Been so busy looking into the technical details of XMP that I almost forgot to check out the current landcsape. Luckily I chanced on these articles by Ron Roszkiewicz for The Seybold Report (and apologies for lifting the title of this post from his last). The articles about XMP are well worth reading and chart the painful progress made to date:
- The Brief Tortured Life of XMP (July ’05)
- [Thought Leaders Hammer out Metadata Standard] (April ’07)
From the earlier characterization of XMP as “underachieving teenager” Roszkiewicz is cautiously optimistic that IDEAlliance’s XMP Open initiative (an initiative to advance XMP as an open industry specification) will help outreach and foster adoption of this fledgling technology.
There has been some activity here. Following on from an industry open day event last year:
* [IDEAlliance XMP Open Day], New York, March ’06</ul>
there have been two metadata summits earlier this year co-sponsored by Adobe Systems and IDEAlliance:
* [Metadata Directions in Advertising and Branding], San Francisco, January ’07
* [Content Metadata Summit 1.1] New York, March ’07</ul>
Promising bestirrings. (And also with the recent public airing of the PRISM 2.0 draft with its support for XMP which was reviewed at the PRISM WG F2F last week for publication as a standard.) But generally the state of XMP-Ville at this time is rather sleepy. There’s not much by way of news on the [XMP Open] website. At least promise, if no promises.
Back to the articles. The really interesting thing of note (to me at any rate) in Roszkiewicz’s review of the last summit is the almost total absence of any mention of the Web. It is as if XMP users (both consumers and providers) would be content to play within the walled garden of the CS3 product portfolio. I don’t get that. The Web changes everything.
Although XMP maps its native data model to RDF (and RDF is an inherently open technology allowing arbitrary schemas to be mixed at will), XMP betrays its application roots by seeming to want to impose some kind of veto on the schemas to be used. Or rather, how they are to be used. It also seems to be all fussed up by centralized notions such as a cross-mapping schema registry. (As if that were part of its remit.)
As Roszkiewicz notes:
> _“The consortia [IDEAlliance and the stakeholders] will have ownership responsibility for name space registry, cross-map definition and support, standards group outreach and coordination, compliance certification and logo and the “XMP Open” brand.”_
> _“So while the standard for XMP might be defined, the data that will be fed into files is not, for want of an IDEAlliance-like standards management body to filter and rationalize the many [schema] into a few.”_
And then more worryingly, this:
> _“That schema should be managed by a government agency such as the Library of Congress which could manage the dictionaries and schema, certify them, register the namespace and provide a centralized location to distribute them.”_
Well, I don’t see what this matters to the core technology of XMP which is just a specification for the sneaking in of an XML document into arbitrary media files. And the use of RDF/XML would seem to be a further indication that XMP is to be independent of the schema used. The use of both RDF plus XML technologies should allow XMP to present itself as a framework or “platform” for metadata exchange and to get out of the way of what is actually carried by the XMP packets. App neutrality, if you will.
Again the notion of Web as just an alternate channel is apparent in the third of the articles where Roszkiewicz talks about the Device Central tool which allows a user of a CS3 product to “Save for Web or Devices…”. This article talks about the clumsy handling of metadata in such device saves, whereby the packet may be abbreviated - and metadata terms dropped - when printing to small footprint devices. Not a feature to be retained for too long, I would hope.
So, where are we currently with XMP? According to Roszkiewicz:
> _“As the developer of a suite of applications that relies on XMP as the vehicle for managing metadata, Adobe has too much invested in its development to allow any substantive changes by outsiders. So “open” primarily will mean open to suggestions, with an official channel in place to process them.”_
And as to that channel?
> _“As the principal conduit to Adobe for changes to XMP, IDEAlliance will act as a gateway and support organization to the user community - a role for which it is well-suited. … As a sponsor-supported, not-for-profit organization, IDEAlliance can serve as a credible buffer for Adobe to the user community and synchronize and standardize third-party development efforts.”_
And goes on:
> _“The principal unanswered questions at this point are: Will the stakeholders represent all of the key industries; will Adobe provide timely support for considering user input and updating the XMP Toolkit; and can Adobe, IDEAlliance and IDEAlliance workgroups manage all of the responsibilities that will fall upon them when the deal is struck. The hand-over doesn’t seem to have taken place yet, and we are still examining the scope and feasibility of the proposal.”_
It seems to me that Adobe is the party girl, IDEAlliance is the special guest, and Crossref publishers are the neighbourly gatecrashers who want to play with the toys. And not perhaps too nicely neither. I just hope that the toys aren’t taken away from us. They’re too much fun.
Ironic really that we’re on the outside of this since scholarly publishers have a very clearcut grasp of what to do with metadata and a ready application in terms of citation linking. XMP is worth it.