Floating point to decimal conversion is not so easy

Tagged:  •    •    •    •  

Russ Cox of Plan-9 and Go fame posted a blog entry titled Floating Point to Decimal Conversion is Easy. While he is usually right, I believe this time he isn't.

Floating point to decimal conversion is easy if you are okay with ugly results. A good conversion routine will print the shortest decimal representation of the floating-point number, that is, the shortest decimal number whose closest floating-point representation equals the original number. You do not want 0.30000000001, you want 0.3, because the number right above 0.3 is 0.30000000003 and 0.30000000001 does not provide any extra precision.

Syndicate content

User login