[GeorgeOrr]

I think the author of this article kind of missed the points being made by Hoffman. For instance, he asks:

"But how did the icon come to look like a snake in the first place?"

Sort of presupposes that it does look like a "real" snake. The icon doesn't necessarily have to look like anything ... we just have to evolve in a way that gets us to jump out of the way. It's the jumping out of the way that is adaptive, not the similarity to anything outside the system trying to survive.

To use Hoffman's analogy, it's like asking why do files on a desktop look like the files in the computer. They don't, they are useful interfaces however.

[nopinsight]

There are many reasons to use a representation that is "similar" in some way to the real thing. These include sharing inference mechanisms to predict characteristics like physical movement, affordance, etc among a number of entities when they are represented consistently. (e.g., snakes are represented with an elongated shape; eggs, circular.)

In theory, it could be possible to devise a system of representations that have such desired properties despite their differences from reality. But I suspect that in practice the most efficient choice would be to mimic certain essential features of reality.

[camelNotation]

You can suspect that, but Hoffman's research doesn't support it.

[jasperry]

Taking the graphical user interface idea a step further, even the best user interface can't help but "leak" some details about what's going on under the hood.

And I'd argue that given the high-level general processing ability of our brains--Turing machines can simulate anything computable--we actually have a tendency to recognize that leakage and build better models of reality, if only a small step at a time.

Most of the time, the "interface" presented to us by our senses is very workable and lets us survive without needing to know too much. But as we keep trying to integrate new pieces of data, we do come to realize that our models are too simplistic, and we have to revise our models. So we do, slowly, over the course of our own lifetimes and over the course of of civilization, dig deeper into the reality behind things. That's what I consider intellectual progress.

Think about how some people develop critical thinking capabilities even though they live in regimes that basically feed them comforting lies 24/7. On the one hand, we often want to hide from reality to stay in our comfort zone, but when reality as currently handled becomes too painful, we do exhibit the ability to learn a better model.

[JoeAltmaier]

To revisit the 'file' idea: the 'save' icon is an idealized floppy disk. But generations(!) have passed who never saw one. So now its just an icon.

This seems to happen to all written symbols. They quickly lose representative form, and become symbols only. I don't think most Chinese characters look anything like a house or a sun any more; the question is moot.

[wlesieutre]

The 'save' icon is fading away pretty quickly in common usage. Modern software is trending toward changes saved in realtime (driven by phones, also seen on OS X and web). When I write a note, I just make a new note, type in it, and close the app.

My favorite example of old-timey icons is 'phone'. https://www.google.com/search?q=phone+icon&tbm=isch

Still used everywhere, but phones just look like ▯ now.

[benjarrell]

The phone icon is from AIGA: http://www.aiga.org/symbol-signs/

[kuschku]

Well, that doesn’t work everywhere, though.

[chc]

How many generations do you reckon have passed since the mid '90s?

[JoeAltmaier]

Well, 20 years, maybe 2? The kids I know have no idea what those icons mean; the young adults I know may have some idea but have never seen the items pictured outside of an old movie.

[hsitz]

"The icon doesn't necessarily have to look like anything ... we just have to evolve in a way that gets us to jump out of the way."

Well, the icon does have to look like _something_, if we're going to see it. So seems good question of why it looks like a snake, rather than something else, whether there's something about the "real" snake that constrains how we must see it if we're to use icon to successfully adapt our behavior. E.g., an icon of a snake that was shaped like, e.g., a sphere or a box, would not be very helpful if we wanted to pick the snake up with a stick to move it out of the way.