Explaining the LRMI Alignment Object
Phil Barker
I've written a post trying to explain the technical aspects of the LRMI
Alignment object in some depth:
http://blogs.cetis.ac.uk/philb/2014/03/06/explaining-the-lrmi-alignment-object/
From implementations of LRMI, I know some people have struggled with
this, I'm interested in feedback on whether this level of technical
detail is helpful. It's aimed at people with a hand in developing
services that expose/embed LRMI in their web pages, so it assumes some
familiarity with schema.org, and definitely isn't meant for
non-technical people trying to make a policy decision on whether LRMI
would be useful.
Let me know what you think.
Phil.
--
work: http://people.pjjk.net/phil
twitter: https://twitter.com/#!/philbarker
facebook: https://www.facebook.com/philbarker66
Ubuntu: not so much an operating system as a learning opportunity.
http://xkcd.com/456/
Steve Midgley
--
You received this message because you are subscribed to the Google Groups "Learning Resource Metadata Initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lrmi+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
Steve Midgley
Stuart Sutton
This is really useful Phil. I've already shared with some groups who are thinking about how to make digital representations of standards and content.The issue that I haven't found an answer to yet is if there is a standard way to represent an curricular framework (in schema.org ideally)? I don't think we have the right keys/vocab yet, let alone examples. For example, Common Core doesn't describe their standards in the same way as Next Generation Science Standards.
That said, we have made progress in creating permalink URL IDs for both, so the schema.org alignment object "works" with both.How important do you think it is to have a way of marking up the curricular frameworks in LRMI or some proposed extensions? I'm talking with a group about working on this, but would love to get people's opinion from this group first. I think ASN had an RDF way of doing it, and GIM (Granular Identifiers Metadata) project built something but never released it..
All input/thoughts welcome,SteveOn Thu, Mar 6, 2014 at 7:14 AM, Phil Barker <ph...@pjjk.net> wrote:
Hello all.
I've written a post trying to explain the technical aspects of the LRMI Alignment object in some depth:
http://blogs.cetis.ac.uk/philb/2014/03/06/explaining-the-lrmi-alignment-object/
From implementations of LRMI, I know some people have struggled with this, I'm interested in feedback on whether this level of technical detail is helpful. It's aimed at people with a hand in developing services that expose/embed LRMI in their web pages, so it assumes some familiarity with schema.org, and definitely isn't meant for non-technical people trying to make a policy decision on whether LRMI would be useful.
Let me know what you think.
Phil.
--
work: http://people.pjjk.net/phil
twitter: https://twitter.com/#!/philbarker
facebook: https://www.facebook.com/philbarker66
Ubuntu: not so much an operating system as a learning opportunity.
http://xkcd.com/456/
--
You received this message because you are subscribed to the Google Groups "Learning Resource Metadata Initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lrmi+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
--
You received this message because you are subscribed to the Google Groups "Learning Resource Metadata Initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lrmi+uns...@googlegroups.com.
Steve Midgley
Stuart Sutton
Oh my bad Stuart on the verb tense. I did not mean to imply that ASN isn't with us - just that the work was completed in the past, but I see my language was definitely wrong.. I'm very excited that D2L has picked up the project and looking forward to working with you and them.Thanks for the summary - it sounds like there are at least 3 models for describing or depicting curricular frameworks:ASN, InLoc, GIM
Is that right?
Yes, and my take is that the metadata wars are basically long over and the lesson learned is that no schema won (or will ever win). So, the most productive conversation would move us from "which do we choose" to what models will take us beyond even metadata interoperability to metadata harmonization (http://www.diva-portal.org/smash/get/diva2:369527/FULLTEXT02).
Should/could any of these be mapped over into a schema.org/lrmi orientation so that both sides of the relationship (per Phil's article) are described with the same metadata structures?
Is your question targeted toward everyone being able to resolve the reference to a framework node and get the exact same serialization of something back? Thus my confusion...
Phil Barker
Hello Steve
On 06/03/14 18:15, Steve Midgley wrote:
This is really useful Phil. I've already shared with some groups who are thinking about how to make digital representations of standards and content.
The issue that I haven't found an answer to yet is if there is a standard way to represent an curricular framework (in schema.org ideally)?
1, a post from a colleague Simon Grant, which addresses how one standard, InLoc, maps to schema alignments (others exist, as I'm sure Stuart will have mentioned)
2, there was a proposal from Dan Brickley a few months ago for a way of encoding vocabulary schemes in schema.org, see https://groups.google.com/forum/#!topic/lrmi/jwgo02DpQzY (though the link to the proposal now doesn't work). I had a look at how this would work for Educational Frameworks and it seemed like a good match, the big question I had in my mind was how to reference it.
I think 2 has possibilities for LRMI, it clarifies how to refer to an educational framework; it wasn't proposed that it would do much by way of expressing relationships in such a framework (though it was a step in that direction), though I am not sure that is a big priority (I think Stuart touched on this as well).
Phil
I don't think we have the right keys/vocab yet, let alone examples. For example, Common Core doesn't describe their standards in the same way as Next Generation Science Standards. That said, we have made progress in creating permalink URL IDs for both, so the schema.org alignment object "works" with both.
How important do you think it is to have a way of marking up the curricular frameworks in LRMI or some proposed extensions? I'm talking with a group about working on this, but would love to get people's opinion from this group first. I think ASN had an RDF way of doing it, and GIM (Granular Identifiers Metadata) project built something but never released it..
All input/thoughts welcome,Steve
On Thu, Mar 6, 2014 at 7:14 AM, Phil Barker <ph...@pjjk.net> wrote:
Hello all.
I've written a post trying to explain the technical aspects of the LRMI Alignment object in some depth:
http://blogs.cetis.ac.uk/philb/2014/03/06/explaining-the-lrmi-alignment-object/
From implementations of LRMI, I know some people have struggled with this, I'm interested in feedback on whether this level of technical detail is helpful. It's aimed at people with a hand in developing services that expose/embed LRMI in their web pages, so it assumes some familiarity with schema.org, and definitely isn't meant for non-technical people trying to make a policy decision on whether LRMI would be useful.
Let me know what you think.
Phil.
--
work: http://people.pjjk.net/phil
twitter: https://twitter.com/#!/philbarker
facebook: https://www.facebook.com/philbarker66
Ubuntu: not so much an operating system as a learning opportunity.
http://xkcd.com/456/
--
You received this message because you are subscribed to the Google Groups "Learning Resource Metadata Initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lrmi+uns...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
--
You received this message because you are subscribed to the Google Groups "Learning Resource Metadata Initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lrmi+uns...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
-- work: http://people.pjjk.net/phil twitter: https://twitter.com/#!/philbarker
Ubuntu: not so much an operating system as a learning opportunity. http://xkcd.com/456/
Steve Midgley
Joshua Marks
All.
Catching up here, but a very interesting conversation. First off, nice job Phil on the blog! We had a very long dialog about the Alignment Object and the type of alignment. Where we ended up was elegant and it seems useful. But, as you and Steve point out, falls a bit short as there is little, er, really no guidance on how to express a framework or taxonomy even though we can effectively point to most any. Even the CCSSO could not get the SIF (LSI/LSD) structure right for corestandards.org.
Stuart, as always you have put some provocative ideas out there. Metadata harmonization and the assertion that framework promulgators do not care about standard representations. History certainly supports you, but I think ‘the times they are a changing.’ Historically frameworks were created as a way to assure that curriculum and instruction achieved some common benchmarks. In the process it often implies or influences pedagogy and methodology. Regardless, there is rarely any intention that the frameworks be used for classification and discovery of materials. But yet, that is the objective we seek to realize. In some sense we are seeking a different kind a framework for a different purpose. Perhaps harmonizing these frameworks might get close, but I am not sure how we do that (I need to read that paper Stuart shared).
Regardless of the structure or intent of a framework, I think we might help folks by providing guidance and a schema.org based approach to expressing any framework. Other then the proposal mentioned, I have not seen a way to do this in Schema, and perhaps this is a contribution we could make. Another approach, taking a clue for Learning Registry, is to provide a free and open source taxonomy creation tool that that automatically publish a well formed and referenceable framework. I have noted that some organizations have created amazingly complex content framework using the Curriki collection tools, such as this one created by the Santa Clara County Office of Education => http://www.curriki.org/xwiki/bin/view/Coll_Group_SantaClaraCountyCCSSCurriculumMapping/Math (Expand the unit maps in each grade to see their structure.) What is cool, is this collection is itself a framework that can be referenced by URL and as XML (e.g. http://www.curriki.org/xwiki/bin/view/Coll_Group_SantaClaraCountyCCSSCurriculumMapping/Math?xpage=xml). This unintended use is quite interesting as the users themselves create and manage the framework and use the framework (TOC) itself to align content. Perhaps this is a more useful form of harmonization; let the community of use create and manage the shared framework.
I think this conversation has some legs. More to come.
-Joshua
All input/thoughts welcome,
Steve
To unsubscribe from this group and stop receiving emails from it, send an email to lrmi+uns...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
--
You received this message because you are subscribed to the Google Groups "Learning Resource Metadata Initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lrmi+uns...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
--
You received this message because you are subscribed to the Google Groups "Learning Resource Metadata Initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lrmi+uns...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
--
You received this message because you are subscribed to the Google Groups "Learning Resource Metadata Initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lrmi+uns...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
--
You received this message because you are subscribed to the Google Groups "Learning Resource Metadata Initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lrmi+uns...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
Stuart Sutton
All.
Catching up here, but a very interesting conversation. First off, nice job Phil on the blog! We had a very long dialog about the Alignment Object and the type of alignment. Where we ended up was elegant and it seems useful. But, as you and Steve point out, falls a bit short as there is little, er, really no guidance on how to express a framework or taxonomy even though we can effectively point to most any. Even the CCSSO could not get the SIF (LSI/LSD) structure right for corestandards.org.
Stuart, as always you have put some provocative ideas out there. Metadata harmonization and the assertion that framework promulgators do not care about standard representations.
I guess I'd go further and say that it isn't LRMI's task to solve the "multiple information models matter" in terms of the frameworks to be referenced. I personally think that getting it clear that the desire for some homogeneous information model on the framework end of the equation isn't an LRMI issue. Then we can stop talking about it "falling short"--LRMI being the "it".
History certainly supports you, but I think ‘the times they are a changing.’ Historically frameworks were created as a way to assure that curriculum and instruction achieved some common benchmarks. In the process it often implies or influences pedagogy and methodology. Regardless, there is rarely any intention that the frameworks be used for classification and discovery of materials. But yet, that is the objective we seek to realize. In some sense we are seeking a different kind a framework for a different purpose. Perhaps harmonizing these frameworks might get close, but I am not sure how we do that (I need to read that paper Stuart shared).
Regardless of the structure or intent of a framework, I think we might help folks by providing guidance and a schema.org based approach to expressing any framework.
But everything doesn't model taxonomically . Take, the U.S. Next Generation Science Standards (NGSS). You can force it into a taxonomic structure if you are unwilling to recognize that the canonical framework is graphing between two distinct frameworks--the NGSS itself and the Framework for K-12 Science Education [3] and are willing to accept the information losses and noise by shoehorning it into a taxonomic structure.
For more options, visit https://groups.google.com/d/optout.
Jade Forester
My name is Jade, I'm the Global Coordinator for Mozilla Open Badges.
My team and I would love to hear more about your work on this alignment object on LRMI in an upcoming Community Call, held every Wednesday at 12pm ET.
These calls are a place for our team and community to share work and gather feedback on badges and related projects, and if you're able and willing, we'd like you to present on a call.
If you are interested, you can email bad...@mozillafoundation.org and we'll find a date that works for you.
Looking forward to hearing from you,
Jade
Global Marketing Coordinator + Liaison
Mozilla Open Badges
@jade_forester
@OpenBadges
Phil Barker
Hello Jade, everyone.
Thank you for you invite, it is very timely, and welcome. I am interested in investigating the overlap between LRMI expressions of educational alignments and Open Badges. Well, I guess it is no accident that the BadgeClassin the open badge spec[1] has an alignment property that can be an array of AlignmentObjects, and there is [possibly] a simple mapping between the name, url and description properties of the targetUrl, targetName and targetDescription of the LRMI alignment object [2]. It's worth double checking on any differences, accidental or deliberate
Phil
1. https://github.com/mozilla/openbadges-specification/blob/master/Assertion/latest.md
2. http://schema.org/AlignmentObject
--
You received this message because you are subscribed to the Google Groups "Learning Resource Metadata Initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lrmi+uns...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
-- <http://www.icbl.hw.ac.uk/~philb/>
Sunday Times Scottish University of the Year 2011-2013
Top in the UK for student experience
Fourth university in the UK and top in Scotland (National Student Survey 2012)
We invite research leaders and ambitious early career researchers to join us in leading and driving research in key inter-disciplinary themes. Please see www.hw.ac.uk/researchleaders for further information and how to apply.
Heriot-Watt University is a Scottish charity registered under charity number SC000278.
Simon Grant
Renato
Design for diversity – multiple alignments (same learning resource, multiple frameworks)
Hello everybody.
I am late here but the issues are still open...
Reaching consensus on a single universal framework to align to would no doubt simplify many problems. Yet the different existing frameworks are an expression of different viewpoints: it might be beneficial aiming intentionally at supporting this diversity – an invaluable asset of the open Internet...
This rises (also) the issue of multiple alignments. One question here.
If we need to align a learning resource to (equivalent) competencies in two different (educational) frameworks, we can either:
(a) embed this information in the learning resource metadata, using two alignmentObjects, or
(b) embed the equivalency information in the frameworks themselves.
ASN is supporting the second approach – which is more efficient if you have many learning resources to (double) align (besides, this solution has other advantages). In order to reduce complexity, ASN makes use of “intermediaries”, that “are services that handle the mapping of an ASN state standards statement URI to a statement URI from another state”. The mapping makes use of rich properties such as broader, narrowMatch, isPartOf... in addition to exactMatch.
Does somebody in the group have experience with or comments on this approach?
Thank you,
Renato
Barker, Phil
Sent: 03 September 2014 13:16
To: lr...@googlegroups.com
Cc: lrmi...@googlegroups.com
Subject: Re: Explaining the LRMI Alignment Object
You received this message because you are subscribed to the Google Groups "Learning Resource Metadata Initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lrmi+uns...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Stuart Sutton
Hello Renato, hello everyone
Not much direct experience, I think we're all exploring possibilities here, but a comment.
Yes, both options are valid.And mapping between compatible frameworks may be more efficient.
But, sometimes educational frameworks are based on such different assumptions about education that a mapping is not viable because some frameworks have such different starting points that they are talking about different classes of things.
But, I'd posit that a mapping assertion implying exact identity is not the only reason one might want to map these two nodes--that's why ASN provides an array of mapping predicates so you can go along a continuum from exact match or more-or-less match. Highly legitimate uses at the more-or-less end of this mapping continuum include resource discovery where a higher degree of noise tolerance is bearable than in what I would call conformance mapping where exact identity between nodes is the goal.
This is the same "more-or-less" tradeoff we have always tolerated in subject indexing where resources tagged with "Cybernetics" from Library of Congress Subject Headings and "Cybernetics" from Dewey are deemed more-or-less likely to be useful for the same purpose when tagging a book on the topic of cybernetics even though the topics likely have different extensions in the two vocabularies. This, by the way, is why standards such as W3C's Simple Knowledge Organization System (SKOS) standard provide mapping predicates running from exact match to more-or-less.
The mappings you finds in the ASN-US data, Renato, are of the exact kind...e.g., mapping an ASN data representation of a node in the U.S. Common Core to that same node in the canonical version. There are no "more-or-less" mappings in framework-to-framework mapping assertions in the ASN-US data. But, the ASN Description Framework supports less than exact mappings.
For example in Scotland the the Curriculum for Excellence[1] focuses not only on competencies and knowledge but on "experiences and outcomes" and has statements such as "I have investigated the everyday contexts in which simple fractions, percentages or decimal fractions are used and can carry out the necessary calculations to solve related problems." [2] This specifies an *experience* so it would be difficult to map from the common core to it. Hence the need for both approaches, I think. Then there is the problem of who in the world cares enough about both Texas Maths Standards and Scottish Curriculum for Excellence to do the mapping where it is possible, after all it would only be more efficient if there were a large number of resources of interest to teachers and pupils in both systems.
Renato
From aligning resources to aligning frameworks
Hello everybody, thank you Phil and Stuart for the very rich replies.
I think we can agree (not necessarily exactAgree, but broadAgree):
- the cost of linking the “Texas framework” and the “Scotland framework” is high;
- the cost of not doing it (opportunity cost) might be even higher;
If we don't link, we would end-up in partitioning the space of learning resources in disjointed “islands” corresponding to unlinked frameworks: the “killer” alignment feature would be powerless in supporting discoverability at full scale.
- we might not be able to align with exactMatching, yet we could align with more-or-less (broad...) matching;
This is explicitly designed in ASN (Stuart). As a reinforcing example, I mentioned elsewhere that OER Commons make use of the extra attribute “degree of alignment”.
- exploiting aligned frameworks requires linked-data oriented consumer applications;
This added complexity brings other advantages, including the potential “harmonization” with other metadata standards.
- directly aligning learning resources is less expensive at an initial stage, aligning frameworks is more efficient on a larger scale.
Could the following “evolutionary” strategy help to gradually move from directly aligned learning resources to aligned frameworks?
Bootstrapping - People start by double aligning a few initial resources to frameworks directly with multiple AlignmentObjects;
Gaining momentum – when somebody aligns a new resource R1 to Node N1 in Framework FA, an “intelligent tagger” would suggest that another resource R2 which is already aligned to the same Node N1 in Framework FA is also aligned to Node N2 in Framework FB (the familiar “people who bought that one also bought these ones”) [btw – this could also work as a useful search expansion technique].
Climax - once we reach a critical number of resources double aligned, harvest those metadata to support the more efficient direct alignment of the corresponding frameworks. That is: If resource R1 is aligned to node N1 in framework FA and to node N2 in framework FB, then NIFA and N2FB could be directly aligned.
The devil, of course, might be in the “details”: exponential explosion of the number of framework-to-framework mappings, degrees of confidence, alignment to multiple nodes within the same framework, granularity, non transitivity of broadMatch...
Might it be wiser, in order to inter-link the different potential “islands”, to use a very coarse form of alignment, e.g. to general subject headings (even built by the community, as Joshua suggested, possibly even within schema.org itself, as mentioned by Steve) rather than fine grained competence standards?
I hope somebody will share his/her thoughts on this. Thank you,
Renato