OMG Teh Geekiness

[andrewducker]

Comments

[khbrown.livejournal.com]

Architecture / word length dependent, amongst other things?

[andrewducker]

Hence the alt text...

[khbrown.livejournal.com]

Didn't view the ALT text, though maybe the adage that a joke which requires explanation (= footnotes / alt / longdesc) is not a joke applies?

[andrewducker]

I thought it was funny before reading the alt text.

And all xkcd's have Alt Text. It's part of the way it works.

[skington.livejournal.com]

A friend at work found xkcd, and started on an archive binge. She was at something like #80 when we told her there were jokes in the alt text. She had to start again.

[dougs.livejournal.com]

Me, I always go from 127 to -128.

[call-waiting.livejournal.com]

If you're implementing it in C, the standard says that signed overflow is 'implementation defined', so if you're incrementing a 16-bit type, it depends on what your compiler does. And what the compiler does is probably largely dependent on what the hardware does, which may or may not be the obvious truncating overflow...

[khbrown.livejournal.com]

Or the standard isn't really a standard, being implementation defined?

[call-waiting.livejournal.com]

Well, it at least defines which things are defined, and which things aren't defined. The standard creates, as Vroomfondel would say, rigidly defined areas of doubt and uncertainty. Which is a lot better than not defining anything at all, and in fact better than defining absolutely everything because it allows the compiler some flexibility in making better use of the hardware.

[khbrown.livejournal.com]

Unless 'everything' contained a sub-clause of leaving certain defined things up to the compiler to define / decide ;-) ?