AmazonFail = TaxonomyFail?

By now, #amazonfail seems like old news (yesterday’s detwitus?), though apparently Amazon’s PR folks are still doing damage control.

But what intrigues me was something in Clay Shirky’s nostra culpa post comparing the collective outrage against Amazon to the Tawana Brawley incident. While the post on a whole did not move me (perhaps because I don’t have any guilt to atone for), I did see a valuable nugget:

The problems they have with labeling and handling contested categories is a problem with all categorization systems since the world began. Metadata is worldview; sorting is a political act. Amazon would love to avoid those problems if they could – who needs the tsouris? — but they can’t. No one gets cataloging “right” in any perfect sense, and no algorithm returns the “correct” results. We know that, because we see it every day, in every large-scale system we use. No set of labels or algorithms solves anything once and for all; any working system for showing data to the user is a bag of optimizations and tradeoffs that are a lot worse than some Platonic ideal, but a lot better than nothing.

Indeed, perhaps the problem is that Amazon relies too mu…

By now, #amazonfail seems like old news (yesterday’s detwitus?), though apparently Amazon’s PR folks are still doing damage control.

But what intrigues me was something in Clay Shirky’s nostra culpa post comparing the collective outrage against Amazon to the Tawana Brawley incident. While the post on a whole did not move me (perhaps because I don’t have any guilt to atone for), I did see a valuable nugget:

The problems they have with labeling and handling contested categories is a problem with all categorization systems since the world began. Metadata is worldview; sorting is a political act. Amazon would love to avoid those problems if they could – who needs the tsouris? — but they can’t. No one gets cataloging “right” in any perfect sense, and no algorithm returns the “correct” results. We know that, because we see it every day, in every large-scale system we use. No set of labels or algorithms solves anything once and for all; any working system for showing data to the user is a bag of optimizations and tradeoffs that are a lot worse than some Platonic ideal, but a lot better than nothing.

Indeed, perhaps the problem is that Amazon relies too much on algorithmic cleverness when it should be taking a more transparent HCIR approach. Perhaps not what Shirky was after, but it’s consistent with all of the versions I’ve heard of what went wrong.

Link to original post