RE: [wsrf] Issue WSRF9

[Steve Graham <sggraham@us.ibm.com>]

yes, that may be true in some cases,
but if the display knew it was maxOccurs>1, the GUI widget might be
different (eg table vs single widget)

sgg

++++++++
Steve Graham
(919)254-0615 (T/L 444)
STSM, On Demand Architecture
Member, IBM Academy of Technology
<Soli Deo Gloria/>
++++++++

"Murray, Bryan P." <bryan.murray@hp.com> 06/22/2004 02:24 PM

To:        Steve Graham/Raleigh/IBM@IBMUS         cc:        <wsrf@lists.oasis-open.org>         Subject:        RE: [wsrf] Issue WSRF9

I think in many cases, a "display
the properties and values" application can display a set of properties
without knowing the cardinality.
 
Bryan
-----Original Message-----
From: Steve Graham [mailto:sggraham@us.ibm.com] 
Sent: Tuesday, June 22, 2004 10:57 AM
To: Murray, Bryan P.
Cc: wsrf@lists.oasis-open.org
Subject: Re: [wsrf] Issue WSRF9


I agree with Bryan that the "display the properties and values"
is the most "obvious" use case to suggest some sort of runtime
support for additional meta-data on "dynamic" resource properties.
 There may be other use cases, but  I again agree with Bryan's
observation of the need for the application to have a-priori understanding
of the semantics of those resource properties. 

So, in the case of "display the properties and values" the application
would need to know not only the xsd details of the GED (eg type), but also
the cardinality of the resource property use of the GED (minOccurs and
maxOccurs).  The xsd details of the GED can be found (via WS-MetaDataExchange,
for example).  It is the cardinality (and maybe other details) of
the resource property that need to be discovered. 

sgg 

++++++++
Steve Graham
(919)254-0615 (T/L 444)
STSM, On Demand Architecture
Member, IBM Academy of Technology
<Soli Deo Gloria/>
++++++++

"Murray, Bryan P." <bryan.murray@hp.com> 06/21/2004 01:07 PM

To:        <wsrf@lists.oasis-open.org>        cc:                Subject:        [wsrf] Issue WSRF9

                 I understand the
need to know which properties are currently
present 
                in the property
document given that the property document
supports 
                dynamic properties.
However, I would like to see a use-case
describing 
                the need to retrieve
at runtime the schema for any properties.

                The only reason
I can think of for retrieving properties which
an 
                application does
not already understand is if the application
simply 
                displays properties
to a human or copies the property somewhere.
If 
                the application
understands the namespace of the property it can

                locate the type
from the WSDL/schema for the namespace.
Otherwise, 
                simple properties
(no child XML elements) can be displayed
without 
                understanding their
type by just treating the text node as a
string. 
                The application
can choose to either not display complex
properties or 
                to list attributes
and child elements with their values.
Grandchild 
                elements or attributes
will ultimately make this too complicated
to do 
                at some point.

                Other, more complex,
applications will need to understand not
only the 
                syntax of the properties
they are retrieving, but also the
semantics 
                of those properties
in order to interact with them
appropriately. The 
                syntax can be found
in WSDL/schema files, but the semantics
needs to 
                be understood by
the programmer writing the property retrieval
code. I 
                am confused as
to how an application could interact with a
property it 
                does not understand
beyond simply displaying or copying its
value. If 
                an application
does understand a property that falls into the
xsd:any 
                portion of the
resource property document schema, it does not
need the 
                schema - it already
knows the type of this property.

                If there is no
need beyond displaying or copying the value of an

                unknown property
I suggest that we not provide the ability to
retrieve 
                the schema at runtime.

                Can someone offer
a use-case that shows where my thinking is 
                deficient?

                Bryan