Re: [pelagios] Digest for pelagios-project@googlegroups.com - 4 Messages in 1 Topic

[Robert Sanderson]

Dear all,

Just to weigh in on the annotation topic ...

Simon Rainer <Rainer...@ait.ac.at> Oct 27 05:41PM   
Anyways: What we do in Pelagios is really best thought of as 'tagging'. We normally have documents (or items), and we're tagging them with a Pleiades URI. OA has an example for tagging (or 'semantic tagging' - since we're tagging with a URI rather than a free-text term) in their docs . If you think of it like that, it's actually pretty intuitive that the "body" is the tag, and the target is the document/item that is the being annotated. In that sense, it's also reasonable to think of the *target* as the thing of which you are the creator or keeper, rather than the body (which is the tagging vocabulary). The problem IMO, is just that the metaphor of "the body being about the target" is a bit shaky when it comes to tagging in general (but not the technical definitions of body and target).

This is precisely the issue.  Is an image a depiction of a landmark (image = body, landmark URI = target), or is the landmark URI a semantic tag for the image?  Either is valid, and the choice depends on the intent of the annotation rather than having an explicit requirement in the model.

As tagging has a long history on the web, I think the current decision to follow that structure makes sense for associating URIs for places with other online documents such as texts and images.

We think of Open Annotation as a framework that communities can adopt, and in doing so there are some decisions as to how the generic framework should apply in the particular use cases.

 The second issue: in Pelagios, we've seen more and more cases where partners use OA to annotate "spatial data", rather than documents (or items). E.g. we have an alignment between Arachne Places and Pleiades Places based on Pelagios annotations. Our Regnum Francorum data is actually an alignment between Johan's places and Pleiades, and there are more examples - including ORBIS I think. The key difference to the "original Pelagios idea" is that these annotations DO NOT annotate documents (like Perseus texts) or items (like Ure Museum objects). I'd argue that they map different gazetteers (or "gazetteer-like datasets") with each other. And here, IMO, the Open Annotation metaphor starts to break. At least it's no longer intuitive to say that "one thing is about the other". (Rather you might say something like "I have a place record that corresponds roughly to Pleiades ID XYZ.)

Indeed.  It seems more like a regular owl:sameAs triple, or conceptual mapping with SKOS.  Not everything is an annotation, although it's tempting to use the same hammer in every situation :)

Hope that helps!

Rob Sanderson

Again, the lines are certainly blurry, and the issue may need more discussion. But I think it may make sense in the future to treat inherently spatial datasets as gazetteers in their own right. And if so, we should align them using the format we drafted in our NYC meeting, rather than using OA.
 
Cheers,
Rainer
 
P.S.: Any feedback on this is highly welcome of course - and the timing would be perfect as we are stilll in the process of drafting both the gazetteer interconnection spec, and the update version of our annotation format!
 
 
 
 
________________________________________
Von: pelagios...@googlegroups.com [pelagios...@googlegroups.com]&quot; im Auftrag von &quot;Karl Grossner [ka...@stanford.edu]
Gesendet: Freitag, 25. Oktober 2013 20:14
An: pelagios...@googlegroups.com
Betreff: [pelagios] OAC body and target
 
Greetings Pelagians and all:
 
When ORBIS published our Pelagios rdf annotations, I was puzzled about the role of Body and Target and now that I'm looking at using the OAC model for some Catalhoyuk stuff, I find I still am.
 
According to W3C, "Typically an Annotation has a single Body, which is the comment or other descriptive resource, and a single Target that the Body is somehow 'about'." So in our case, ORBIS asserts that certain Pleiades Place records are "about" sites in ORBIS. It seems to me the relationship is the reverse; that a gazetteer like Pleiades provides the blank node and core references for a Place that we all annotate. Aren't ORBIS assertions concerning routes to, from and through a place annotations to ("about") that Place record. Why isn't the Pleiades URI the target?
 
from the Cookbook:
"In Pelagios, we use this generic mechanism to "tag" arbitrary online content with Pleiades URIs: the content is the annotation target, the Pleiades URI the annotation body."
 
Speaking generically, if I tag my resource with the Pleiades URI I am somehow associating all current and future annotations the Pleiades URI is "about" with my resource in a different way than if the roles were reversed. The Pleiades core record has references, name variants, connections, and spatial information. Seems to me that record is my target--I want to annotate it with info I consider "about" it; essentially, "my info is about that place." How is it I get to assert the Pleiades record is about anything? I am not its creator nor keeper!
 
Does this make sense? What am I missing? I'm guessing there was discussion at the time of course--any record of it?
 
all best and cheers,
Karl
 
------------------
Karl Grossner, PhD
Digital Humanities Research Developer
Stanford University Libraries
Stanford,CA US
www.kgeographer.org

[Karl Grossner]

Dear Rainer, Leif, Rob, all ---

Thanks, I figured there was a rationale and it is useful to hear it. No need to visit Pelagios' dusty archives! ;^)

Pelagios has been the most successful implementation of linked data in this domain, and that counts for a lot.

Rainer, you say

"And here, IMO, the Open Annotation metaphor starts to break. At least it's no longer intuitive to say that "one thing is about the other". (Rather you might say something like "I have a place record that corresponds roughly to Pleiades ID XYZ.)"

and Rob has said

"Indeed.  It seems more like a regular owl:sameAs triple, or conceptual mapping with SKOS."

Simply, that was the point I was raising. I guess I have a place-centric view, particularly when it comes to gazetteers, imagining they are (bottom-up, community-built) authorities in some sense for particular regions/periods. Also, I'm a literalist -- when it comes to semantic computing anyway ;^) So the issue of what a scope note says comes up for me all the time. A good example is the CIDOC definition of Place; it is defined by its authors as purely physical space but frequently sub-classed in ways that are counter to its scope note.

Of course Pelagios has worked splendidly in any case, so I guess my concern has more to do with semantic web dynamic "architecture" writ large than any one implementation.

Again, thanks for the considered responses!

best

Karl

------------------
Karl Grossner, PhD
Digital Humanities Research Developer
Stanford University Libraries
Stanford,CA US
www.kgeographer.org

--
You received this message because you are subscribed to the Google Groups "pelagios" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pelagios-proje...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.