> urticator.net

  About This Site
> Glue

  The Mind
  The Body
> Other (2)

  Don't Fight Your Mind
  Thoughts About Email
> How I Cleaned My Room
  Household Memes

> Separation of Functions
  Multiple Passes
  De-Sentimentalization (1)
  Disposing of Things
  De-Sentimentalization (2)
  Other Approaches

  Classification of Functions
> Footnote


A physical object can't be in more than one place at the same time, but an information object can.

To make any sense of that claim from the parent essay, you need to be familiar with my peculiar ideas about information objects, also known as persistent objects. If you want the long version, here are the two definitive essays, each several pages long.

Concepts for Persistent Objects
Objects and Identity

If you want the short version, well, I think I can explain the relevant point in just a few sentences. Suppose I have a box of identical cheap blue pens. I could say that they're all instances of a single object, the Platonic cheap blue pen of that particular design, but it would be weird, and nobody would take me seriously. But, suppose I have a file on a computer and I copy it to a flash drive, a laptop, and a web server. In that situation it's perfectly natural to talk about the file and regard all the copies as instances of a single object. That's how an information object can be in more than one place at the same time.

I didn't make the right distinction there, however. The point isn't that there are different kinds of objects, it's that there are different frameworks, different ways of thinking about objects. An information object can always be thought of as some number of physical objects, and a physical object can always be thought of as part of some information object. Which view is more compelling depends on the situation.

Thinking of information objects as physical objects is usually fairly dull, but the reverse can be fun which is why I already wrote a whole essay about it (Objects and Identity). Here are two more examples that I thought of just now.

  • Faced with a locked door, it's perfectly natural to ask who has the key, even though there are almost certainly several physical keys.
  • When a book is published, it's natural to talk about the new book by whoever, even though that one book is instantiated in perhaps millions of physical objects.

The point is, files on computers aren't the only things that we view by default as information objects.

(When I wrote Objects and Identity, I thought that the physical instances of an information object ought to be physically identical, but at the moment that seems like a mistake. It doesn't matter if a key is chrome-plated or not, if a book is hardcover or paperback, or if a file is stored as magnetic spots on a disk or as electrical charges in a chip, what matters is that the pattern is the same.)

Now, where were we? Objects have functions, function implies location, so an object with many functions needs to be in many places at the same time, which is problematic unless the object is an information object! If you have a file that's a reference manual, and you need it in more than one place, you can just make sure there's an instance in every place, and you're all set. (And what if you print it out? And make photocopies?)

There are two interesting complications I'd like to mention.

First, if an information object is in many places, and is changing over time, then it takes effort to keep the instances in sync. That effort may or may not be made. In the reference manual example, suppose you post the manual on the internet. If you release a minor update, most people won't notice, and those who do may not care. (And, those who care might also keep a copy of the old version for reference!)

I also like to think about how the changes occur. They might flow outward from an authoritative central source, as in the reference manual example. (Virus definitions are another good example here.) They might originate in several places and be routed through a central location, as in a version control system with several users. Or, they might be totally chaotic, with the object kept in sync only by occasional peer-to-peer communication.

The second complication is that it's not really true that function implies location. The pen and paper I use for my grocery list aren't in the same location as the phone, but they're only a few feet away, and that's close enough for taking notes when I'm talking. So, what matters is location and distance. (In fact, the whole idea of being in the same location as the phone is fundamentally wrong. The phone occupies a volume, not a single point, and it occupies it, so that nothing else can be in the same volume.)

However, distance doesn't have the same meaning for information objects as it does for physical objects. An information object can be broken up into pieces and sent at the speed of light over optical fiber, and at that speed, everything on Earth is close to everything else! In other words, what matters isn't distance in space but rather distance in time or effort. This isn't just some abstract theory, by the way. More than once, I've wanted to check a fact, known the exact book and page where I could find it, and still found it quicker and easier to get the fact off the internet than to get up and walk over to my bookshelf.


  See Also

  Separation of Functions

o May (2007)
@ December (2007)