Talk:376: Bug
That is why on Unix epoch (the time_t type) is signed type, and covers dates before epoch. --JakubNarebski (talk) 19:52, 5 July 2013 (UTC)
- Ohh, and much more is missing. I did mark it as incomplete. We also have to talk about the time frame the 32bit epoch does cover, and what would be changed by using a 64bit variable. What will happen on 19 January 2038?--Dgbrt (talk) 20:17, 5 July 2013 (UTC)
- The general hope, it appears, is that 64-bit integers will be firmly in place, having ousted the feeble 32-bit integers from the system time. As has been demonstrated in innumerable instances, it's rather difficult to eliminate legacy code from systems due to attempts to support older systems in a backward-compatible methodology. In short, however, it will take time to resolve time. Thokling (talk) 05:34, 22 September 2013 (UTC)
- Rewrite needed
The current explanation barely has to do with the actual topic of the comic. Instead it explains several unrelated qualities of Unix time, and petty much skips over the actual epoch thing. Needs a rewrite. --NeatNit 141.101.99.58 05:59, 8 February 2015 (UTC)
Why would anyone want to take the square root of a timestamp? It is much more likely that Cueballs program just handles negative time values incorrectly. Condor70 (talk) 07:15, 14 November 2016 (UTC)