Metadata Working Group

[Martin_Doudoroff]

I wanted to draw everyone's attention to today’s announcement that the Metadata Working Group (Adobe, Apple, Canon, Microsoft, Nokia & Sony) have released their first specification (1.7MB PDF) for metadata handling. The document is written for software developers, but is an interesting skim for everybody because it does discuss some of the pervasive problems with image metadata and will give you some perspective on how metadata is handled differently across different file formats.

This first specification focuses on how software should preserve and modify metadata when it loads and saves files. One of the pervasive problems this specification is intended to address is the loss of metadata because one software tool neglects to copy it from the source file to the output file, or otherwise corrupts it while the user is processing the file.

What the spec does not do is attempt to standardize any of the data itself, although pages 16-18 contains a general catalogue of the gruesome mess that many of the most common metadata fields have become (e.g., keywords), so they're certainly acknowledging there are more problems to solve.

As far as it goes, there's nothing in this spec that jumped out at me as particularly self-serving or lowest-common-denominator. My biggest concern is that even if the biggest players are serious about adopting this spec and conforming their own products to it, smaller players, third party libraries and open source libraries might not reach parity for years, if ever, simply because of how software development works (or doesn’t). The good news is that the major OS developers (Microsoft, Apple) are increasingly providing sophisticated image file i/o APIs that other software developers can use. If Microsoft and Apple are diligent about keeping these APIs up-to-date and in following the specs of their own working group, then there will be less need for third party solutions and new standards like these will have a prayer of success. (Of course, legacy software will continue to butcher metadata until abandoned.)

[larry_angier]

As far as metadata, early versions of Lightroom would strip out the IPTC data while building web galleries for smaller files. This happened during the "Orphan Works" legislation discussions in Congress in the past couple of years. I and many others alerted Adobe about this oversight and now the option is to leave all the metadata intact.

Just getting the metadata embedded in a file is a hassle as it is and then having the software strip it out is unconscionable. My own solution is Photo Mechanic at ingest. IngesterPro as a front end for Lightroom is also good and I'm sure there are other good software packages that do the same.

Metadata is one of the most important aspects of digital photography. Without at least minimal metadata, your image is invisible and cannot be found, let alone attributed to your creation. At the very least, you need to put your name and contact info within each image. Better yet is to caption, keyword, put in the location, and rename each image.

As an example of how important just renaming a file is, do a search for "DSC", the default name of nearly every digital camera, in Google and you'll get 58 million images. Which one is yours?

One multi-organization group discussing and setting the standards for metadata is UPDIG.com. A good reference is Peter Krogh's The DAM Book in conjunction to his website with more current info on this necessary aspect of digital.

It's good that software and hardware manufacturers are getting on board. The easier it is to implement metadata the more we'll all use it and the more likely hood that our images won't end up lost in cyberspace.

[Martin_Doudoroff]

Thank you for the info, Larry; the UPDIG group is actually at Universal Photographic Digital Imaging Guidelines - UPDIG.