Geek twitch - Grin with cat attached — LiveJournal
Previous Entry Next Entry
Geek twitch Oct. 13th, 2008 05:11 pm
"The Long Now Foundation was established in 01996".

NOO! You can't have 9's in octal numbers! Bad numerologists!

(Very cool project, yes, but you don't get long-term thinking by trashing data-representation standards)

From: barbedwirekiss
Date: October 13th, 2008 - 06:36 pm (Link)
Errmm... from http://www.longnow.org/about/

"The Long Now Foundation uses five digit dates, the extra zero is to solve the deca-millennium bug which will come into effect in about 8,000 years."

So, they're not attempting to use octal-based numbering just trying to avoid a Y10K bug! :)
From: wechsler
Date: October 13th, 2008 - 06:38 pm (Link)
Yeah, I saw the excuse, it still makes my geek-sense twitch! ;)
From: mooism
Date: October 13th, 2008 - 10:12 pm (Link)
But they remain oblivious to the Y100K bug.

Alternatively, we already solved it by having numbers (including year numbers) terminate with the first non-digit. They’re just being pretentious. Leading zeroes are only needed in abbreviations.
From: kekhmet
Date: October 13th, 2008 - 10:49 pm (Link)
yes, I was just thinking about this the other day. I had just commented to someone that, having worked on trying to prevent the Y2K bug rearing its head in a complex and aging billing system at the utility company I worked at in the years just before 2000, I cannot bring myself to write years as abbreviated to just the last 2 digits. This made me wonder, briefly, if I ought to start left-padding my dates with 0's. I concluded no, since the leading 0's are implicit. The actual thing needed to prevent a y10K (or Y100K, or...) bug is having *space* for greater than 4 digit years. The y2K bug wasn't just - or indeed even chiefly - about lazy people only writing the last 2 digits of a year that had leading non-zero digits in it after all. It was about all the code running at the time that was still storing dates in such a way that there was only room for two digits in the year! Admittedly, it does seem like a better bet that we really truly will not still be running old code from now in the year 9999 than it turned out to be for the folk who wrote all that code for mainframes back in the 1970's never expecting that anyone would not have moved on from it by 1999. But you never know (and one hopes people in oh, say, the 9990's will have the sense not to write *new* code that can't handle 10000, but plenty of fools didn't think ahead to 2000 even in the early 1990's ;-/ people have a proven track record for being a bit idiotic ;-P )