Hacker News Reader

If it's okay to screw programs that make nonportable assumptions, then int should just be 64 bits. After all, int is supposed to be the machine's natural integer.

The current fashion is that we like programs that assume int is 32 bits, and don't like ones that assume wraparound arithmetic.

View on HN

32-bit ints have a more compact encoding in x86 because it was assumed they would be the most common. The "native" size isn't a simple matter these days. Back in the old dark days of 8 and 16-bits common values regularly exceeded the size. But 4 billion is large enough to still be the default.

Reply on news.ycombinator.com