Translate

Monday, April 28, 2014

In data identification?

Had an idea yesterday about sites adding a user's identifying identification onto a photo that is sent back to that user, and began to wonder this morning if it's a new idea--which I doubt--and also to consider how it can be done with just about any data.

The reality when you use the web is that your computing device sends a request for data to some server somewhere. That server if it honors that request sends you the requested information back. It conceivably could add information identifying you into that information, if it had it, within the data it sends back to you!

Like with my original photo idea, your user ID if you were known to the site could be imprinted on the photo, either visibly, or near visible, or invisible as a watermark. That would only matter if you shared as it could let people trace back sharing to the source, if it were copyrighted material.

One cool thing is that there are no privacy issues, as it's the user's own identification being sent back to them. If data isn't being encrypted, then that probably isn't a problem as if they requested the data via an unencrypted stream then they already sent that exact same type data. So you're just reflecting back, removing privacy concerns. If they share later then they broke their own privacy, not you.

Of course one might argue that people could simply remove the data before sharing.

Yup, if they knew how!

If everybody did it differently, then it would be difficult for some app to be built that could remove every way it might be done, and for most users, sites could have a reasonable expectation that they could trace back illegal sharing rather accurately, with a narrow group of superusers at the fringe with whom there would be a constant war to figure out ways to imprint they couldn't find.

But that's just what I'm thinking this morning after an idea yesterday which occurred to me while I started thinking about the subject wondering how you might keep up with who is sharing your photos.

I have these kinds of ideas ALL the time. I don't necessarily take them too seriously, but will just toss them up, as kind of a brainstorming exercise.

If not new, ok. If is new, why? It's so simple!

Idea is free.

Hmmm...have decided to call it: in-Data Dynamic Identification. Or IDDI for short.


James Harris

No comments: