Economics of Trust and Permanent Metadata

Recently, there have been some great concern regarding Orphan Works legislation within the photography community.  There is a lot FUD going around and attempts to address it - some of these discussion ultimately come down to the need for "permanent metadata" or more security ala DRM.

But this isn't an issue about security - the issue is about business - which is about trust between buyer and seller - making sure that the transaction can occur as smoothly, easily and robustly as possible.

Remember the time when there weren't UPC labels (I barely do) - people went around tagging assets in the stores with prices, someone had to retype each tag into another machine to process a transaction. Time consuming but even price tagging was a revolution in metadata and efficiency over haggling over the price of an asset. UPC labels took it a quantum step further so that even today there are self-scan, self-pay stations - talk about trust!

Yeah, there still are thieves and people who switch tags but there always will be - it's a cost of doing business. It doesn't make
economic sense to shop for groceries in a bank vault.

There is a lot we can do to move the industry from haggling to UPC automation - so to speak. But it will involve opening up the pathway for information flow throughout the Adobe toolset and making it robust but only to the point it makes economic sense.

This is where the DRM argument comes into context - it doesn't make economic sense depending on the business transaction - In the grocery analogy, my oranges don't have a security tag around it but my $100 bottle of champagne might (if you live where I do). But ultimately it's the seller's choice.

Increasing the information flow mean preserving the metadata wherever possible, across file transformations, copy paste, compositing, etc.

Making it robust means, adding IDs (like UPC labels) that stay with the file and can be resolved to an owner. It may also mean storing those IDs in more robust places in the file in case the metadata gets accidentally stripped.

Yes ubiquity and security pull in opposite directions, but it's about finding the right tension based on the economics of business which is ultimately the economics of trust.

What do you think Adobe could do to move the industry forward?

Technorati Tags: , , , ,

8 Responses to Economics of Trust and Permanent Metadata

  1. I like your thinking. A single user identifiable DRM signature. Wouldn’t that be great. I could tag all my images and simply look for them in Google. Yeah sure there would be ways of circumventing it but something like this should be possible.I don’t agree completely with your analogy for security tags as they simply cost too much and it takes time for staff to put them on and take them off again.DRM written at the time of taking or post processing, sounds Great..

  2. Whether it’s permanent metadata or DMR, I think a system of permanently tagging images with ownership information (well, as permanent as possible, all systems can be circumvented) would be a good thing. While this will not stop image piracy, perhaps it will make it harder to do and easier to prove ownership of pirated images.

  3. Adobe could be talking with Google to push to make metadata searchable.It is weird that it isn’t! and Google’s image search kind suffers because of this.If Adobe was like – hey Google, we’re working on encouraging folks who use our products to embed metadata by making it super easy to do so, and if loads of people are doing this, as we think they will be, Google might want to consider reading image metadata as part of the indexing that it does because ultimately it will improve search results for images on Google….[Gunar] We’ve made them aware of XMP and the FREE toolkit to allow them to read and write XMP across multiple file formats. Don’t put the blame on Adobe for not doing enough.The issue is with spam. Google doesn’t look at the tag in web pages for good reason. Why should they look at the metadata in XMP?

  4. Chiz Dakin says:

    One thought I’ve had along these lines in relation to making the copyright owner of an image permanent in metadata to prevent photographers images getting “accidentally” or otherwise orphaned – especially by metadata stripping – is the following…When you get a new (or second hand) camera, part of the registration process would involve plugging the camera into the PC via USB/firewire cable. Via a browser connected to the manufacturers registration site, a java etc widget reads the camera’s serial number and through a secure connection adds the copyright symbol and your name (as entered on the registration form) into a “no-other-way-editable-possible” field within the camera’s software (similar but more robust example to how its possible to add the users name in Canon software at the mo (iirc?))Doing it this was means that the person changing the info has to have the actual camera – so it should make it much harder to do fraudulently? And then its permanently tagged to any images taken with that camera until it’s sold/given away and the new owner registers their details.The only problem with this theory is in the rare occasions where the copyright is legitimately transferred – but this could perhaps be dealt with via the paper trail from the photographer (or old copyright owner) to the new copyright owner stating the images for which copyright had been sold.It might be possible to do this as a registration widget in photoshop/lightroom? But however, and whoever provides the software that does the registration, I really think it should require the user to have the actual camera plugged in and the serial number read and then the copyright data stored within the camera itself to reduce the risks of fraudulent data.This may well mean its only possible via the camera manufacturers, can Adobe help encourage the manufacturers to take this or a similar idea up, and provide the extra bits in the image metadata to read this info?[Gunar] Great ideas. I’ve also heard that each camera has it’s own noise signature that could also be used as a fingerprint. We are continually talking with the camera manufacturers – this would be an interesting topic.

  5. Rich Wagner says:

    Gunar,I would refer you to “A Metadata Manifesto” at http://metadatamanifesto.blogspot.com/Ownership metadata must never be removed.We need to institute standards and best practices in order to protect and preserve critical metadata. Information that identifies the copyright holder must be treated as “read only” or “write once” data, and must never removed by image distributors and users. The only exception would be changes done with the explicit consent of the copyright owner.Automated systems for creating and managing digital files need to honor and assist implementation of this principle. Most critically, these systems need to preserve ownership metadata by default and discourage removal of other metadata by warning users about the legal implications of removal. Right now, metadata is not preserved throughout the Adobe CS3 Suite and its individual apps. The recent changes to Photoshop’s Save for Web are an improvement, but there’s still a long way to go. For example, Acrobat Pro v8’s Extract TIFF command will strip all metadata when a TIFF is extracted from the supposedly robust PDF. That’s not good! The default behavior should always be to preserve metadata.Thanks for discussing this issue on your blog – as the dominant player in the imaging market, the default behavior of Adobe’s apps can go a long way towards establishing best practices for others to follow.[Gunar] Thanks for the reference. Metadata has become a business critical technology – but with anything digital it can be tampered with or removed. Adobe can do more to plug the holes and ensure that the business information can flow more freely through the media and tools – otherwise we begun to erode our relevance and trust with our customers.

  6. Rich Wagner says:

    One more related comment, stimulated by some writing of David Riecks, who provided me with the refs above (I’d forgotten about them!).I agree that the solution will not be completely technological – although technology can go a long way towards helping us out – just like the little sensors in packages that trip the alarm when someone walks out the door without deactivating the device at the cashier’s counter. In fact, a classic experiment showed that making products easily accessible to the public (not locked in a glass case) and using the sensors (called “source tagging”) increased sales.With digital images, the image data still needs to be easily accessible, to encourage use, but we could use the development of a “sensor tag” to reduce the “shoplifting” of images. It would also help those who acquire images some distance downstream to determine the source. I hope some of the bright minds at Adobe are mulling this one over…Technology; Putting the Tag on Shoplifters – NY Times, May 16, 1993http://tinyurl.com/5brucj–Rich[Gunar] Interesting article, I think we are at the beginning of understanding what ownership, sharing and stealing means for digital media from a cultural perspective. Especially when the store is always open and goods are everywhere.

  7. Jim Pogozelski says:

    Copyright/original owner should be permanent. But some “softer” fields should still be changeable (like keywords or location — what if there was a mistake in adding them?). Although even then, maybe a checkbox to lock it ALL down…Gunar There are simple UI tricks we can do in our applications to improve “permanence” like adding locks but a hacker with a text editor can undo all that.

  8. PiP says:

    If Adobe applications [and others] retained image metadata for all saving/extracting operations, this would eliminate most “Orphaned Images”. Just like UPC labels, this will not prevent deliberate malicious tampering, but is a very economical solution against “casual copying” and does not implement draconian DRM or needlessly complex encryption/authentication schemes.[Gunar] Exactly my point. DRM can be dangerous for the implementer and artist if it is too draconian – see Sony’s rootkit debacle – where was the net economic benefit there? But, in some cases it is economical, as in the need to lock down media that contain sensitive info or trade secrets – like in the pharma industry or movies in production. However, these examples are within a closed environment where there is value in limiting distribution rather than an open marketplace where the pull it towards promotion and access.