Code Page

Imagine old-time spies each holding a different codebook. Give them the same coded message, and they'd read it as completely different words. A computer showing text on screen works much the same way. A computer doesn't know letters at all. It's a machine that only calculates with numbers made of 0s and 1s.

So developers created a basic rule called ASCII, which gave each letter and symbol its own number. But ASCII could hold only 128 characters, such as English letters and digits. That was nowhere near enough for other languages. In the end, each country started filling the leftover 128 slots with its own language.

The United States and Western Europe filled them with special symbols and accented letters. Places like Korea and Japan have thousands of characters, so 128 slots weren't enough. They used pairs of numbers from the leftover slots to stand for a single character. A standard that pairs numbers with characters for a particular language or region is called a code page. In effect, computers in each country opened a different codebook to read the same text.

You may have opened an old file and seen a flood of strange symbols and bizarre characters. The computer or the file isn't broken. The computer that wrote the document and the computer that opened it were using different code pages.

For example, Korean Windows uses a code page called CP949 to turn Korean characters into numbers and save them. But if you open that document on Japanese Windows, the Japanese code page (CP932) kicks in. It turns the very same numbers into completely different Japanese characters and shows those on screen.

When the writer's number table and the reader's number table don't match, you get garbled text. Every time people swapped files between language regions, alien writing popped up and caused headaches.

To end the mess of every country using a different number table, Unicode, which holds every character in the world in one giant table, arrived along with UTF-8, which handles it. Thanks to them, we can now visit foreign websites in other languages and read them without garbled text.

But a new technology doesn't make the old problems vanish overnight. The only thing the widely used UTF-8 can read as-is from the old days is English letters and symbols in the ASCII range. Old text documents saved with a Korean code page, such as the older precomposed Korean encoding (wanseonghyeong) or CP949, still look garbled if you read them as UTF-8 without changing the encoding.

A code page is less a single method than a technical product of an era, when computer memory was far too scarce and people somehow had to represent each region's language. Even now, the rules of code pages are still alive deep inside Windows and in old programs at public agencies.