by Michael S. Kaplan, published on 2006/02/24 00:45 -05:00, original URI: http://blogs.msdn.com/b/michkap/archive/2006/02/23/538455.aspx
Earlier today in the post On an upgrade, we maintain, I talked about how we have a very strong desire to retain a user's locale preferences on upgrade.
It may just be that extra three inches from the ground that would make people feel a little uncomfortable about (see the post for more on that reference!), but we really want to keep the results as consistent as possible.
But at the close of the post I said there was an exception to this rule.
That exception?
People made many guesses and one even came close, but no one guessed it.
So I thought we could try and reason it out. And I will use The Poor Man Institute's approach to using SCIENCE to do so, described as follows:
SCIENCE is nothing to be afraid of - it is merely a method of inquiry which makes use of empirical data about the world and fits it into an abstract, predictive model. For example, suppose you ask me the question: “what is the volume of an average human being?” This is a very stupid and pointless question, exactly the sort of question I would expect someone like you would ask. Why do you care? If I refuse to answer your question, you may become violent, so I will attempt to do so, quickly, by making a few simplifying approximations. First, in order to make the math simpler, I will assume that the average person is a uniform sphere, 3 feet in diameter. Why, when I look at the problem that way, it turns out that I’m really quite extraordinarily tall and svelte! Indeed, I’m far too attractive a physical specimen to have to answer your damn fool questions, so I roll you out the door like a beachball full of cottage cheese and have the chicks from “Coyote Ugly” over for a week-long orgy. All thanks to SCIENCE!
So, by using these powerful analytic tools, let's see what we can find out....
Think for a moment about what are the types of problems that have really galvanized people's willingness to ignore their usual tendencies. I can think of three:
1) SECURITY: Obviously if there were some sort of security concern, there might be a reason to make a specific targeted change. There were, however, none of these related to locale settings. So that would not be it.
2) TERRORISM: Obviously the terrible events of 9/11 have been the source of many changes in the way people live their lives. However, the implied timeline of change would really predate the exception. So that would not be it, either.
3) Y2K: The average person probably thinks of this whole thing as the biggest wet firecracker in history, though even from my relatively uninvolved position I was indirectly to tests which, had they not been run in order to find problems, would have resulted in disastrous situations. Just as Jonah was hated for hie false prophecy (even though it was his warning that caused the people to clean up their act and thus be able to defeat the prophecy!), many consultants were not given the thanks they deserved.
(Of course many other bilked their clients and used scare tactics to make heavy profits, so it all evens out, I guess!)
In any case, it was indeed the Y2K issue that led to the exception being made. The short date format was automatically updated to use a four digit year on upgrade (using the special /U switch to intl.cpl's unattend format), as follows:
rundll32.exe shell32,Control_RunDLL intl.cpl,,/U
The general feeling of unease simply seemed to make such a change feel worthwhile as the year 2000 loomed near and a version of Windows was about to ship....
So there you have it -- the exception that proces the rule, worked out with the help of SCIENCE! :-)
This post brought to you by "Ｕ" (U+ff35, FULLWIDTH LATIN CAPITAL LETTER U)
referenced by