--On Tuesday, June 23, 2015 03:35 +0000 John Levine
<johnl(_at_)taugh(_dot_)com> wrote:
If the EU debt met all of the criteria identified in 3797, I
would see no reason to avoid using it. I don't know whether
there is such a thing as "the EU debt"
There is no such thing as "the EU debt." Each EU member
country has its own national debt. Some of that debt is
denominated in Euros, some in national currencies such as UK
pounds, Danish Krone, Romanian Lei, and so forth. The Danish
Krone and Bulgarian Lev are tightly tied to the Euro, while
...
Thanks. That was approximately what I had assumed from assorted
news reports and a vague understanding of the system, but I
didn't know offhand and didn't have the time to check it out.
The US debt happens to be unusually easy to use as a data
source since it is a single country's debt in its own currency
that no other country of any size* uses, and the US publishes
very good economic data. Given its advantages, and the points
you made about the importance of larger amounts of data rather
than perfect randomness of individual sources, I'd want
something more persuasive than loud unsupported assertions to
switch to something else.
As Joel pointed out, the specs leave the choices up to the
Nomcom Chair each year as long as specified criteria are met.
There is really no basis, other than expressions of personal
opinion (and political preferences) for criticizing the choices
unless those criteria have not been adhered to. We have not
heard a single claim that the criteria were violated. Harald
explained the reason for his choices (which he was certainly not
obligated to do), so I suggest we move on.
One suggestion for those who don't like the choices (for
whatever reason): If someone feels passionately enough about
this to want to do some work rather than complaining, I think
doing the research and posting a list of a hundred or two
plausible choices would be useful. The list should not be
normative or otherwise restrict Nomcom chair choices, but I
infer from Harald's comment that one of the reasons for making
particular choices is precisely that it was easy and the
alternatives would require significant research work, so making
a long list of possible options, with URLs and whatever
historical, statistical, or other explanations are needed might,
over time, improve the diversity of the selections made just by
making more options readily available. I do believe, however,
that any such document should either make clear that
diversifying the choices is an aesthetic or political statement
rather than a way to significantly improve either randomness or
lack of predictability of the final results unless the latter
can be demonstrated on a sound statistical basis.
john
R's,
John
* - yeah, I know about Panama, Ecuador, the Bahamas, and the
BVI