Enriching a curriki-page

[Paul Libbrecht]

Hello LRMI group,

following the announce of Greg, we have been investigating annotating an existing Curriki resource information page with LRMI annotations.
This is incomplete still but it would serve a discussion.

My principle in doing this was to make no visible change to the existing Curriki pages, to follow as much as possible the principle of "schema.org property values are displayed text",  and to change the least possible the HTML structure.

The result is taken from one resource page saved a few months ago:
http://www.hoplahup.net/tmp/LRMI/MySpaceinDemocracymine.html
then saved as-web-page-complete (which, thus, misses a few pictures), then rearranged.

Could people here comment on this attempt?
What kind of interoperability could we expect from there?
Can one of you with a repository sketch how it could land into it?

The search engine interoperability is one that many people hope out of schema.org, so let's see what one tool gives us: We've been trying, the Google Structured Data Testing Tool. it's not perfect:
http://www.google.com/webmasters/tools/richsnippets?url=http%3A%2F%2Fwww.hoplahup.net%2Ftmp%2FLRMI%2FMySpaceinDemocracymine.html&html=

Among others, I see the following issues:

- almost everything about author is wrong there, the HTML5 tags used to make parts of the page be "tangential" are ignored by this tool... I wonder what to do there.

- the "publisher" property does not seem to be properly exploited. Am I doing something wrong here? 
(asking the Google tool above suggests me Google+... not really what I have!)

- how is typical age range going to be exploited? 
(the fact that the two lines become a single sounds worrying)

- the "created" date sounds to display the text and to ignore the value of the content attribute, this sounds wrong. Any clue?

Moreover, I intend to enrich the output with standards information (available on the source:http://www.curriki.org/xwiki/bin/view/Coll_sriii2000/MySpaceandDemocracy?bc=&viewer=standards)

thanks in advance for your thoughts and comments.

Paul

[Phil Barker]

Hello Paul, everyone.

Looks like a good start, for example, good to see the URL property being
used to show that this is a description of a resource held elsewhere,
not a description of the page itself.

Comments:

Itemscope itemtype property should be a URL, e.g.

itemscope itemtype="http://schema.org/Collection"

However collection isn't a recognised schema.org itemtype. You would be
better using http://schema.org/CreativeWork


Where does itemprop="topic" come from? Shouldn't it be itemprop="about" ?

Also I would mark up each subject individually, ideally only marking up
the end node of the classification, i.e.

<li>Educational Technology&gt; <span itemprop="about">Integrating Technology into the Classroom</span></li>

but at the very least

<li itemprop="about">Educational Technology&gt; Integrating Technology into the Classroom</li>

That would align with other metadata schema, e.g. LOM and Dublin Core,
where each subject goes in its own tag, and I guess with how the data is
stored in most catalogues.


Likewise, repeat the typicalAgeRange itemprop for each age range

Also there is the whole grade level as framework for educational
alignment Vs typical age range issue. Something like this would use
both to their full potential.

<li>
<span itemprop="educationalAlignment" itemscope itemtype="http://schema.org/AlignmentObject">
<meta name="alignmentType" content="educationLevel">
<meta name="educationalFramework" content="US Grade levels">
<span itemprop="targetName"> Grades 6-8 </span>
</span>
/
<spanitemprop="typicalAgeRange">Ages 11-13</span>
</li>
<li>
<span itemprop="educationalAlignment" itemscope itemtype="http://schema.org/AlignmentObject">
<meta name="alignmentType" content="educationLevel">
<meta name="educationalFramework" content="US Grade levels">
<span itemprop="targetName">Grades 9-10 </span>
</span>
/
<spanitemprop="typicalAgeRange">Ages 14-16</span>
</li>

(hmm, interesting, the Google structured data testing tool doesn't show
the content of the meta tags...still it's there for people who do trust you)

Phil

On 26/09/12 23:10, Paul Libbrecht wrote:
> Hello LRMI group,
>
> following the announce of Greg, we have been investigating annotating
> an existing Curriki resource information page with LRMI annotations.
> This is incomplete still but it would serve a discussion.
>
> My principle in doing this was to make no visible change to the
> existing Curriki pages, to follow as much as possible the principle of

> "schema.org <http://schema.org/> property values are displayed text",

> and to change the least possible the HTML structure.
>
> The result is taken from one resource page saved a few months ago:
> http://www.hoplahup.net/tmp/LRMI/MySpaceinDemocracymine.html
> then saved as-web-page-complete (which, thus, misses a few pictures),
> then rearranged.
>
> Could people here comment on this attempt?
> What kind of interoperability could we expect from there?
> Can one of you with a repository sketch how it could land into it?
>
>
>
> The search engine interoperability is one that many people hope out of

> schema.org <http://schema.org/>, so let's see what one tool gives us:

> We've been trying, the Google Structured Data Testing Tool. it's not
> perfect:
> http://www.google.com/webmasters/tools/richsnippets?url=http%3A%2F%2Fwww.hoplahup.net%2Ftmp%2FLRMI%2FMySpaceinDemocracymine.html&html=

> <http://www.google.com/webmasters/tools/richsnippets?url=http://www.hoplahup.net/tmp/LRMI/MySpaceinDemocracymine.html&html=>

>
> Among others, I see the following issues:
>
> - almost everything about author is wrong there, the HTML5 tags used
> to make parts of the page be "tangential" are ignored by this tool...
> I wonder what to do there.
>
> - the "publisher" property does not seem to be properly exploited. Am
> I doing something wrong here?
> (asking the Google tool above suggests me Google+... not really what I
> have!)
>
> - how is typical age range going to be exploited?
> (the fact that the two lines become a single sounds worrying)
>
> - the "created" date sounds to display the text and to ignore the
> value of the content attribute, this sounds wrong. Any clue?
>
> Moreover, I intend to enrich the output with standards information
> (available on the
> source:http://www.curriki.org/xwiki/bin/view/Coll_sriii2000/MySpaceandDemocracy?bc=&viewer=standards)
>
> thanks in advance for your thoughts and comments.
>
> Paul

--
Ubuntu: not so much an operating system as a learning opportunity.
http://xkcd.com/456/

[Joshua Marks]

Phil (et al),

Thank you for the detailed and helpful input. I wanted to talk over some
points of best practice that go to the issues of the relationship of the
visual presentation and the meta-tagging. What follows is Curriki specific,
but generalizable to many learning resource repositories, and most
particularly to those looking to provide modular reusable content. Please
review the live version of the resource Paul is using here ->
http://www.curriki.org/xwiki/bin/view/Coll_sriii2000/MySpaceandDemocracy

I would like to point out a few things about this resource and its
presentation. First the resource itself has a number of different views (In
a semi-restful way). The default view is the primary content view itself
(The Content Tab). If you add a parameter to this URL (or click on the tabs)
you get the other views:

- Information ->
http://www.curriki.org/xwiki/bin/view/Coll_sriii2000/MySpaceandDemocracy?bc=
&viewer=info (The one Paul enriched)
- Standards ->
http://www.curriki.org/xwiki/bin/view/Coll_sriii2000/MySpaceandDemocracy?bc=
&viewer=standards (More specific information),
- comments ->
http://www.curriki.org/xwiki/bin/view/Coll_sriii2000/MySpaceandDemocracy?bc=
&viewer=comments

There are yet other views and modes under this as well such as the history,
the history compare, embeddable students viewers, and a native XML view as
well.
- History ->
http://www.curriki.org/xwiki/bin/view/Coll_sriii2000/MySpaceandDemocracy?bc=
&viewer=history
- Student/LMS view ->
http://www.curriki.org/xwiki/bin/view/Coll_sriii2000/MySpaceandDemocracy?bc=
&viewer=embed
- Data view ->
http://www.curriki.org/xwiki/bin/view/Coll_sriii2000/MySpaceandDemocracy?xpa
ge=xml

Now all of these views are of the same resource and have the common header
section with attribution, and when you hover a mouse over the title, a
summary of the metadata is displayed (Form the info tab, but not all of it).
If we put all the information and content on one page, well no human would
like to look at it all. This brings some questions related to your statement
about the info tab being a descriptive record but not the content itself.

1) Is the resource the base URL and the info and other tabs just a mode of
the same view, or is each tab a separate page that should express only the
metadata displayed within and its reference to the content tab?
2) If there is additional metadata that is hidden until some user action,
should that be tagged with LRMI tags and will the search engines index or
ignore such information because it is not directly visible?
3) Should each of the tabs and modes have the complete resource record
metadata expressed in the meta header as LRMI/Schema tags or just he inline
elements that are displayed? Should nothing be expressed in the meta-header
and only he inline visible elements should be tagged?

Now the next item to observe is that this resource is really a collection of
other resources (As depicted in the Table of Contents). As you click into
this TOC, you will see that each element of this "Unit of instruction" is
itself an independent and independently discoverable resource. Such as this
one ->
http://www.curriki.org/xwiki/bin/view/Coll_sriii2000/Mini-LessonsandActiviti
es?bc=;Coll_sriii2000.MySpaceandDemocracy;Coll_sriii2000.MySpaceinDemocracy%
2DPart1 . There is no way in LRMI to indicate such a relationship (e.g.
"This resources belongs to or is a part of this other resource") This leads
to a more problematic issue of common practice: what is the appropriate
granularity to express things as independent resources? This however is a
question that perhaps is out of scope. Many, if not most LMSs or
repositories only express the container (Course) as the resource, not the
elements within. As a practice, the molecules (chapters and units) and atoms
(assets) might be made independently discoverable as they are in Curriki.

LRMI and Schema were devised to classify any form and structure of web
content. The more consistently learning resource sites can both tag,
structure and organize their resources, the more useful the discovery will
be at the item level in constructing personalized learning sequenced or
assembling topical materials collections.

More to come...

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.

[Phil Barker]

On 27/09/12 23:41, Joshua Marks wrote:
> Phil (et al),
>
> Thank you for the detailed and helpful input.

You're welcome. Sorry, I only have time for sketchy input today, on a
couple of your points.

>
>
> 1) Is the resource the base URL and the info and other tabs just a mode of
> the same view, or is each tab a separate page that should express only the
> metadata displayed within and its reference to the content tab?
> 2) If there is additional metadata that is hidden until some user action,
> should that be tagged with LRMI tags and will the search engines index or
> ignore such information because it is not directly visible?
> 3) Should each of the tabs and modes have the complete resource record
> metadata expressed in the meta header as LRMI/Schema tags or just he inline
> elements that are displayed?

That's quite a deep series of questions. I know some people smarter than
me who take a web architecture sort of stance that different views like
this are just different representations of a resource, and others who
say some of these views are of a metadata record which is a distinct
resource. Me, I don't know. But from a schema.org/LRMI point of view I
think that if you give the URL of the item as
http://www.curriki.org/xwiki/bin/view/Coll_sriii2000/MySpaceandDemocracy
then the right thing would be found. I would hope that application
consuming metadata found on different pages for items with the same URL
would aggregate that metadata, but I don't know of any existing
practice.[--I would love to hear of some.] Site maps should help avoid
the problem described in point 2 in making sure that an indexer can
reach every URL that users could see through some action.

> Should nothing be expressed in the meta-header
> and only he inline visible elements should be tagged?
>

See http://support.google.com/webmasters/bin/answer.py?hl=en&answer=1093493
Is your marked-up content hidden from users? for the answer from
Google. Of course other consumers of schema markup may have different
criteria for what metadata they trust.

> Now the next item to observe is that this resource is really a collection of

> other resources (As depicted in the Table of Contents). [snip] There is no way in LRMI to indicate such a relationship (e.g. "This resources belongs to or is a part of this other resource")
Yes, I had a look at this yesterday. The only concept of collections I
could find in schema.org was for http://schema.org/WebPage which can
have a property isPartOf a http://schema.org/CollectionPage
Of course a lot of the components of many learning resource aren't
webpages, they are powerpoints etc., and it would be a shame to describe
them as webpages. Perhaps it is worth suggesting that schema.org move
the isPartOf itemprop up to creative work?

Phil

[Greg Grossmeier]

<quote name="Phil Barker" date="2012-09-28" time="10:48:48 +0100">

> Perhaps it is worth suggesting that
> schema.org move the isPartOf itemprop up to creative work?

I didn't realize isPartOf was on WebPage not CreativeWork. Good
suggestion.

Greg

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

[Peter Pinch]

In responding to Paul Libbrecht's test markup for curriki, Phil Barker
suggested using an itemtype of CreativeWork.

In my earlier tests for MIT OpenCourseWare, I used WebPage because that
seemed closest to what I have.
(http://www.hawklake.com/courses/electrical-engineering-and-computer-scienc
e/6-003-signals-and-systems-spring-2010/

Looking at WebPage, it seems unlikely I'm going to use any of its
properties, although it seems accurate enough. Is there any best practice
regarding specificity of itemtype?


And then, at some point, I think Aaron Bradley suggested a new type of
"course" but I don't think that went anywhere.

- Peter

-----Original Message-----
From: Phil Barker <phil....@hw.ac.uk>
Reply-To: "lr...@googlegroups.com" <lr...@googlegroups.com>
Date: Thursday, September 27, 2012 5:38 AM
To: "lr...@googlegroups.com" <lr...@googlegroups.com>
Subject: Re: Enriching a curriki-page

[Phil Barker]

Side-stepping any ontological arguments about what is a course, it has
been suggested that LRMI properties are added to the schema.org Event
type (as well as CreativeWork). Both Greg and Dan Brickley seemed to
agree that was feasible. Is it going ahead?

Phil

[Greg Grossmeier]

<quote name="Phil Barker" date="2012-10-08" time="18:40:16 +0100">

>
> Side-stepping any ontological arguments about what is a course, it
> has been suggested that LRMI properties are added to the schema.org
> Event type (as well as CreativeWork). Both Greg and Dan Brickley
> seemed to agree that was feasible. Is it going ahead?

I believe so.

I've updated the phrasing on the proposal to be explicit:
"The schema mainly adds new properties (typically but not necc. on
CreativeWork and Event):"

I'll let you all discuss if you think all of LRMI applies to Event.