W3C home > Mailing lists > Public > public-openannotation@w3.org > March 2013

Re: May a Specific Resource be object of the oa:hasScope relationship ?

From: Lutz Suhrbier <l.suhrbier@bgbm.org>
Date: Mon, 18 Mar 2013 14:27:50 +0100
Message-ID: <51471656.60900@bgbm.org>
To: public-openannotation@w3.org
Hi Rob,

I don't think that our interpretations of the multiple body concepts are 
different.
Only regarding choices, I understood that at maximum one of the specific 
resources can be selected by the client application. Your definition 
below sounds like it may be more than one ?

As in my use case, all SpTargets must be applied, and also all SpBodies 
must be applied as well, the Composite appears to be the right concept 
to express this.

What's about the interpretation of oa:hasContext related to a SpTarget 
within SpBodies in general, and within SpBodies in a Composite in 
particular. As there are no samples in the spec or the cook book, my 
interpretation is about the same as the image in a web page example. 
Here, the image specified in the SpTarget must be seen by the client 
application in the context of the web page. If the application's body 
would be applied to the same image, but within another web page, than 
the annotation consuming client application might conclude invalid 
results from that annotation. If I related now a Body to a SpTarget, 
than it would mean that the content of that body must be seen within the 
context of that SpTarget, or not ?
I think, that use case has never been discussed in the mailing list. As 
the spec states, that oa:hasContext can be applied to any SpResource, 
somehow, the question is, what do I express, when I use oa:hasContext 
with an SpBody and if I related the context to a SpTarget ?

Lutz

> Sorry, I wasn't clear.  All of the resources apply together as a
> single composite resource.
>
> The distinctions, given two bodies B1 and B2 and a single target T1:
>
> * Multiple uses of hasBody:   Each applies individually.
>      B1 annotates T1, B2 annotates T1
> * Choice:  All apply, but only one needs to be treated at any point in time.
>      (B1 or B2) annotates T1
> * Composite: All apply together as a single whole.
>      (B1 and B2) annotates T1
> * List: All apply together as a single whole, and in a particular order
>      (B1 then B2) annotates T1
>
> Thus (B1 annotates T1 and B2 annotates T2 and-not B1 annotates T2) is
> not possible to express within a single annotation. It is necessary to
> express this using multiple annotations, even if the annotator and
> annotation creation time and process is the same for both of them.
>
> Rob
>
> On Fri, Mar 15, 2013 at 5:19 PM, Lutz Suhrbier <l.suhrbier@bgbm.org> wrote:
>> Hi,
>>
>>> Hi Lutz,
>>>
>>> That's not possible.  The semantics of a oa:Composite is that *all* of
>>> the resources apply.
>> Yes, that is what I need. I want that all Specific Bodies will be applied.
>> But I need to prevent, that each body is applied to all Specific Targets.
>> Only to that target identifying the data record element, the user entered
>> the specific comment stored in that body for.
>> Beyond that, an annotation in our use case can comprehend multiple data
>> record elements which we can not be subdivided in multiple annotations,
>> because the are semanitcally related to the context of that one annotation
>> (e.g. Geolocations). Also, they are entered by the user all together as a
>> single annotation to a data record.
>>
>> I expected, the hasScope property as an adequate means covering exactly
>> those relationships, considering it like a kind of "if"-clause (if context
>> then apply body). Obviously, I misinterpreted that property.
>>
>> Unfortunately, I do not see any other options in the current spec to express
>> that use case.
>> So, looks like either I apply this in my domain specific way but OA
>> unconformant way, or I do have to invent my own "body language" in terms of
>> a dataset, defining a list of XPath expression - value pairs and let clients
>> apply that list to the source document. Hmm, would be also highly domain
>> specific and .... bad news.
>>
>> May be, does anybody have an idea to make our use case conformant with OA
>> 1.0 ?
>>
>> Lutz
>>
>>
>>
>>> You would need to mint multiple annotations in order to have "aaa"
>>> apply only to doc#/a and "bbb" apply only to doc#/b.
>>>
>>> You could have multiple bodies, but then each of aaa and bbb would
>>> apply separately to /a and /b.
>>>
>>> Rob
>>>
>>> On Fri, Mar 15, 2013 at 4:28 PM, Lutz Suhrbier <l.suhrbier@bgbm.org>
>>> wrote:
>>>> Hi Paolo,
>>>>
>>>> Hi Lutz,
>>>> I am not sure I understand completely what you are doing as I see only an
>>>> abstract example.
>>>>
>>>> I would probably use a SpecificResource
>>>> http://www.openannotation.org/spec/core/specific.html#Specific
>>>> in combination with a FragmentSelector
>>>> http://www.openannotation.org/spec/core/specific.html#FragmentSelector
>>>> therefore I would use a oa:hasSource rather than oa:hasScope.
>>>>
>>>> Oh yes, I overlooked  the possibility of using XPointer as fragment
>>>> selector. I am currently not quite sure if it is allowed to be used with
>>>> URNs too(we can not use URLs to identify data records in or use case),
>>>> but
>>>> at a first glance, it looks that the standard does not forbid that. So, I
>>>> don't need to define a XPath Selector anymore. Good News.
>>>>
>>>>
>>>> A typical use of oa:hasScope is when you are annotating an image (with a
>>>> URI) displaying in a webpage.
>>>> The oa:hasSource is the image, the selector is for instance a
>>>> oa:SvgSelector
>>>> and the oa:hasScope is the webpage.
>>>>
>>>> Hmm, I know that my use case is not quite typical as all recent
>>>> discussions
>>>> went around that image in a web page example. That's why I am asking to
>>>> the
>>>> list, if it would be "forbidden" to use oa:hasScope in the way I
>>>> described ?
>>>> So, can I use oa:hasScope within a Specific Body  to point at the
>>>> (resolvable)URI of a Specific Target, which defines somehow the context
>>>> in
>>>> which the scoped body content should be seen by a client (application) ?
>>>>
>>>> May be a short example:
>>>> Lets have a XML-document urn:uuid:doc01 as source. In my annotation I
>>>> define
>>>> 2 Specific Targets aggregated in a Composite,
>>>> urn:uuid:doc01#xpointer(/a)
>>>> and urn:uuid:doc01#xpointer(/b).
>>>> Now I have two Specific Bodies also aggregated in a composite, one is
>>>> holding contentAsText:"aaa", the other "bbb" and I want that the one
>>>> holding
>>>> "aaa" should only be applied in context with Specific Target
>>>> urn:uuid:doc01#xpointer(/a) , the one holding "bbb" only with Specific
>>>> Target urn:uuid:doc01#xpointer(/b).
>>>> Finally, I want to add property "oa:hasScope urn:uuid:doc01#xpointer(/a)"
>>>> to
>>>> Specific Body "aaa" and "oa:hasScope urn:uuid:doc01#xpointer(/b)" to
>>>> Specific Body "bbb" in order to express this with OA 1.0.
>>>> Is this possible ?
>>>>
>>>> Lutz
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> Hope it helps,
>>>> Paolo
>>>>
>>>> On Fri, Mar 15, 2013 at 10:06 AM, Lutz Suhrbier <l.suhrbier@bgbm.org>
>>>> wrote:
>>>>> Hi,
>>>>>
>>>>> in our project, we are dealing with record data which is usually
>>>>> exchanged
>>>>> in domain specific XML documents. The system enable users to annotate
>>>>> these
>>>>> record data, where one annotation may comprehend  multiple record data
>>>>> elements. Therefore, annotations will be recorded as a set of XPath
>>>>> expressions, each of them mapped with annotated value and comment
>>>>> information. (same use case as discussed in
>>>>>
>>>>> http://lists.w3.org/Archives/Public/public-openannotation/2012Sep/0017.html).
>>>>>
>>>>> Now, with OA 1.0 released, i started to remodel our use case with the
>>>>> new
>>>>> version. The new multiplicity concepts were very helpful for this and
>>>>> solved
>>>>> most problems arising in the last discussion from September.
>>>>> So,  the list of XPath expressions for a given target can now be
>>>>> organised
>>>>> by a Composite of Specific Target resources, where I introduced a simple
>>>>> XPath-Selector to hold the XPath expression addressing  the annotated
>>>>> record
>>>>> data element. Likewise, I organised Specific Body resources in a
>>>>> Composite,
>>>>> where each Body holds the corresponding annotated value and comment
>>>>> information for one of the Specific Targets.
>>>>> Now, I want to use the hasScope relationship to outline the Specific
>>>>> Target a Specific Body relates to. Or in terms of OA, to specify a
>>>>> Specific
>>>>> Target as object of the oa:hasScope relationship that somehow scopes or
>>>>> provides the context for the (Specific Body) resource in this Annotation
>>>>> (i.e. Xpath expression within a record data document).
>>>>>
>>>>> Finally, as all samples contain just webpages or images as resources, my
>>>>> simple question is if my interpretation of the OA 1.0 document is right
>>>>> and
>>>>> using it that way would not contradict to the intended usage of the
>>>>> hasScope
>>>>> property ?
>>>>>
>>>>> Please find attached two figures, one depicts our data model, and the
>>>>> other our corresponding OA representation.
>>>>>
>>>>> Best regards
>>>>> Lutz
>>>>>
>>>>
>>>> --
>>>> Dr. Paolo Ciccarese
>>>> http://www.paolociccarese.info/
>>>> Biomedical Informatics Research & Development
>>>> Instructor of Neurology at Harvard Medical School
>>>> Assistant in Neuroscience at Mass General Hospital
>>>> Member of the MGH Biomedical Informatics Core
>>>> +1-857-366-1524 (mobile)   +1-617-768-8744 (office)
>>>>
>>>> CONFIDENTIALITY NOTICE: This message is intended only for the
>>>> addressee(s),
>>>> may contain information that is considered
>>>> to be sensitive or confidential and may not be forwarded or disclosed to
>>>> any
>>>> other party without the permission of the sender.
>>>> If you have received this message in error, please notify the sender
>>>> immediately.
>>>>
>>>>
>>
>
Received on Monday, 18 March 2013 13:28:29 GMT

This archive was generated by hypermail 2.3.1 : Monday, 18 March 2013 13:28:29 GMT