OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

office message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: RE: [office] FW: XMP on the front burner as usable subset of RDF

For the record:

I realize after reading my post below that it may be perceived that I am in a potential conflict of interest situation WRT discussions within this TC adopting RDF, DC and/or XMP, given that XMP is an Adobe creation based on RDF.  

Please note that I am taking every caution to wear my standards hat in this TC and will always do what is best for the TC from a "technically pure" perspective.  I will also voluntarily remove myself from *any* vote or process in a situation where it may be perceived that I have a conflict of interest.  If such a situation arises and I miss it, please let me know and I will abstain.  I will not feel at all upset by anyone requesting this.

Adobe has the highest ethical standards WRT conducting business and I feel both honored and compelled to uphold these when I work in standards.

You should also note that Adobe does not "sell" XMP, it is licensed under RF terms.  

Duane Nickull 

-----Original Message-----
From: Duane Nickull [mailto:dnickull@adobe.com] 
Sent: Tuesday, September 27, 2005 12:25 PM
To: office@lists.oasis-open.org
Subject: [office] FW: XMP on the front burner as usable subset of RDF


I had a meeting with the XMP and RDF gurus in Adobe's San Jose HQ.  They were kind enough to provide some feedback and I think we can seek further aid as required.

Feedback on proposal for RDF in OD is below.  Where I paraphrased, I used [] as a convention.

"There are a lot of devils in the details, often missing from the
proposal and often misleading from implications in the proposal. It
is not a great idea to "push RDF". I don't know what the current Open
Office standard says about metadata. I think this proposal needs a
lot of work and clarification."

DN: this was largely in response to my proposal and I acknowledged that we had not contemplated all the details yet.  I think this is a warning we should heed.  He continues:

"When dealing with XMP it is very important to first understand the
XMP data model. XMP ***USES*** a small subset of RDF. XMP is
***NOT*** equivalent to RDF. Another way to say this is that the XMP
data model can be expressed using a small subset of RDF. 

[Paraphrased] The adoption and use of XMP could be
considered. It really depends on the Open Office folks. If they like
RDF, fine. If they don't like RDF because it does not play well with
XML Schema, then XMP is an option. I have not looked at the
Relax NG schema language, I don't know if it works with RDF or how it
would work with Plain XMP."

DN: I really felt like this was an honest and unbiased opinion of how we should move forward based solely on a quest for technical excellence.

"[Further clarification on Duane's proposal:]

> Dublin core is a metadata language  that can expressed in XML.

Dublin Core is not "a metadata language". RDF is a metadata language.
HTML META tags are another metadata language. Dublin Core is a set of
properties that can be expressed in a variety of metadata languages.
I don't think there is any single official way to express the Dublin
Core properties in RDF, see the example below. At any rate, XMP
certainly differs from common DCMI usage of RDF.

> XMP is used to wrap around RDF, Dublin core and many other forms of
> metadata.

XMP does not "wrap around" RDF. XMP is expressed in a small subset of
RDF. All valid XMP can be expressed in RDF. There is a lot of RDF
that is not valid XMP."

DN:  I had stated this before.  It was an error.  He is correct.

"> Adobešs XMP SDK contains a complete Dublin Core schema and can be
> used immediately to accomplish the task.

The XMP expression of Dublin Core differs from that typically shown
by the DCMI. This is a direct fallout of the fact that the XMP data
model does not encompass all of RDF. At this point reasons are moot,
but one big reason for this was probably restrictions to make XMP
reasonable to implement. (I was not around when XMP was created.)

A prime example of this is that XMP does not allow repeated
properties, it requires explicit use of arrays. While the DCMI folks
commonly use notation like (ignoring xmlns declarations):

         <rdf:Description rdf:about="...">

In XMP this would have to be:

         <rdf:Description rdf:about="...">

[There is] a lot of software in the field that wants the 2nd form.
Bridging these differences is technically possible, but is the kind
of thing that requires planning up front and long rollout periods to
get past incompatibility humps.

------ End of Forwarded Message


Based on this feedback, I favor a more tightly constrained set of RDF to be specified for OD metadata.  I am aware that this may be perceived as a conflict of interest coming from an Adobe employee.  Please be assured that I am making every attempt to remain as open and neutral as possible to the best technical solution for our TC's requirements.  I will also voluntarily remove myself from any voting where there may be a conflict of interest or perception thereof.  TC members should also be aware that we do not charge any money for XMP and the license for its' use is very open and RF.

Moving forward:

Does the use of a constrained, syntax-specific subset of RDF make sense for OD?  I would state yes since we need a few items such as a consistent representation of metadata, preferably in plain text format.


To unsubscribe from this mail list, you must leave the OASIS TC that
generates this mail.  You may a link to this group and all your TCs in OASIS

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]