Wednesday, October 19, 2005

Data compression technique and length compression (The Story of Locapoint: 3)

The method to shorten display length using large base of radix (more than ten) is called digit length compression. I wrote about it in my previous post.

These is a similar word, "data compression". (Concept is different, though).
I consider about losslessly data compression techniques large file could be compressed a lot. In fact, there are lots of compression techniques in IT world. zip, cab, or tgz are a very popular extension for compressed file.
However, latitude and longitude for a point with 1 meter precision is only 50 bit. In byte, it only 7 bytes. Latitude and longitude for any place on earth means, data could be random value. Then 50 bit is minimum requirement. Data compression technology can't be used for single location data.

There is another way to 'fake' compression. Like G-Code on video recording, assign small number for frequently used object. For example, major channel, high viewing rate time, fixed record time (like one hour, two hours) will be assigned small number. When G-code is "23", it originally comes from "0000000000023" etc. (I don't actually know about G-code length). If you want record a minor TV-program, its G-Code should be long and close to original length.
Similarly, Major street in Paris is assigned a short digit, so address will be short and simple.

With this concept, Location code could assign short and convenient code to major cities, and assign longest code for ocean area.
"MAPCODE" designed by Denso Co. in Japan use this concept. Assign short number to Tokyo area, so any location in Tokyo can be expressed in 4 digits of numbers. But in country side, it needs 10 digits.

This method needs "assign" information between area and codes to encode/decode. In fact, if you want to use MAPCODE, you need to buy data from Denso Co.


I have written about demerit of latitude/longitude format. "too long".
However, what is the Merit of traditional latitude/longitude format ?

When someone is in distress in sea or mountain, Lat/Lon format is generally used for SAR(Search and rescue) operation. If you send a mayday with "MAPCODE" and SAR entity didn't buy a assigned data?
The biggest merit of Lat/Lon format is its generality.


I tried to create a code for "easy usage" but now I have to define what is "easy usage".


Next, I will write about my definition of "easy usage" for geographic location code.

Continue...