June 17th, 2003


Ideogram Soup

Jackals roam the rooftops of the Bay Area, fangs bared, eyes alert for solitary prey. They leap from roof to roof. When you look at them, reflected in the moonlight in a puddle of oil, you will see human forms.

One such jackal spoke to me.

"The Japanese language, at least Kanji, uses ideograms instead of alphabet. Each symbol represents a full word."

Neither I nor the jackal knew any Japanese, but the jackal continued.

"Today, though, there are problems with this. Using Japanese on computers means implementing a much larger character set than a normal alphabet. Different people found different ways to implement this set, and they're not necessarily compatible. When you use the wrong characters, it's called ''Mojibake.' "

The jackal paused, sniffing the air, then returned to what it was doing.

"Mojibake literally means 'haunted characters'. It's nonsense Japanese. Sometimes it just shows up as Yen signs or character with Umlauts, but sometimes, it shows up as totally different word-jumble Japanese characters. Mojibake chracters."

The jackal sat, and in the moonlight, looked noble and poised.

"I propose divination by Mojibake. The query, or a statement, is entered in one Kanji-based character set. Then the data is read as another character set. The nonsense is pieced together and meaning is placed on it. By this, the postmodern diviner can discern hidden truths."

I gurgled a bit, then said to the jackal, "You're just scavenging."