[ExI] y2k bug

[Harvey Newstrom]

On Saturday, July 12, 2014 2:09 PM, Anders Sandberg wrote,
> Are there any good post-mortem whitepapers that analyzed the total
> impact, the outages, and especially the what-if outages that would have
> happened without the effort? Y2K seems to have been both a potential risk
> that was largely averted, and a risk people decided to forget about as soon as
> the hangovers lifted. There is probably plenty to be learned from this.

The United States Air Force Experience with Y2K:
http://www.nap.edu/openbook.php?record_id=11999&page=1
"clear that enough problems were experienced in the course of the Y2K rollover to demonstrate the reality of the problem and the importance of remediation efforts (GAO 2000). Serious known disruptions were avoided in the banking and insurance sectors, two NATO nation spy satellites went down for two days, and numerous other documented failures were either avoided or responded to in real time during rollover."

Gartner research:
http://ecommerce.hostip.info/pages/1114/Y2K-Bug-IN-HINDSIGHT.html
"Gartner Group, for instance, held that the catastrophe-free rollover was a product of the extensive and successful preparation companies undertook; Gartner Group analyst Dale Vecchio remarked, "people didn't spend $300 billion on a problem that didn't exist."

Wikipedia lists some documented problems:
http://en.wikipedia.org/wiki/Year_2000_problem#Documented_errors
- Incorrect medical test results sent to pregnant women, causing two mistaken abortions.
- Four Down's syndrome babies were born who were mis-screened.
- Radiation monitoring equipment failed at Ishikawa, Japan.
- Onagawa, Japan nuclear power plant sounded false alarms.
- Japan's largest cellular carrier had phones deleting new messages rather than saving them.
- Australian bus-ticket machines stopped working in two states.
- One location of slot machines in the U.S. stopped working.
- U.S. Naval Observatory official government time clock reported invalid year of 19,100 disrupting all government clocks referencing this source for official government time.

This website documents many Y2K problems that actually occurred before the date was reached and were fixed by the time Y2K arrived:
http://www.greenspun.com/bboard/q-and-a-fetch-msg.tcl?msg_id=000RSP

One common misconception is that programs would merely miscalculate dates after 2000 as being dates in the 1900s.  But in reality, many of the flaws produced invalid dates, such as January 0, or the year 100 or 19,100.  In such cases, the programs didn't merely produce the wrong date, but instead, they aborted or had math errors and just plain would not run.

Such common programs included Microsoft Excel spreadsheets, the C programming language date libraries, Perl scripts and Java programming language, both used extensively on the web.  Most credit cards had to be reissued before Y2K, and most ATMs and software accepting cards had to be upgraded.  

Many people think there was not a problem because they did not have to do anything to fix the problem.  Many do not realize that their PCs and applications were often triggered to update themselves with patches to solve this problem.  All Windows (3.1 and up), Macintosh, most Linux's, most Cisco network devices were all patched long before Y2K, and would have failed if they were not.

So imagine if these were not fixed and most spreadsheets, credit cards, ATMs, gas-pumps, web pages, home computers, and networking equipment all quit working on the same day.  It could have been bad if they were not fixed in advance.

--
Harvey Newstrom   www.HarveyNewstrom.com