[oslc-core] Questions/remarks about the delegated UI for resource selection

Steve K Speicher sspeiche at us.ibm.com
Thu Oct 13 16:44:30 EDT 2011


Hello,

I have provided some remarks below...

Thanks,
Steve Speicher | IBM Rational Software | (919) 254-0645

> From: Thunissen Marc <marc.thunissen at oce.com>
> To: <oslc-core at open-services.net>, 
> Date: 10/12/2011 05:38 AM
> Subject: [oslc-core] Questions/remarks about the delegated UI for 
resource selection
> Sent by: oslc-core-bounces at open-services.net
> 
> Hello,
> 
> I have a few remarks about the delegated UI for the resource selection:
> 
> 1.       The specs are unclear about the possibility to select one or 
more resources.
> The explanations are “allow a user to pick a resource” (one resource), 
but 
> the example show a response with two resources.
> My feeling about that is that the UI consumer should be able to choose 
the 
> preferred behavior: single or multiple selection.
> Can we imagine an option in the UI URI, such as 
“oslc.options=singleSelection” or “
> oslc.options=multipleSelection” ?

This is always possible, though I wonder what the scenario is that 
requires this.  I have not heard the demand for this but do recall it 
being discussed at one point.

> 2.       As a UI consumer, I also have a need to restrict the resource 
selection.
> I would like to allow the user to select an asset, but only for a 
specific asset type.
> By instance we are in a workflow to create a request for tender and I 
want 
> to allow the user to select a template for the document to create.
> I could be useful to allow the “oslc.where” in the UI URI.
> Something like and oslc.where=dcterms:type=”Request for tender template” 

> could be a solution.

I would have thought that you could accomplish this by listing multiple 
oslc:selectionDialogs (one per resource type).  The consumer would pick 
the right selection dialog based on the desired type.  Does this not work 
for you as is?

> 3.       A common design pattern is to build the UI on top of the API.
> It could be interesting to be able to create a generic UI layer using 
> metadata coming from the API to build the creation and selection HTML 
forms.
> We can use the resource shape, but I cannot find a kind of “display 
name” 
> for the properties to create the form labels (and of course we need to 
care 
> about the localization).

These are available from resources from their optional oslc:instanceShape 
reference and from that you can use oslc:Property's dcterms:title
See: 
    
http://open-services.net/bin/view/Main/OSLCCoreSpecAppendixA#OSLC_Properties
    
http://open-services.net/bin/view/Main/OSLCCoreSpecRDFXMLExamples#Shape_Resources

> 4.       I have some oslc resources for which some properties must be 
> specified when creating a new resource but cannot be changed later.
> I am not sure how to declare this in the shape.
> Should I have to write “oslc:readOnly=true”, “oslc:occurs=http://open-
> service.net/ns/core#Exactly-one” and no “oslc:defaultValue”.
> But there are also some resources, such as the “dc:identifier”, that can 

> neither be set at creation time nor changed later.
> How can I mark the difference in the shape (still trying create the UI 
> automatically) ?

Yes would need different shapes for different purposes

For the shape associated with the creation factory, setup the requirements 
as:
   “oslc:readOnly=false”, 
“oslc:occurs=http://open-service.net/ns/core#Exactly-one” and no 
“oslc:defaultValue”

Then for the query and/or instance shapes do this:
   “oslc:readOnly=true”, 
“oslc:occurs=http://open-service.net/ns/core#Exactly-one” and no 
“oslc:defaultValue”
(mark readOnly as true)




More information about the Oslc-Core mailing list