W3C home > Mailing lists > Public > public-openannotation@w3.org > October 2012

Re: F2F Decision: Multiple Resources

From: Bob Morris <morris.bob@gmail.com>
Date: Mon, 29 Oct 2012 16:12:09 -0400
Message-ID: <CADUi7O6MrejdTJdLpZun2Rx79ohnr6NnF=6Ppzr-q3E8-2fm7Q@mail.gmail.com>
To: Jacob Jett <jgjett@gmail.com>
Cc: Robert Sanderson <azaroth42@gmail.com>, Leyla Jael García Castro <leylajael@gmail.com>, public-openannotation@w3.org
Not quite sure what you are suggesting. I \think/ it is exactly what I
meant: the second annotation hasTarget the first annotation.

If, on the other hand you are suggesting that the second annotation
has the same Target as the first but merely adds Body assertions, then
I think the model is delicate. We have actionable annotations
semi-autonomously flying around a P2P network with a semantic Pub/Sub
facility. We normally try to resist intentionally adding explicit
triples to a published annotation, because it can sometimes makes it
hard for the subscription notification component to detect that
something relevant happened. (We haven't explored that problem very
much yet....).  Also, there is an issue if several response
annotations enter the network, each with different assertions about
which is the right choice.  Although I think it's not restricted to
oa:Choice issues, we know that actionable annotations are sometimes
subject to semantic or chronological race conditions, and that the
annotation provenance is quite critical to agents that change their
back-end data (for us, physical specimen metadata).

On Mon, Oct 29, 2012 at 3:22 PM, Jacob Jett <jgjett@gmail.com> wrote:
> ...but couldn't this already be done in the model by annotating the initial
> multi-body (oa:Choice) annotation? More specifically, couldn't we annotate
> one of the choices and use the context predicate "oax:annotatedIn" to
> capture the annotating an annotation bit?
>
> It seems like this might be very useful for your use case where, if we were
> to model the different distinct possibilities as one oa:Choice, you could
> then add an annotation targeting one of the objects of an oa:item predicate
> within that choice, e.g., 'this is our default choice and here is why', and
> use oax:annotatedIn to denote that this annotation noting a default choice
> and the reasoning why (i.e., the evidence), was made in the context of an
> oa:Choice.
>
> The modeling is complex but it seems like this would service the needs of
> your use case without adding additional properties to the specification.
> Does that seem sensible? Seems like there are some other ways to do this
> too, including adding additional properties.
>
> Regards,
>
> Jacob
>
>
> On Mon, Oct 29, 2012 at 1:46 PM, Bob Morris <morris.bob@gmail.com> wrote:
>>
>> A typical use case for us would be a taxonomist's annotation of an
>> image, or of a morphological description,  of  an organism for which
>> the annotator is offering several different possible distinct species
>> as an identification, but with no ability to say which of those
>> species it may be.  In such cases, an "arbitrary" choice on the part
>> of the consumer is not going to be based on preference, but on some
>> kind of scientific evidence which possibly arises from related
>> resources, or even from part of an annotation dialogue.  A second
>> annotator, not the Target publisher,  who acquires the first
>> annotation might well launch an annotation of the form "In Annotation
>> X, I don't know what species the Target is, but I know it isn't
>> Alternative A, and here's my evidence for that."   This is also likely
>> to be a typical biomedical application where the resources are patient
>> examination data and the goal is a medical diagnosis. In fact,
>> taxonomists often refer to the descriptive data that distinguishes one
>> species from another as"diagnostic characters."
>>
>> (Remark: in science, all choices between hypotheses are based on
>> evidence, which is why I'm gathering use cases to make a case for
>> adding Evidence modeling to OA.  I'd even go so far as to suggest that
>> \all/ scholarship needs Evidence sooner or later to support its
>> assertions....)
>>
>> Bob Morris
>>
>>
>> On Mon, Oct 29, 2012 at 11:33 AM, Robert Sanderson <azaroth42@gmail.com>
>> wrote:
>> > The XOR or Choice is to select one and only one of the resources.  For
>> > example, if there are three translations of the same comment, a system
>> > should display only one of them as appropriate for the user's
>> > preferences (and potentially allow the user to se the other options).
>> > On the other hand, given an oa:Set of three comments, all three should
>> > be displayed.
>> >
>> > Hope that helps,
>> >
>> > Rob
>> >
>> > On Mon, Oct 29, 2012 at 4:44 AM, Leyla Jael García Castro
>> > <leylajael@gmail.com> wrote:
>> >> Hi Bob,
>> >>
>> >> Do you have a use case for the ao:XOR? Not so sure whether I understand
>> >> it.
>> >>
>> >> Cheers,
>> >>
>> >> Leyla
>> >>
>> >>
>> >> On Sun, Oct 21, 2012 at 11:16 PM, Bob Morris <morris.bob@gmail.com>
>> >> wrote:
>> >>>
>> >>> With respect to the Multiple Resources model[1] that emerged in
>> >>> Chicago
>> >>>
>> >>> 1. It would be nice if the Issues List reflected what Rob's initial
>> >>> proposal morphed into, and the discussion continued there. (Rob: I'll
>> >>> have a
>> >>> try if you want...)
>> >>>
>> >>> 2. oa:Set and probably oa:List can profitably be applied to  a
>> >>> collection
>> >>> of oa:Annotations.  The use case is actionable annotations that are
>> >>> delivered to remote agents,  and upon which collections of expected
>> >>> actions
>> >>> must taken, possibly in a prescribed order.  This is particularly
>> >>> needed
>> >>> when actionable annotations will generate response annotations (e.g.
>> >>> "Agent
>> >>> Smart accepted all of your corrections in the oa:Set :mySet1 except
>> >>> the
>> >>> oa:item :mySet1.item10.").  If a collection of actionable annotations
>> >>> travels in a disconnected fashion, the annotation publisher can not
>> >>> easily
>> >>> (at all?) convey that a coordinated action is desired.  There may be
>> >>> an
>> >>> argument for ao:XOR on collections of annotations also.  It's likely
>> >>> that
>> >>> none of these collection types should be restricted to Target, Body,
>> >>> and
>> >>> Specifiers, as is perhaps being suggested in [1]
>> >>>
>> >>> 3.  Probably oa:List objects cannot(?) survive being put in a triple
>> >>> store, since order of identified nodes is not defined in the graph.
>> >>> [2] is a
>> >>> proposal to address the issue, but it is unclear how much traction it
>> >>> has.
>> >>> This means that  processing order for oa:List will depend on the
>> >>> serialization, not on the RDF.  I vaguely recall this was raised in
>> >>> Chicago,
>> >>> perhaps tabled for more discussion.
>> >>>
>> >>> [1]
>> >>>
>> >>> http://lists.w3.org/Archives/Public/public-openannotation/2012Oct/0004.html#start4
>> >>> [2] http://www.w3.org/2009/12/rdf-ws/papers/ws14
>> >>>
>> >>> Bob Morris
>> >>>
>> >>> --
>> >>> Robert A. Morris
>> >>>
>> >>> Emeritus Professor  of Computer Science
>> >>> UMASS-Boston
>> >>> 100 Morrissey Blvd
>> >>> Boston, MA 02125-3390
>> >>>
>> >>> IT Staff
>> >>> Filtered Push Project
>> >>> Harvard University Herbaria
>> >>> Harvard University
>> >>>
>> >>> email: morris.bob@gmail.com
>> >>> web: http://efg.cs.umb.edu/
>> >>> web: http://etaxonomy.org/mw/FilteredPush
>> >>> http://www.cs.umb.edu/~ram
>> >>> ===
>> >>> The content of this communication is made entirely on my
>> >>> own behalf and in no way should be deemed to express
>> >>> official positions of The University of Massachusetts at Boston or
>> >>> Harvard
>> >>> University.
>> >>>
>> >>
>>
>>
>>
>> --
>> Robert A. Morris
>>
>> Emeritus Professor  of Computer Science
>> UMASS-Boston
>> 100 Morrissey Blvd
>> Boston, MA 02125-3390
>>
>> IT Staff
>> Filtered Push Project
>> Harvard University Herbaria
>> Harvard University
>>
>> email: morris.bob@gmail.com
>> web: http://efg.cs.umb.edu/
>> web: http://etaxonomy.org/mw/FilteredPush
>> http://www.cs.umb.edu/~ram
>> ===
>> The content of this communication is made entirely on my
>> own behalf and in no way should be deemed to express
>> official positions of The University of Massachusetts at Boston or
>> Harvard University.
>>
>



-- 
Robert A. Morris

Emeritus Professor  of Computer Science
UMASS-Boston
100 Morrissey Blvd
Boston, MA 02125-3390

IT Staff
Filtered Push Project
Harvard University Herbaria
Harvard University

email: morris.bob@gmail.com
web: http://efg.cs.umb.edu/
web: http://etaxonomy.org/mw/FilteredPush
http://www.cs.umb.edu/~ram
===
The content of this communication is made entirely on my
own behalf and in no way should be deemed to express
official positions of The University of Massachusetts at Boston or
Harvard University.
Received on Monday, 29 October 2012 20:12:37 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 29 October 2012 20:12:37 GMT