Last week, I went to an art show locally that contained many types of artwork, including several pieces of glass art. I used to work for Corning Incorporated (formerly Corning Glass Works) and learned about glass art from people who really knew their stuff, so I suspect that my view on one piece was very different than many people.
The piece was a blown glass water pitcher and carried a $100 price. A friend (who is also an artist) wanted to talk about it. The first thing that struck me was that there was an obvious bubble that did not belong (a flaw). A flaw would decrease the price. She, however, had picked it up (something I considered a no-no) and noticed the weight of the piece. It was heavy. I know that those who are relatively new at blowing glass have a hard time creating pieces that are delicate. This piece looked delicate, but it wasn't. My suspicious is that the price related not only the flaw but also its weight (which we might also consider a flaw).
If you had seen a photo of the glass water pitcher, you would have liked it. And I imagine that very descriptive metadata could be created for it. Even though the metadata could include info on the flaw and on the piece's weight, holding it made the story real. So here is another example of where a digital surrogate would not do the original justice.