A Thousand Words

April 28, 2008 – 12:02 am

We've all heard the old adage: "a picture is worth a thousand words." I thought I'd use that as a jumping off point for today's post. If a picture is worth a thousand words, what's a thousand words worth... and where is that going to take us?

The first thing to do is figure out the weight/value of 1,000 words. For it's own estimations, Wikipedia uses an average of 6 characters per word (5 letters plus a space). So, for our purposes, our first Rough Equivalent will be 6,000 characters. We're also going to use another Rough Equivalent that will be familiar to any student who's reached a certain level in their schooling, which is its Rough Equivalent in typed, double-spaced 8.5" x 11" pages... 4.

Now the first thing to do is test whether a picture is truly worth a thousand words using an objective unit of measurement that both words and pictures can be broken down into. I'll prove that words are cheap later on in this post, so I'll avoid comparative monetary values for now and go geekier with bytes. How many bytes are required to store 6,000 characters?

If they're simple ASCII characters, then 6,000. And that's not six kilobytes. Six kilobytes is 6,144 bytes because a kilobyte is actually 1,024 bytes, not 1,000 (and this is why a "200 gigabyte" drive shows up as around 186 gigabytes, because the drive's manufacturer says 1 billion bytes is a gigabyte, but your computer says a gigabyte is a smidge under 1.074 billion bytes). But the ASCII character set is very limited. It's an eight-bit encoding system with room for only 256 characters and pretty much limits us to Latinate alphabets.

If we're going to allow for various foreign characters and symbols, that leads us to encoding standards like UTF-16 and UTF-32. These are 16-bit and 32-bit encoding systems, respectively, which try to make room for every letter and common symbol in every alphabet known to humankind. And since a byte is eight bits, a 32-bit encoding system would make each letter worth 4 bytes and give us a maximum weight of 24,000 bytes.

Now, we could add weight by putting the words into a document format that carries certain amounts of overhead, but that's dishonest, because it's not really part of the 1,000 words. It's just a container. It's like weighing yourself with your shoes on.

So at 24,000 bytes, is a picture worth a thousand words?

In the above image, on the left, you'll find 6,000 characters, generated for us by the Lorem Ipsum Generator. It's actually 897 words, but I attribute that to the fact they're in pseudo-Latin rather than English. On the right, you'll find a photo that I shrunk to just over 24,000 bytes (using the "60" level of JPEG compression in PhotoShop and messing with the dimensions to drop bytes).

As you can see, to fit 1,000 words or a Rough Equivalent thereof into the space occupied by a photo that uses the same amount of storage, the words have to be shrunken down until they're illegible. The photo is clear and readable, while the 1,000 words aren't. Now you could argue that the photo is only worth as many words as could legibly fit into the same space, or you can argue that it's worth more than 1,000 words because it can do the job in a smaller space than 1,000 words needs (which is probably the original meaning of the old adage). If you're inclined to argue that, scroll down and use the comments section.

But the really fun Rough Equivalents come when we start doing the silly math with 1,000 words being 4 sheets of 8.5x11 paper. Standard 20 lb. bond copy/printer paper in the 8.5" x 11" size weighs 20 pounds per 2000 sheets, or 100 sheets per pound. That means a good conversion metric might be 1 pound = 25,000 words (in double-spaced typed pages). It also means 1,000 words roughly equals 2.6 square feet.

Here are a few Rough Equivalents:

• If I go back to my ton of money post, a ton of pennies is worth $3,628.74. A ton of words would amount to 50,000,000 (25,000 * 2,000 lbs). That's roughly 137.78 words to a penny. Proof that words are cheap.
• An acre of words (43,560 square feet) would be roughly 16,753,846 words.
• The Space Shuttle could carry approximately 1,342,500,000 words into orbit.
• It would take a 84,615 page document, comprised of approximately 21,153,750 words to carpet all the floor space in the White House.

What are some Rough Equivalents you can come up with?