RE: Difference between data and metadata

I have found that meta data is hierarchical in nature, the hierarchy is 
related to the context of the data. A number of descriptions may 
contribute to the meta data associated with a specific piece of data.
A web service message exchange may have meta data expressed directly in 
associated WSDL, or even with the contents of the message depending on the 
implementation semantics, there may be higher level orchestration meta 
data as well and there may be higher level business functions above that. 
And there may be other meta data that at the service level represented 
that currently is specific to the business but outside the scope of 
current standards.
Where do I locate this meta data, architecturally I figure that WSDL is a 
good entry point to this meta data stack from which you could drill up or 
down for additional information. There are other starting points you could 
take as well.


Gavan





Jonathan Marsh <jmarsh@microsoft.com>
Sent by: www-ws-desc-request@w3.org
09/27/2006 04:02 AM

 
        To:     Ramkumar Menon <ramkumar.menon@gmail.com>, "Rogers, Tony" 
<Tony.Rogers@ca.com>
        cc:     "paul.downey@bt.com" <paul.downey@bt.com>, "christian.au@gmail.com" 
<christian.au@gmail.com>, "www-ws-desc@w3.org" <www-ws-desc@w3.org>
        Subject:        RE: Difference between data and metadata


Good point.  It illustrates that the data/metadata relationship holds 
within a context.  In the context of web service message exchanges, you 
could say the messages being exchanged are data and the descriptive 
artifacts such as WSDL and Schema are the metadata.  But within the 
context of a single message, there may data (like a payload) and metadata 
about that data (like, whether it is authentic).  Another good example is 
xml:lang.  It provides metadata about data within the context of an XML 
element ? and the XML element itself can be thought of as metadata about 
the structure of the data it contains.
 

From: Ramkumar Menon [mailto:ramkumar.menon@gmail.com] 
Sent: Monday, September 25, 2006 7:31 PM
To: Rogers, Tony
Cc: paul.downey@bt.com; Jonathan Marsh; christian.au@gmail.com; 
www-ws-desc@w3.org
Subject: Re: Difference between data and metadata
 
Hi All,
 
My 2 cents.
 
1) In the context of web services, Messages that are sent between, or 
among endpoints could be termed as data. 
2) Metadata describes the what, how and where for the data [message].Note 
that in this case, metadata should be interpreted as more of metadata for 
the interaction, than about just being [but includes] the 
metadata/description for the data[message]. For instance, where the 
message is originating from, where its heading to, how to secure the 
message, description of the message,how to perform message non-repudiation 
etc are considered metadata. 
3) Wire messages could contain data, possibly along with the metadata of 
the data that is being sent across, as headers, as a part of the message 
itself, or as references that can be resolved by the consuming 
entity/intermediaries. 
 
With this perspective, 
a) WSDL is not really ". the data which is enhanced via metadata", as Christian refers to, but is 
the metadata itself, since it describes the message, describes the service 
that supports the message exchange etc, and describes where the message 
can be addressed to etc. 
b) WS-Security too, with the support of point in (2) and (3) are metadata 
.
3) WS-CDL, BPEL etc are metadata of your business processes/collaborations 
that define an observable message exchange [exchange of data] among 
parties in a distributed environment. Hence, they too fall under metadata. 
 [of the interaction]
 
-Ram
 
On 9/25/06, Rogers, Tony <Tony.Rogers@ca.com > wrote: 
I was with you most of the way, but not all the way to the end. :-)
 
I consider metadata as the descriptive information about the data, rather 
than just the stuff that isn't data.
 
If we consider a packet travelling over the Net, carrying the data, we 
don't consider the start and end markers as data, but we don't consider 
them metadata, either. You can argue that the address fields in the packet 
are metadata (they "describe" the destination), but mostly they would be 
left out of consideration, too. The metadata here is more like the 
standard describing the IP part of TCP/IP. 
 
Along those lines, I'd consider the element names to be metadata when they 
appear in the XSD (for XML schema), but not when they appear on the wire 
(on the wire they are data for the XML processor, allowing it to separate 
the pieces of the XML data). 
 
Hmm - not my most articulate description, but maybe I've explained how I 
see it?
 
I definitely (and precisely) agree that /precise/ agreement isn't easy 
here :-)
 
Tony Rogers
CA, Inc
Senior Architect, Development
tony.rogers@ca.com
co-chair UDDI TC at OASIS
co-chair WS-Desc WG at W3C
 

From: www-ws-desc-request@w3.org on behalf of paul.downey@bt.com
Sent: Mon 25-Sep-06 21:05
To: jmarsh@microsoft.com; christian.au@gmail.com; www-ws-desc@w3.org
Subject: RE: Difference between data and metadata

 
 
I'd agree and warn it's something which we are all unlikely to
/precisely/ agree upon.

WSDL seems like classic meta-data in that the receiver
is oblivious of how the sender came to know how to 
produce correctly formatted messages - the knowledge may have
been obtained using any one of a number of different
out-of-band methods, including consuming a WSDL document.

However, HTTP Content-type might also be considered by some to be 
meta-data, as might XML element names and other descriptions
which appear 'on the wire' and directly influence processing.

Paul

-----Original Message-----
From: www-ws-desc-request@w3.org on behalf of Jonathan Marsh
Sent: Fri 9/22/2006 10:11 PM
To: Christian Au; www-ws-desc@w3.org 
Subject: RE: Difference between data and metadata

>From a Web Services point of view, I believe what travels on the wire is 
data (e.g. SOAP messages in HTTP envelopes).  What describes what goes on 
the wire is metadata (WSDL, WS-Policy, CDL).  In a few instances metadata 
might appear on the wire, but generally only to facilitate future 
interactions, not to make that particular exchange work. 

I believe WSDL is purely metadata, through the thought experiment.  If I 
took the WSDL away, would the described on-the-wire exchange still be able 
to occur?  In almost all cases an omniscient coder could still craft the 
required messages and successfully exchange them.  The metadata, in its 
various flavors, is not strictly necessary to the successful exchange, but 
describes different parts of the exchange in formats that allow an 
ignorant coder ( e.g. a toolkit) to raise it's level of awareness towards 
omniscience.

We all know there are limits to today's metadata formats, especially in 
terms of legal obligations and business purpose, that will continue to be 
indescribable.  This information is missing (in machine-readable form) and 
yet the exchange can still take place, thus it also falls under my 
definition of "metadata". 

Hope this helps.

________________________________
From: www-ws-desc-request@w3.org [ mailto:www-ws-desc-request@w3.org] On Behalf Of Christian Au
Sent: Thursday, September 21, 2006 1:48 AM 
To: www-ws-desc@w3.org
Subject: Difference between data and metadata

Dear all,

I am kind of confused about the definition of metadata in the context of 
web services and would be very grateful for any clarification. 

First question would be, if the WSDL-specification already is metadata 
regarding to my web service or primary a definition of my web service ( 
i.e. the data which is enhanced via metadata)?  WS-Policy and WSDL-S 
annotations clearly seem to be metadata, but what about a description of 
the choreography with ws-cdl or a definition of security mechanism with 
ws-security? data or metadata? 
how would you classify standards like ws-management and 
ws-reliablemessaging?

In the standard I could not find any clear specification of this term...

Thanks and kind regards,
Christian






-- 
Shift to the left, shift to the right!
Pop up, push down, byte, byte, byte!

-Ramkumar Menon
A typical Macroprocessor 

Received on Tuesday, 26 September 2006 21:47:13 UTC