning content. This will involve creating new markup for all
our learning content, including but not limited to all our learning
content for schools. So, it's kind of a big deal for us. We're working
on what the markup should actually be right now.

I'm afraid we have come to the conclusion that we can't use the LRMI's
'educational alignment' field.

To be useable, we would need educational alignment to include both
'level' (catered to in 1.0) and 'entity that defines the level' (not
catered to in 1.0).

The reason is this:

In the United Kingdom, we have four nations. They each have their own
national curricula.

The four national curricula use the same terminology, but in relation
to different levels. i.e. The terms do not align.

'Key Stage 2' in Northern Ireland refers to different ages/years of
school than 'Key Stage 2' in England.

Therefore, if the specification does not allow us to specify which
nation is defining 'Key Stage 2', any content marked up as 'Key Stage
2' is intrinsically misleading - it will, inevitably, be accurate for
one group of users and inaccurate for the other three.

Under this circumstance using 'educational alignment' is worse for us
than storing the information in e.g. <h3>

This is just one example. I've tried, and I've not been able to think
of any instances of assumed alignment where this problem does not
occur. (Including level data but not including reference to the
defining entity is by definition assumed alignment.)

The alignmentObject property at http://wiki.creativecommons.org/LRMI/Properties/1.0/AlignmentObject
defines only:

- alignmentType
A category of alignment between the learning resource and the
framework node. Recommended values include: 'assessed', 'teaches',
'requires', 'textComplexity', 'readingLevel', 'educationalSubject',
and 'educationLevel'.

- targetDescription
The description of a node in an established educational framework.

- targetName
The name of a node in an established educational framework.

- targetUrl
The URL of a node in an established educational framework.

- We will not use alignmentType as it is too broad to have semantic
utility.
- TargetDescription and targetName have no value for us without the
ability to specify what the 'established educational framework' is. We
will not use them as they are currently defined.
- We have no plans to use targetURL.

We would probably use alignment if it included (to pull terminology
but not requirement from the air) targetOrigin, where 'origin' is the
entity responsible for the 'established educational framework'.

I am now convinced that creating the impression of level equivalence
(alignment) where it does not exist - which is what the current
specification does - will reduce accuracy so far that using it will
make resource discovery worse.

I also note that, even if equivalence did exist by accident, 'Grade
three history' in Belfast has minimal (if any) content crossover with
'Grade three history' in Vancouver. If the LRMI is to work across
nations it must be capable of capturing nation data. There is no such
thing as a universal curriculum.

Here is a document showing the complexity of level alignments in the
UK. Note that this is for four 'level defining entities' (in this
case, Nations), where in the scope of the LRMI, there are thousands.

https://docs.google.com/spreadsheet/ccc?key=0AsqEpfKQI9UsdFpxSlJfQ0k2aGYxXzhwYWRwZHVKVHc

I don't mean to be horrible, but I really do want to see the LRMI
succeed and for us, this has emerged as a blocker. I don't want it to
become a blocker for other potential adopters.

Cheers, Zoe

> The diff:http://www.w3.org/wiki/index.php?title=WebSchemas%2FLearningResources...

[Jim Goodell]

If the targetURL always pointed to the authority for the framework and the targetOrigin were discoverable, along with all other metadata about the node and framework, then LRMI could work as is. This does however require standards for the sites referenced by targetURL.

[Joshua Marks]

Hmmm, so it sounds like the solution to this challenge is to either extend
EducaitonalALignment, add EducaitonalLevel as a different type of alignment,
or define this in terms of the Audience?

[Greg Grossmeier]

<quote name="Joshua Marks" date="2012-07-27" time="16:53:58 -0700">

> Hmmm, so it sounds like the solution to this challenge is to either extend
> EducaitonalALignment, add EducaitonalLevel as a different type of alignment,
> or define this in terms of the Audience?
>

educationLevel is already a type of alignment.

alignmentType
Text

"A category of alignment between the learning resource and the framework
node. Recommended values include: ‘assessed’, ‘teaches’, ‘requires’,
‘textComplexity’, ‘readingLevel’, ‘educationalSubject’, and
‘educationLevel’."

http://www.lrmi.net/the-specification/alignment-object

[Joshua Marks]

Greg,

Indeed, but it fails for Zoe and for Monty in that the targetURL is not sufficient to discriminate two different definitions for the same educational level (term). They seem to want something like '3rd grade as defined by State A vs. State B' WITHOUT having a URL from those authorities for the leveling framework or some discriminating element in the targetDescription or targetName. Now if every authority publishes a target URL for an alignment node that unambiguously indentified the authority, then I would argue it would be sufficient. But this is clearly not the case, and that, I believe, was Jim Goodell's point. So his note suggest a possible extension of AlignmentObject with an additional targetOrigon (Which might better be named framework publisher or authority). Adding something like this to AlignemntObject might satisfy Zoe et al.

But this misses my prior point... If we are to adopt the Audience proposal, how do things like age of audience and level of alignment match up? And what about Parent's of Child as described in the W3C proposal?

Joshua Marks
CTO
Curriki: The Global Education and Learning Community
jma...@curriki.org
www.curriki.org

I welcome you to become a member of the Curriki community, to follow us on Twitter and to say hello on our blog, Facebook and LinkedIn communities.


-----Original Message-----
From: lr...@googlegroups.com [mailto:lr...@googlegroups.com] On Behalf Of Greg Grossmeier
Sent: Friday, July 27, 2012 8:02 PM
To: lr...@googlegroups.com
Subject: Re: Audience proposal for LRMI for Schema.org inclusion

[Brandt Redd]

EducationalAlignment was set up to manage alignment to "a node in a learning framework." Frameworks include learning standards like the Common Core, curricula like Scotland's Curriculum for Excellence, text complexity frameworks like MetaMetrics Lexiles, etc.

A critical design decision which distinguishes LRMI from predecessors like LOM is that LRMI does not attempt to define or represent the framework only to align to such a framework. The moment you start to define the framework or taxonomy it introduces a tremendous amount of complexity that violates the simplicity of the Schema.org context in which we operate.

Education Levels are no different. Zoe has started to show how complex it is to represent Education Levels. Without the suggested "Target Origin" the same term (e.g. "Key Stage 2") has ambiguous meaning. But this is only the tip of the iceberg. To properly represent a taxonomy of education levels you would also need the expected age range, the day of the year on which the age is measured, rules about stage mobility, and ultimately what is expected to be taught in that stage.

To manage complexity, we chose to reference taxonomies using the TargetURL property rather than attempt to represent taxonomies in some universal way. Given that Zoe and the BBC "have no plans to use targetURL" it's not surprising that the schema fails to meet their needs. They've chosen not to use the one property that's intended to solve this kind of problem.

Attempting to use Audience to solve the problem will ultimately end up with the same question -- do you define an audience schema complex enough to describe Education Level taxonomies or do you reference external taxonomies?

In my opinion, the BBC should simply assign URLs to the taxonomies they find relevant. For example:

http://bbc.co.uk/uri/England/keystage2

http://bbc.co.uk/uri/NorthernIreland/keystage2

If, on some future date, the relevant educational authorities assign their own URLs for the same levels then they could substitute the new ones. Or, the BBC may have enough clout to convince authorities to create URLs like these immediately. Keep in mind that the URLs don't have to point at any real content. They can even return 404 errors. They just have to be unique and someone has to make a definition of their meaning. For an example of how this is done in a different domain you can look at how the W3C uses URLs to identify encryption algorithms. (http://www.w3.org/TR/xmlenc-core/#sec-AlgID). In their case, the URLs redirect to the relevant standards document.

Yes, this is a messy process. In the US, each state that adopts the Common Core State Standards can make up to 15% additions to the standards. And some states (notably California) are changing the standards (which isn't supposed to be done) rather than just adding to them. In addition to the core standards (which are only just barely getting identifiers) all of the state additions and changes must be assigned identifiers. And then there are the states that haven't adopted the common core. Our plan is to use the Learning Registry as a clearinghouse and cross reference for all of these identifiers. (http://learningregistry.org, also see the common core identifier project at http://corestandards.org/developments-on-common-core-state-standards-identifier-and-xml-representation).

Your thoughts?

Brandt

[Zoe]

Hi Josh -

You write:

"Hmmm, so it sounds like the solution to this challenge is to either
extend EducaitonalALignment, add EducaitonalLevel as a different type
of alignment, or define this in terms of the Audience?"

Three options worth discussing, but I'm not sure they quite do the
job.

Just to be *really* explicit about use cases - I'm working on an
assumption that 'if the (person making markup) has to look up the
documentation to know what goes in the field, the field isn't
usable'.

This assumption is based on a) my tech comms background b) my
experience as an English teacher and c) my experience in ASPECT
(Adopting Standards and Specifications project) where, across 20+
participating educational publishers, I saw universal confusion and
misuse of fields in LOM metadata, including for fields that I thought
were fairly self-explanatory (and I don't think educationalAlignment
is self-explanatory, it is very complex). We can rely on our adopters
being experts in content creation, but we can also rely on them *not*
being experts in standards, metadata, or best-practice markup. Folk
who work in educational publishing are very rarely technologists by
inclination, and they need a lot of help. (See also: teacher-creators,
who also make fantastic learning resources.)

'Level' is not going to do the trick as it does not yell 'add an
authoritative entity to me!' If I was a marker-upper, I would happily
add 'Grade three' without specifying e.g. either 'Australia' or
'USA' (and no, these two instances of 'grade 3' do not apply to the
same age ranges or areas of study.)

'Audience' isn't quite right because it's not an audience issue -
there is no universal 'Grade three' which has different audiences. The
concept 'grade three' is dependent on an authoritative entity saying
'Grade three is X' (where X might be 'the school year eight year olds
are in'). That's why we need the authoritative entity explicitly
identified - they are the *only* thing that determines the meaning of
the level.

'extending educationalAlignment' I don't think works either (I'm
making an assumption here that by 'extend' you mean 'add another value
to AlignmentObject/alignmentObject' - please correct if wrong). I
don't think that works because it doesn't meet the requirement of
being self-explanatory. AlignmentType is full of recommended values -
'teaches', 'readingLevel', 'educationalSubject' are all in there. It's
a jumble, and a very confusing semantically undifferentiated jumble at
that. (Can you imagine a creator of grammar games looking at
alignmentType and saying 'oh I see, of course 'requires' is in the
same place as 'textComplexity', that makes perfect sense'? I can't.
They're more likely to say 'this is weird... I don't get it. I don't
think I'm going to use this standard' and walk away.)

So I do think that, same as my earlier post, we still really do need a
dedicated field to specify the educational authority that defines the
level.

On Jul 28, 12:53 am, "Joshua Marks" <jma...@curriki.org> wrote:
> Hmmm, so it sounds like the solution to this challenge is to either extend
> EducaitonalALignment, add EducaitonalLevel as a different type of alignment,
> or define this in terms of the Audience?
>
> Joshua Marks
> CTO
> Curriki: The Global Education and Learning Community

> The alignmentObject property athttp://wiki.creativecommons.org/LRMI/Properties/1.0/AlignmentObject

> https://docs.google.com/spreadsheet/ccc?key=0AsqEpfKQI9UsdFpxSlJfQ0k2...

[Zoe]

Hi Brandt,

Excellent summary - thank you.

First things first: I see a clarity error on my part. We don't have
plans to use authoritative URIs for our four national curricula
because they don't exist. There are no centrally (i.e. government)
maintained URIs for the outcomes specified in the four national
curricula.

We can - and are currently - making out own, and yes, we're going to
use those, extensively.

We can do that, but as a solution, that isn't much use to (for
example) a small/niche educational publisher, a publisher who hasn't
got the expertise or resources to create/maintain good URLs, or a
teacher-creator. And the teacher-creators especially make some really,
really great stuff - not to mention, their stuff is more likely to be
both free and remixable!

If we restrict the list of potential LRMI adopters to those who both
understand URIs and can afford to maintain their own URIs, we lose
many of the world's best content creators.

I'm very wary about any specification that could - even inadvertently
- exclude smaller players. I might be missing a point here though -
can anyone assuage this worry for me?

Zoe

On Jul 30, 9:58 am, Brandt Redd <bra...@redd.org> wrote:
> EducationalAlignment was set up to manage alignment to "a node in a
> learning framework." Frameworks include learning standards like the Common
> Core, curricula like Scotland's Curriculum for Excellence, text complexity
> frameworks like MetaMetrics Lexiles, etc.
>
> A critical design decision which distinguishes LRMI from predecessors like

> LOM is that LRMI does not attempt to *define* or *represent* the framework

> only to align to such a framework. The moment you start to define the
> framework or taxonomy it introduces a tremendous amount of complexity that
> violates the simplicity of the Schema.org context in which we operate.
>
> Education Levels are no different. Zoe has started to show how complex it
> is to represent Education Levels. Without the suggested "Target Origin" the
> same term (e.g. "Key Stage 2") has ambiguous meaning. But this is only the
> tip of the iceberg. To properly represent a taxonomy of education levels
> you would also need the expected age range, the day of the year on which
> the age is measured, rules about stage mobility, and ultimately what is
> expected to be taught in that stage.
>

> To manage complexity, we chose to *reference *taxonomies using the
> TargetURL property rather than attempt to *represent* taxonomies in some

> universal way. Given that Zoe and the BBC "have no plans to use targetURL"
> it's not surprising that the schema fails to meet their needs. They've
> chosen not to use the one property that's intended to solve this kind of
> problem.
>
> Attempting to use Audience to solve the problem will ultimately end up with
> the same question -- do you define an audience schema complex enough to
> describe Education Level taxonomies or do you reference external taxonomies?
>
> In my opinion, the BBC should simply assign URLs to the taxonomies they
> find relevant. For example:
>

> http://bbc.co.uk/uri/England/keystage2http://bbc.co.uk/uri/NorthernIreland/keystage2

>
> If, on some future date, the relevant educational authorities assign their
> own URLs for the same levels then they could substitute the new ones. Or,
> the BBC may have enough clout to convince authorities to create URLs like
> these immediately. Keep in mind that the URLs don't have to point at any
> real content. They can even return 404 errors. They just have to be unique
> and someone has to make a definition of their meaning. For an example of
> how this is done in a different domain you can look at how the W3C uses
> URLs to identify encryption algorithms. (http://www.w3.org/TR/xmlenc-core/#sec-AlgID). In their case, the URLs
> redirect to the relevant standards document.
>
> Yes, this is a messy process. In the US, each state that adopts the Common
> Core State Standards can make up to 15% additions to the standards. And
> some states (notably California) are changing the standards (which isn't
> supposed to be done) rather than just adding to them. In addition to the
> core standards (which are only just barely getting identifiers) all of the
> state additions and changes must be assigned identifiers. And then there
> are the states that haven't adopted the common core. Our plan is to use the
> Learning Registry as a clearinghouse and cross reference for all of these
> identifiers. (http://learningregistry.org, also see the common core

> identifier project athttp://corestandards.org/developments-on-common-core-state-standards-...
> ).

> >https://docs.google.com/spreadsheet/ccc?key=0AsqEpfKQI9UsdFpxSlJfQ0k2...

> ...
>
> read more »

[Dan Brickley]

On 30 July 2012 13:00, Zoe <zoe.f...@gmail.com> wrote:
> Hi Josh -
>
> You write:
>
> "Hmmm, so it sounds like the solution to this challenge is to either
> extend EducaitonalALignment, add EducaitonalLevel as a different type
> of alignment, or define this in terms of the Audience?"
>
> Three options worth discussing, but I'm not sure they quite do the
> job.
>
> Just to be *really* explicit about use cases - I'm working on an
> assumption that 'if the (person making markup) has to look up the
> documentation to know what goes in the field, the field isn't
> usable'.

Do you think autocompletion facilities offer scope for improving the
situation here? i.e. something that looks like a text field box, but
in fact when they start typing, it's matching against remote database
entries and recording a canonical URL ID "behind the scenes"?

Dan

[Joshua Marks]

Zoe and Brandt et al,

The "Extension" of AlignmentObject (An admittedly abstracted construction) I
was thinking about is to add another Property to specifically qualify the
authority of the term(s) used in the alignment, even if there is no URL to
point to. The targetSource/targetOrigin/targetAuthority idea. This would
allow Webmasters and educators to tag and classify (disambiguate) alignments
that might use the same term differently from one region, authority or
system to another without the benefit of a standardized URL reference. It
might also be a way to deal with the thorny issue of multiple promulgators
of a standard by a specific authority (e.g. two different URLs that point to
the same node of the same framework in a different database). So the idea of
explicitly appending the "In the UK" or "In Ireland" to the alignment might
solve the issue, and bring other benefits without the burden of requiring
formal and managed URLs.

Zoe, I am glad you are about to solve the issue for the UK systems and their
leveling frameworks by providing a reference lookup. Perhaps you can share
that with this group as a reference for others to use? It seems with this
you can use AlignmentObject for and alignementType of educationalLevel
without extension. However, you are entirely correct that this is not going
to be the case for a lot of the best content producers, and particularly
those grass roots producers we cater to. We can and do provide these
frameworks in a URL look up (The Common Core, and all core standards in all
50 states of the US at least) and a way for users to create an mange their
own reference leveling, sequencing, and other types of frameworks. So you
can rest assured that solution providers will enable more automated tagging
and classification solutions and can and will leverage the abstracted
alignmentObject in LRMI...

All, I am still waiting on a discussion regard the Audience proposal. I
agree that level describes the resource and is part of Alignment as
designed, but there is a corollary in the audience as defined in the
proposal, and some other issue in the present proposal to consider. (Parent
of child is just one.)

[Brandt Redd]

Thanks, Zoe, for two very valuable clarifications -- use of URIs by the BBC and your use case.

I think it's Zoe's use case that we need to consider directly -- her assertion that, "if the (person making markup) has to look up the
documentation to know what goes in the field, the field isn't usable." If we adopt that standard it certainly would send us in a different direction from the multiple uses of EducationalAlignment that we've been favoring.

So far, we've been adopting a strategy more like the one Joshua talks about -- developing tools that support those that are tagging content thereby making it clear what they're asserting and ensuring that consistent values and URIs are used.

Thanks,

Brandt

[Zoe]

Hi Josh,

"We can and do provide these frameworks in a URL look up (The Common

Core, and all core standards in all 50 states of the US at least)...

So you can rest assured that solution providers will enable more
automated tagging and classification solutions and can and will
leverage the abstracted alignmentObject in LRMI... "

I just need to check, Josh - are you suggesting that the Common Core
standards are usable for alignment creation by other English speaking
nations (e.g. Australia, New Zealand, UK, Canada)?

If yes, I'd suggest that might be a little outside the scope that the
education authorities in those nations might be comfortable with -
there's no such thing as a universal curriculum, after all...

If you *are* suggesting that all English speakers will be able to use
the Common Core URLs, I'm not sure that's going to meet all the use
cases - if a teacher in Queensland made a great history resource about
the past and future of Palm Island (former penal colony for indigenous
Australians), what in the Common Core could she ever align that to?

I'd also note that meets the requirements for very common subject
areas, but not for the breadth of educational resources in the world -
there's a lot of not-school formal learning out there, PADI diving
certificates, Diplomas, Certificates... one of my friends has a
qualification in Pet Store management (she fixed the nitrate levels in
my fishtank recently :) ) but I can't imagine that will ever have a
canonical URL in the Common Core!

Zoe

> jma...@curriki.orgwww.curriki.org
>
> I welcome you to become a member of the Curriki community, to follow us
> on Twitter and to say hello on our blog, Facebook and LinkedIn communities.
>
>
>
>
>
>
>

> > >jma...@curriki.orgwww.curriki.org
>
> > > I welcome you to become a member of the Curriki community, to follow
> > > us on Twitter and to say hello on our blog, Facebook and LinkedIn
> communities.
>

> ...
>
> read more »

[Zoe]

(thank you for bringing 'Audience' back up, I have a response for
tomorrow.)

On Jul 30, 7:22 pm, "Joshua Marks" <jma...@curriki.org> wrote:

> jma...@curriki.orgwww.curriki.org
>
> I welcome you to become a member of the Curriki community, to follow us
> on Twitter and to say hello on our blog, Facebook and LinkedIn communities.
>
>
>
>
>
>
>

> > >jma...@curriki.orgwww.curriki.org
>
> > > I welcome you to become a member of the Curriki community, to follow
> > > us on Twitter and to say hello on our blog, Facebook and LinkedIn
> communities.
>

> ...
>
> read more »

[Joshua Marks]

Zoe,

Yes anyone can align to and find materials based on the US Common Core in
Curriki (Or any official US state standard current and past). However, you
can also construct your own alignment framework and align content to it in
the Curriki collection editing tools. Each node of a hierarchical collection
in Curriki has a title, description and a URL/URI (And a complete set of
LRMI compatible metadata). Build a collection tree in Curriki, and you have
a reference framework with URL/URL look-up. If you have a XML export of a
framework (Basically any, but particularly in the SIF
LearnignStandardsDocument and LearningStandardsItem format), we can load it
into our framework alignment and browsing tool. There are lots of others who
build repositories, referitories and curriculum mapping solutions that
provide similar capabilities.

I am not, and would never suggest we have a single curriculum or even a
single way to define and measure outcomes and levels of achievement (And
according to Brant, perhaps the unification in the US K-12 around the common
core is a bit shaker then we would like). However, common structures, common
terms, and forms of interoperability enabled by standards like LRMI can
support a robust cross-flow of content, ideas and processes from one
locality and pedagogy to another. This is why I am involved in the TWG.

[John Fontaine]

Instead of teacher, student and parent how about making a less k-12 focused terminology of "instructor", "learner", "observer". The person observing may be a parent, mentor, manager, or other person beyond just a parent.

Further I propose these other items as part of a regular data profile:
"content evaluator" -- this tags content focused on supporting those who are evaluating the materials and might include reviews and other non-core supporting materials related to the content.
"grader" -- this is material for those who are grading a specific exercise. It might include rubrics, answer sheets and other data
"leader" -- for materials such as a group exercise where there is no teacher but someone is leading the session
"participant"-- for other participants where a leader is described.

John Fontaine

[Eric Weiss]

I am the project manger for the LRMI Tagger and Search tools. I've been monitoring (lurking) this group to see how progress is made on this topic as it affects our design and development of future versions of the LRMI Tagging Tool. After speaking today with Michael Jay of Educational Systemics about how to modify Tagger to support both age and grade level, I thought it may be beneficial to circulate the approach here for your feedback. 

What we are contemplating is:

1. User can select to tag for Age Level or Grade Level.
2. If Age, than use simple entry process of selecting an age range.
3. If Grade Level, user would need to select their geographic basis for grade level (could be set as a default value or changed on the fly) and then select the grade level(s) that the educational resource they are tagging is aligned to
4. If they tag to Grade Level, Tagger would automatically map to age, based on the selected geography. If they tag to Age, we would map to the default grade, also based on geography. Tagger would need to maintain a matching table to align age with grade level, similar to the following (I know this is non-trival):
1. https://en.wikipedia.org/wiki/Educational_stage
5. The output from Tagger would align to the existing LRMI 1.0 spec as follows:
1. The metadata for age would sit under the TypicalAgeRange property. 
2. The metadata for grade would site under the alignmentType property as suggested by Greg G. 

For publishers who may typically align their content to grade level, especially in the US, they would be able to leverage their existing metadata. For smaller publishers or those creating OER, they would have a choice to make of specifying age or grade, at their discretion. If they are using Tagger, than they would be sure to have their tagged content properly aligned to LRMI such that their selection would be immaterial as far as Tagger is concerned. If not using Tagger, we would still be able to publish our best practices for tagging. It also retains a significant level of granularity which I believe is desired by publishers and content creators to make their content discoverable for very specific searches.

Looking forward to your feedback.

Eric Weiss

LRMI Search and Tagger Project Manger

Agilix Labs Inc.

[Greg Grossmeier]

Sorry all for my apparent relative silence on this thread. I've been
talking with various individuals privately. My apologies.

<quote name="Zoe" date="2012-07-30" time="05:00:53 -0700">

> So I do think that, same as my earlier post, we still really do need a
> dedicated field to specify the educational authority that defines the
> level.

(I'm stealing the idea from others on this list, of course, Joshua in
particular).

Would something called "targetAuthority" suffice? That could be a text
value (eg: "Council of Chief State School Officers", or "CCSSO"? that's
where a text field gets tricky, especially with organizational names
that are commonly abbreviated, but abbreviations are not canonical).


That would extend the current list of properties of AlignmentObject to
be (new one highlighted):

alignmentType
Text

A category of alignment between the learning resource and the framework node. Recommended values include: 'assessed', 'teaches', 'requires', 'textComplexity', 'readingLevel', 'educationalSubject', and 'educationLevel'.

targetDescription
Text

The description of a node in an established educational framework.

*******************
targetAuthority
Text
The educational authority responsible for the definition of the
alignment target (eg: educational standard, grade level, text
complexity, etc).
*******************

targetName
Text

The name of a node in an established educational framework.

targetUrl
URL

The URL of a node in an established educational framework.

Word-smithing of that definition is appreciated.

Best,

[Paul Libbrecht]

Zoe,

this is soooo funny, to hear "universal confusions" among, basically, a group of worldwide experts into metadata modeling (I believe the ASPECT project, the EU project right?, was indeed made of such).

Maybe this is a good reflection that, simply said, educational systems are just not fully comparable!

In the intergeo project, we admitted full incomparability, including differentiating the "siebte Klasse" of two neighbouring states in Germany or other countries. That means that an assessment of educational level is great if it matches, but it rarely does. Some kind of generalization procedure would be desirable...

This is all living in an ontology with educational regions, pathways, and levels; all best practice in principle but we had a hard time completing the set of educational levels in Europe (and sadly enough the Eurydice organization has not taken that role).

Le 30 juil. 2012 à 17:20, Dan Brickley a écrit :

Do you think autocompletion facilities offer scope for improving the
situation here? i.e. something that looks like a text field box, but
in fact when they start typing, it's matching against remote database
entries and recording a canonical URL ID "behind the scenes"?

Dan, we tried that for http://i2geo.net/, just try the search box there to search resources for that educational level, and it  works for some educational regions where there are really few but for others it is harder. We have not yet been able to fully polish things.

For Germany, for example, it is almost impossible to know which state you are talking about when using a word such as "siebte Gymnasium": it can mean any the seventh grade of any the states and for each there's several pathways for each, "Gesamtschule", "Realschule", "Gymnasium". For some educational regions, where there's not too many, auto-completion helps to disambiguate (that works for "troisième") but for some it has failed thus far. We have an unachieved vision for enhancement... having a location at hand might considerably help.

paul

[Joshua Marks]

+1 for me (I often agree with myself ;-)

Joshua Marks
CTO
Curriki: The Global Education and Learning Community
jma...@curriki.org
www.curriki.org

I welcome you to become a member of the Curriki community, to follow us
on Twitter and to say hello on our blog, Facebook and LinkedIn communities.


-----Original Message-----