<?xml version="1.0" encoding="UTF-8"?>
 <rss version="2.0" 
    xmlns:content="http://purl.org/rss/1.0/modules/content/"
    xmlns:wfw="http://wellformedweb.org/CommentAPI/"
	xmlns:dc="http://purl.org/dc/elements/1.1/">
<channel>
	<title>I can't believe it's not content!</title>
	<link>http://fluff.info/blog</link>
	<description>Yet another blog, with a focus on economics and political science.</description>
	<language>en</language>

<item>
        <title>Metro signage</title>
	<content:encoded>
See, it's not just computer geeks who come up with pedantic and anti-intuitive customs. This entry is not because I've been screwing up a lot. I'm actually pretty good about this stuff, though it takes a few seconds of effort, calling up my years of experience, every single time I board a train.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000258.htm</guid>
        <link>http://fluff.info/blog/arch/00000258.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>2 February 2010 00:00:00 +0000</pubDate>
</item>


<item>
        <title>Republican census</title>
	&lt;p&gt;
	<content:encoded>&lt;a
	href="http://www.politico.com/blogs/bensmith/0110/GOP_census_in_a_census_year.html"&gt;A
	blogger&lt;/a&gt; points out that the Republican Party is sending
	out a fundraising letter with the header 2010 CONGRESSIONAL DISTRICT CENSUS commissioned by the Republican Party.
	The cover letter uses the word "census" 13 times, by my count.
	It's just another fundraising letter, being mailed to the usual "select few" targeted for this round of fundraising campaign.
	&lt;/p&gt;
	&lt;p&gt;
	Are they trying to unfairly mislead? On the one hand, the word "Census" is just a dictionary word, not trademarked by anybody, that anybody is free to use; on the other hand, there is certainly a strong association between the word and the official, mandatory headcount run by the U.S. Government. To break the tie, some definitions:
	&lt;/p&gt;
	&lt;p&gt;

Survey: a sampling of the total population, maybe checking .01% or 10% but
definitely nowhere near 100% of the population. The usual statistical techniques are run after the survey to
determine characteristics of the whole.
	&lt;/p&gt;
	&lt;p&gt;

Census: a counting of every single head in the population, shooting for 100% minus unavoidable nonresponse. No
post-canvassing statistical magic needed, because pains were taken to cover the entire population.
	&lt;/p&gt;
	&lt;p&gt;

	The U.S. Census is appropriately named because it is an effort to count every head. The Supreme Court has ruled that it must be a census, not a survey, though some statistical massaging to fix some specific types of nonresponse is OK. The Republican Party mailing is not a census, because it is mailed to a "select few" recipients, and has no goal or intent of counting every head.
As the letter explains to the recipient, "Your opinions will represent literally &lt;u&gt;thousands&lt;/u&gt; of or Republicans in your Congressional District..."
	&lt;/p&gt;
	Final verdict: the mailing is misleading, because not only does it make repeated use of the potentially misleading word "Census", but it does so in regards to something which is in no way a census.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000257.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>22 January 2010 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Be your own chocolatier</title>
	<content:encoded>
I'm at a loss for the policy implications of this entry.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000257.htm</guid>
        <link>http://fluff.info/blog/arch/00000257.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>11 January 2010 00:00:00 +0000</pubDate>
</item>

<item>
        <title>It's getting chilly, so I've broken out the tealights.</title>
	<content:encoded>
It took millennia for we, as animals and humans, to work out fire, and
Prometheus has suffered ever since for revealing to us such magic. It
made civilization possible, it allowed us to live in places where we
have no business living, it has destroyed entire cities, and it is the
symbol of all things mysterious, ever-changing, and powerful.&lt;p&gt;

You can get a hundred tealights from the drug store for $4.99. Each is
identical in size (1.5 inches diameter) and behavior (ASTM standard F2417).
Encased in a standardized tealight holder, they are our little, abject victory over everything that
is full of wonder because it is more powerful than we are.&lt;p&gt;

I have lousy circulation, so I use them to warm my hands.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000256.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>27 November 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>It's getting chilly, so I've broken out the tealights.</title>
	<content:encoded>
It took millennia for we, as animals and humans, to work out fire, and
Prometheus has suffered ever since for revealing to us such magic. It
made civilization possible, it allowed us to live in places where we
have no business living, it has destroyed entire cities, and it is the
symbol of all things mysterious, ever-changing, and powerful.&lt;p&gt;

You can get a hundred tealights from the drug store for $4.99. Each is
identical in size (1.5 inches diameter) and behavior (ASTM standard F2417).
Encased in a standardized tealight holder, they are our little, abject victory over everything that
is full of wonder because it is more powerful than we are.&lt;p&gt;

I have lousy circulation, so I use them to warm my hands.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000256.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>27 November 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Picking a cancer</title>
	<content:encoded>
Looking forward to comments about how I oppose breast cancer research and/or hate women, because Balkanization encourages those sorts of arguments.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000256.htm</guid>
        <link>http://fluff.info/blog/arch/00000256.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>14 November 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Suspicious Closures</title>
	<content:encoded>
Feel free to add to the list in the comments.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000255.htm</guid>
        <link>http://fluff.info/blog/arch/00000255.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>24 October 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Enforcing normalcy</title>
	<content:encoded>
I tried writing a piece about gardening, but this is what turned up. Maybe next time.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000254.htm</guid>
        <link>http://fluff.info/blog/arch/00000254.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>4 October 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Amtrak policy on folding bikes</title>
	<content:encoded>
A policy piece in which I mostly argue by example.&lt;p&gt;

Sorry about the hiatus, by the way. I've been kinda down and out lately.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000253.htm</guid>
        <link>http://fluff.info/blog/arch/00000253.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>26 September 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>The `Net's most conservative site</title>
	<content:encoded>
Rest assured that I do get it, and have garnered amusement from pretty much every example I present. 
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000252.htm</guid>
        <link>http://fluff.info/blog/arch/00000252.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>8 July 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>In the land of invented languages</title>
	<content:encoded>
Not really a book review; more just riffing on the author's lead.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000251.htm</guid>
        <link>http://fluff.info/blog/arch/00000251.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>22 June 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>My headphones: the social and political implications</title>
	<content:encoded>
This is post number 250. &lt;tt&gt;wc&lt;/tt&gt; tells me I've written about 250,000 words since September 2003. What have I done with my life.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000250.htm</guid>
        <link>http://fluff.info/blog/arch/00000250.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>15 April 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>My headphones: saving my life even more effectively</title>
	<content:encoded>
Because reading while biking remains a practical impossibility, I have overthought everything that I deal with while biking.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000249.htm</guid>
        <link>http://fluff.info/blog/arch/00000249.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>4 April 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Better of SXSW `09</title>
	<content:encoded>
There were a dozen more worthy tracks after my faves from last time, so I put up another file, at  &lt;a href="http://drop.io/fluffinfo2"&gt;drop.io/fluffinfo2&lt;/a&gt;. At some point I might let the nonsubscribers know too.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000247.htm</guid>
        <link>http://drop.io/fluffinfo2</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>3 April 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>The password arms race</title>
	<content:encoded>
Little sociomathematical problems turn up everywhere. Sometimes life is fun like that.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000248.htm</guid>
        <link>http://fluff.info/blog/arch/00000248.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>26 March 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Best of SXSW `09</title>
	<content:encoded>
Don't bother reading these notes. Just go straight to the &lt;a href="http://drop.io/fluffinfo"&gt;music download&lt;/a&gt;.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000247.htm</guid>
        <link>http://fluff.info/blog/arch/00000247.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>18 March 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Updates</title>
	<content:encoded>
Hi. I haven't been updating in a while because what I intend to be the
next post---the annual round up of the SXSW torrent of pop---takes a
ton of research. I'm most of the way through, so keep an eye out. Subsequent posts should come a little more frequently.
&lt;p&gt;
In the mean time, those of you who read this blog for the mathy philosophy of science stuff might want to 
have a look at this
&lt;a href="http://modelingwithdata.org/"&gt;blog on computational statistics&lt;/a&gt;.
It's by Ben Klemens, with whom I share a close relationship. E.g., we as a unit hacked together the code underlying fluff.info's blog, now recycled for the Modeling with Data blog. [Its core back-end is a modded version of very useful and sadly
&lt;a href="http://compliticytheory.vox.com/library/post/noah-grey-is-leaving-the-internet.html"&gt;disappeared&lt;/a&gt; Greymatter CMS].
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/20000246.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>10 March 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Cut-to-keep ratio</title>
	<content:encoded>
In the &lt;a href="http://fluff.info/terrible/"&gt;Why Word is a terrible
program&lt;/a&gt; section of this site, the author mentions the wonderful
joy of having a document-writing system that lets you leave comments
that don't appear the final document. Besides the general utility of
comments, this means that you can `comment out' drafts that you worked
on but which might not quite be ready.
&lt;p&gt;
I have read many a review of a B-side compilation that used a phrase like
`[My favorite band]'s cutouts are better than most bands' best works'. I
like to think that that sort of thing more-or-less applies to my 
writing. This may sound cocky, but that attitude helps to
keep my writing from overflowing. If there's a borderline
paragraph, I reassure myself that even if my B-sides are still be pretty 
danceable, they still have to go. So complete crap
just gets deleted; B-sides get commented out in case of later regret.
&lt;p&gt;
I wrote a little ruby script to elide the commented segments from 
my documents, and used the &lt;tt&gt;wc&lt;/tt&gt; utility to compare word counts with and
without. It turns out that 20% of what I wrote for the textbook didn't
get published. That is, for every four paragraphs of final product, I
have a paragraph that was basically OK but not quite good enough to use. Of
course, the true ratio of cut-to-keep is larger because of stuff that was so
bad it wasn't even worth commenting out, but I think this ratio of
B-sides to final product is a better measure of how well I'm self-editing (and is, um, measurable).
You get to decide for yourself if this is high or low, and compare to your own cut-to-keep ratio.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000246.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>5 February 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>The journalist and the horse race</title>
	<content:encoded>
Perhaps I should've posted this one before the election--I wrote it about a year ago--but the situation isn't changing any time soon.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000246.htm</guid>
        <link>http://fluff.info/blog/arch/00000246.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>2 February 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>My office</title>
	<content:encoded>
You may be wondering where I do all my work, being that I'm unemployed.
So here's a low-quality photo from my `office', where I am writing you 
(and blatantly avoiding work) now. As you
can see, I'm big on windows, so I've got the desk pulled right up to
them. I'm told that this is bad chi, so I'm living dangerously. But I do get
to watch the birds (which my telephone reduced to a few blobs in the
tree at left), watch the snow, and delight in not having to leave the
house.
&lt;p&gt; &lt;img src="http://fluff.info/blog/asst/office.jpg"&gt;&lt;p&gt;
&lt;p&gt;
The monitor is mounted on an IV pole, which allows me to mount the thing
sideways, so I can see full pages from PDFs and long streams of code.
Writing that last book would've been heck without a paper-shaped monitor.
Peeking out from behind the monitor is the antenna from the wireless router that is hanging from
the same IV pole. Having it right against the window makes it easier to share wireless with the neighborhood.
&lt;p&gt;
Where's the keyboard? I
&lt;a href=http://fluff.info/blog/arch/00000082.htm&gt;posted on keyboard setup&lt;/a&gt; a long time ago.
&lt;p&gt;
Next time: content. But for now, seeing that it is a soft January afternoon with snow curled about the house, it's back to coffee and bird-watching.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/20000345.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>27 January 2009 00:00:00 +0000</pubDate>
</item>


<item>
        <title>Rubber foot protection</title>
	<content:encoded>
	I sent in my laptop for warranty service, with of a list of ailments
	gathered over years of heavy use of a device that is in the end just a
	piece of plastic with a hard drive. Broken this, wobbly that, port
	stopped working.
	&lt;p&gt;
	Best Buy got it back a month late, but in much better shape.
	They seem to have disregarded my list of ailments and run down their own
	checklist, because they did things that I hadn't asked for, like replacing
	the keyboard, and didn't do one or two things
	I had listed, like replacing the laptop's little rubber feet. The feet
	sound silly but are
	actually pretty important: without them, the machine wobbles when you
	type. The guy at Best Buy said the parts department would be happy
	to send me a set of feet and I can stick `em on myself.
	&lt;p&gt;
	Today the UPS guy delivered them. Each foot was in a separate ziplock bag. Each ziplock
	bag was in its own padded envelope. The four padded envelopes were
	put in a large cardboard box. The box was then filled with foam peanuts.
	&lt;p&gt;
	&lt;img src="http://fluff.info/blog/asst/boxedfeet.jpg"&gt;&lt;p&gt;
	A set of four feet is a little under 2 cubic centimeters (ccs). The box
	is about 30 x 23 x 15 cm, or 10,600 ccs, and so is 5,600 times larger than
	the set of items it is holding. Thanks to such care, none of the feet broke
	during shipping.
	&lt;!-- box is 12 x 9 x 6 inches. one pair of feet is .9 x .75 x .75 cm. --&gt;
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/20000245.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>25 January 2009 00:00:00 +0000</pubDate>
</item>


<item>
        <title>Parade of motorcades</title>
	<content:encoded>
I attended a little inaugural party over on the fourth floor of an
apartment overlooking Mass ave.  On the night of the election, when Obama
won, that intersection was taken over by whooping pedestrians until 2 AM. 
But on the inaugural night, the evening gridlock was punctuated only by
the sound of motorcades.
&lt;p&gt;
Most days in DC see few motorcades. Are we really going to occupy a
dozen emergency vehicles every time the Secretary of the Interior goes
out to lunch? So I suspect that the stream of Inauguration day motorcades
were more for prestige and a glam night out. The president himself
attended a dozen and a half balls, and his motorcade passed by twice
that day. By odd coincidence, the DC police made &lt;a
href="http://www.washingtonpost.com/wp-dyn/content/article/2009/01/20/AR2009012004139.html?hpid=topnews"&gt;no
arrests&lt;/a&gt; that day.
&lt;p&gt;
Motorcades have a special sound. When an emergency occurs, police
from all over have to turn around and get going, but a motorcade radios
ahead to clear the path and travels as a pack, where everybody works as a
rapidly-moving unit. The police seem more comfortable honking their horns.
A motorcade begins and ends with motorcycles, which in a group also
provide a distinct sound. So you get the sound of police blocking
traffic, silence for a few minutes as the streets are clear, then the
motorcycles and sirens.
&lt;p&gt;
The Presidential motorcade is worse: you have at least four motorcycles
on front and back, and a whole lot of black vehicles in between,
including a spare limo, a few vans and SUVs, and what looks like an 
all-black ambulance, I presume carrying equipment. Two blocks behind,
in with traffic, there'll be three or four more police.
&lt;p&gt;
The party I was at was nothing unique: that part of Mass Ave is two 
walls of apartments facing each other, and I'm sure many a party was
being thrown that night. In the valley between the cliffs, the motorcades
ran all night, as those in the motorcade set travelled among their own
ground-level parties.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000245.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>22 January 2009 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Columbia</title>
	<content:encoded>
If you really think its proper name is "District of Columbia", then I expect you to refer to "The District of District of Columbia" from time to time.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000245.htm</guid>
        <link>http://fluff.info/blog/arch/00000245.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>24 December 2008 00:00:00 +0000</pubDate>
</item>


<item>
        <title>A high school geometry lesson</title>
	<content:encoded>
It must be hard to write high school math books. You have to make the subject interesting without making it
&lt;em&gt;too&lt;/em&gt; interesting.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000244.htm</guid>
        <link>http://fluff.info/blog/arch/00000244.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>28 November 2008 00:00:00 +0000</pubDate>
</item>


<item>
        <title>Open comments</title>
	<content:encoded>
In late 2003, I did a post entitled "&lt;a href="http://fluff.info/blog/arch/00000021.htm"&gt;Should Israel exist?"&lt;/a&gt; It's still the first hit on Google for that phrase. The post will be as interesting or not as any of my present posts, and Google tells me the page has only seen 400 readers to date in 2008. But the comments to this post are a li'l train wreck. After all, you don't type "Should Israel exist" into a search engine without having a strong opinion about the subject. So I recommend paging down to the comments for what you will find to be either profoundly funny or profoundly depressing. It's a three-part essay, so continue through to #23 to get to comments like fearsomefire@gmail.com's "Okay I hate jews to the core."
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000243.htm</guid>
        <link>http://fluff.info/blog/arch/00000021.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>5 November 2008 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Capital, liquidity, and the crash</title>
	<content:encoded>
A guest blog today: a macroeconomist who works with issues of bank stability responds to my microeconomist-dabbling-in-macro post from last time.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000243.htm</guid>
        <link>http://fluff.info/blog/arch/00000243.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>24 October 2008 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Velocity, risk, and the crash</title>
	<content:encoded>
I only worked at the brokerage firm for about a year and a half. I
didn't last long partly because I was much more interested in how the various
systems of instrument trading worked than the part about maximizing the company's
profits.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000242.htm</guid>
        <link>http://fluff.info/blog/arch/00000242.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>10 October 2008 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Causality and ethics</title>
	<content:encoded>
Thesis sentence: the perception of causality is what makes us human, in both positive and negative ways.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000241.htm</guid>
        <link>http://fluff.info/blog/arch/00000241.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>24 September 2008 00:00:00 +0000</pubDate>
</item>

<item>
        <title>The Cult of the Amateur, by Andrew Keene</title>
	<content:encoded>
The Cult of the Amateur, Andrew Keene
&lt;p&gt;
The thesis of this book is that we need to maintain a division between
the producer of a work and the audience. If the amateurs in the audience
start producing the work, then we just get amateur results, and the
world fills with noise.
&lt;p&gt;
Mr. Keene has run a few web sites, including a traditional curated-content
music site, where you'd buy Bach or Dylan CDs. But this is not a tech
book, because it is only looking at the consumer face of the `Net; his
tech-world expertise and experience is therefore not super-relevant.
No, this is a Sociology book, on how technology shapes culture, and
how different organizational structures lead to different societies
(and vice versa).
&lt;p&gt;

In the world of Sociology, Mr. Keene is an amateur. He has no formal
sociological training. His citations top out at a few books
that might appear in a first-year Sociology course. His authority on
questions of how society shapes itself is exactly as authoritative as
any other curmudgeon with a `Net connection. As an amateur, he doesn't
really have the training to hold the debate on his own, as we expect a
book to do: we want him to present his argument, then tell us what the best of the best on
the other side think, then tell us why he believes his arguments
should win. There are good pop press books out there that really peg
this dialectic, often by academics.
&lt;p&gt;

My own experience writing books via the traditional
publishing scheme, with an editor who played gatekeeper and coach,
showed me the work that goes into making sure every single sentence is
accurate and backed up (which frankly does not go into many online
books, but we all know that already). On the level of attention to
the details of writing, Mr. Keene is something of an amateur as well,
who lowers the quality of his work with unfounded offhand comments,
like how Chuck D of Public Enemy was the "first serious rapper", or
how the respect of property and not stealing is "Judeo-Christian", or
how that Judeo-Christian belief applies directly and unequivocally to
intellectual property. [I could go on all day with examples.]
&lt;p&gt;

But you don't need to be a published author to recognize bad writing
and bad logic, and that's another key failing of the book: generally,
people know crap when they see it, and the problem of filtering out crap
exists even in plain print. Let us say that only professionals with
certification may have license to write (or sing or dance). 
Just reading the ten or so high-quality peer-reviewed academic journals in a field, cover to cover,
would literally be a full-time job. The world blew past our ability to be on top of all of everything
all the time, and if you don't design a good filter to work out what to
focus on, you're screwed. This isn't about the Internet or the amateur,
but about a world where education is common, the population is rising,
and the day remains stubbornly stuck at 24 hours.(*)
&lt;p&gt;

We have a lot of mechanisms to sort out what is or is not crap, including
popularity among the masses, popularity among friends, gut feeling, and
recommendations by authorities and gatekeepers.  Mr Keene insists that
the only correct method is recourse to recognized authorities. This
is another place where the literature is dense, and Mr Keene cites
none of it, perhaps because there are so many benefits to the other
reasons that he can't counter. Because the problem is identical for both old and new media,
we can give an example from old media: some people will pick up this
book just because they'd heard of it (recourse to broad popularity,
more or less), or because a friend handed them a copy (that's my story), or because the
theme that the Internet is making the sky fall appeals to their gut. I
expect that few will pick it up because of recommendation by qualified
authorities.(**)
&lt;p&gt;

But hey, Mr. Keene had the opportunity to publish, even via bound-paper
means, as he prefers. I can only hope that he recognizes the wonderful
empowerment that he has gained from a society where any amateur such as
himself has a shot at presenting his or her work.
&lt;p&gt;

(*) See also David Foster Wallace's essay in &lt;em&gt;The Best American
Essays 2007&lt;/em&gt;. He curated the volume, and was thus the Decider whom the reader
contracted to select which essays he or she would read. He
found the setup to be slightly uncomfortable, and I think he summarizes
the emotional state that Mr. Keene probably had when writing his book: "...we are
starting to become more aware of just how much subcontracting and
outsourcing and submitting to other Deciders we're all now forced to do,
which is threatening (the inchoate awareness is) to our sense of
ourselves as intelligent free agents."
&lt;p&gt;

(**) Head over to 
&lt;a href="http://www.amazon.com/Cult-Amateur-Internet-Killing-Culture/dp/0385520808/ref=si3_rdr_bb_product"&gt;Amazon.com&lt;/a&gt;, 
and you'll see that Publisher's Weekly
hated it, and the short back-cover platitudes are from folks like
Larry Sanger, who is explicitly praised as a Solution in this book.
[Notice how you could easily crop the clearly negative Publisher's
Weekly comment into praise on par with the back-cover lines, which is
a reminder that we should never trust a review that consists only of a
sentence fragment, because the next word could so easily be &lt;em&gt;,
but...&lt;/em&gt;.] No comments on Amazon from any sociologists.

</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000240.htm</guid>
        <link>http://fluff.info/blog/arch/10000240.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>17 September 2008 00:00:00 +0000</pubDate>
</item>


<item>
        <title>Google OS (aka Chrome)</title>
	<content:encoded>
What happens when I knock out a blog entry in an hour or so. Could've tied in the history of server/client relations a little better, but it does make a comeback at the end.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000240.htm</guid>
        <link>http://fluff.info/blog/arch/00000240.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>3 September 2008 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Statistics as unbearable longing</title>
	<content:encoded>
This one is a bit meta-, in that it comes close to but doesn't quite get at what I'm trying to express. Maybe next time, when I elaborate on the story of Little Red Riding Hood.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000239.htm</guid>
        <link>http://fluff.info/blog/arch/00000239.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>26 August 2008 00:00:00 +0000</pubDate>
</item>

<item>
        <title>The NYT takes a side on the copyright war</title>
	<content:encoded>
The link is to a New York Times article on Girl Talk, a DJ who does very heavy sampling---running hundreds of samples for maybe only two or three seconds each---and how the record labels claim that every last sample used needs to be paid for. The question is whether a de minimus sample (in a transformative work at that) counts as fair use of a copyrighted work, or whether even the tiniest clip is somebody else's work that needs to be paid for. Now, your newspaper typically tries to be objective and let the sides take their own sides. But here, the NYT did something interesting: they gave us a multimedia box that lets us listen to a minute and a half of Girl Talk's work, and therefore the work of the twenty people whom Girl Talk sampled in those short clips. It is very hard to argue that Girl Talk is infringing the copyright of, say, LL Cool J, and yet when NYT plays that exact same clip in its entirety, the NYT is not.  [Yes, I do know how one would go about making that argument; it's not interesting.] By posting the clip, the NYT's lawyers are directly asserting that they believe the labels are wrong and that de minimus use is fair use. 
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000238.htm</guid>
        <link>http://www.nytimes.com/2008/08/07/arts/music/07girl.html</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>6 August 2008 00:00:00 +0000</pubDate>
</item>


<item>
        <title>Two sides of the statistical war</title>
	<content:encoded>
I think that one of the main reasons that people don't know what's going on with regards to modern statistics is that there are actully two fields tied together. Schools teach them as one field, and many people talk about them as one, but their goals and methods often directly conflict. And for those of you who find stats to be boring and are just here on the Internet looking for porn, I also threw in a full-color figure of what may be genitalia.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000238.htm</guid>
        <link>http://fluff.info/blog/arch/00000238.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>6 August 2008 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Press siete to borrar.</title>
	<content:encoded>
I'm not really bilingual, but I've lived in Spanish-speaking countries
long enough that I can hold most of a conversation, and I'd like to not
forget too much of it. So, given an option, I put various silly things in
Spanish. Instead of telling me to press seven to delete the message, my
voicemail suggests that I marque el siete. It's not like any of us really listen to
those messages after about the third time anyway.
&lt;p&gt;
Using most web sites in not-English is a pretty terrible experience. Few
pages are entirely in English, and you frequently get bounced back to
all-English pages. I can't say what the average non-English
speaker can do, but I frequently hit decision points on allegedly
Spanish-ized sites that require a lot of English. My favorite new toy
of late has been last.fm, but if you change your language of choice to
Spanish, your account gets transferred to last.fm.es, which seems to
be a version or two behind on the technology. Why does "musica" get a
clunkier interface than "music"? Last.fm is just a toy, but other dificil-a-navigar
sites are a little more on the side of
vital infrastructure for a person living in modern society.
&lt;p&gt;
When you call a corporation's customer service department, the person you
talk to will frequently go to your account on the web and read your data
to you. I've always been annoyed by that---if I'm on the phone, then I've
already spent ten minutes clicking things---though I realize that for some
people, reading their own data to them is necessary.  Anyway, it's pretty
amusing when your service rep (who is much more likely to be in  Oklahoma
than in India) gets stopped cold trying to select between "Mi perfil",
"mi cuenta", and "ayuda". Ha ha, you can't access key information via 
any means besides the web site anymore, either.
&lt;p&gt;
As another point that reveals the workings behind the curtain, I've
changed the background color of my browser to a sort of wheat color,
which is not as staring-into-light-bulb as the default white. [Right
now, I'm typing
this text over a pale blue background.] But that change reveals all
sorts of interesting choices made by web designers. Some override all
colors on the entire page---you're getting a white background whether
you want it or not. Some are inconsistent, and have white-background
pictures but transparent-background text, or vice-versa. A reasonable
compromise is the white box on your choice of background, like the set of of squares
on yahoo.com. I also tend to keep my text zoomed fifty or sixty percent
larger, again to be easier on my eyes. This used to have absurd effects
on most web pages, though designers seem to be learning on this front. 
&lt;p&gt;
OK, so some of this, like getting your customer service rep to actually
confess that they have no more information than you do, is just
amusement.  The background color thing is cute, and maybe a warning to
those of you who are web designers, but I don't care that web builders don't
design around my eccentric color preferences or concern themselves with
those who have less-than-perfect vision. The language problem, however, is a bigger deal,
and shows that as fabulous as this Internet thing may be, we're still a
long way from equal information access, even regarding
infrastructure-level services. Lots of web sites have contracted
a translator to translate most of the terms on their sites, but few web
sites truly commit to hosting multiple languages.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/20000237.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>21 July 2008 00:00:00 +0001</pubDate>
</item>

<item>
        <title>A history of telephones</title>
	<content:encoded>
&lt;img src="http://fluff.info/blog/asst/phone_history.jpg"&gt;&lt;p&gt;

These are all of the cellular telephones I've had, up to the rightmost 
one that I got last month. [The screen on that one is bright because I 
was on hold with the IRS when I took the picture (and as I write 
this).]&lt;p&gt;

All have more-or-less different brand markings: Qualcomm/Kyocera, Samsung,
Handspring Treo, PalmOne Treo, American Telephone and Telegraph by Palm.
You can see when whip antennas fell out of fashion, and when people
started thinking keyboards on a telephone could work (the middle one 
has a keyboard under the lid).  But the software is basically the same:
they all run the Palm operating system.&lt;p&gt;

Palm as a company has been a mess, with endless merging and submerging 
and emerging and remerging: bought by US Robotics, then 3COM, then spun off, then
split, then merged, ... and I'm certain that for every step along the
way the CEOs got richer while the product got more rickety. You have no  
guarantee that the company that sold you Palm hardware this year will
keep doing so next year.&lt;p&gt;

But right from the start, before the mergers, the software people got a number of
things right: the data is in a flat database whose format is at this
point simpler than the company's merger history, the operating system's
functioning is equally simple, and the
company (still) is open about all the software's workings.
It's so easy to write a program for the Palm OS that even I've done 
it.&lt;p&gt;

So that's allowed for commodification of the hardware.
Large-form computers (i.e., desktop and laptop machines) are already
commodified:
the operating system runs the same programs that a hundred other computers from a hundred other
brands can run. When your machine breaks, pull out the hard drive and
plug it into another.&lt;p&gt;

People aren't so interested in the
commodification of small-form computers (like telephones and music
players), but it provides just as much benefit. Just load the
old software and databases onto the new machine and pretend you didn't
drop your telephone trying to open the flip lid while biking. 
No matter how badly-made the plastic junk, no matter how much the guy on
ebay lied about how it's in TOP SHAPE---MINT!!, you have
complete continuity.&lt;p&gt;

I've had a Palm machine since a day or two before 28 May 1998, which I
can tell you because my current telephone says that I had a doctor's
appointment that day at 2:20 PM. Need 
the telephone number of anybody I've met since 1998? I've got it for ya,
such as somebody named "C" in the 626 area code. I've added `file
form 8889 so the IRS will give me back my $300' to the to-do list, so that
I'll be able to fondly look back on the time the IRS said I had a math error.
And that's how continuity across hardware has enriched my life.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000237.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>7 Jul 2008 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Still just parametrized models</title>
	<content:encoded>
Last month, Wired ran a big thing praising SUVs and not-organic food; everybody was irate but I didn't have much to say.
But this moth, they ran a cover story about how parameter searches over large data sets is a paradigm shift from older methods of modeling.
&lt;i&gt;Fuck that.&lt;/i&gt; 
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000237.htm</guid>
        <link>http://fluff.info/blog/arch/00000237.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>2 July 2008 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Data is typically not a plural</title>
	<content:encoded>
	Until writing this, I hand't realized what a font of conflict this grammatical quirk is.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000236.htm</guid>
        <link>http://fluff.info/blog/arch/00000236.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>26 June 2008 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Bloggers through history</title>
	<content:encoded>
	I might pick up on this further next month in the context of my trying to get a nice, traditional job.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000235.htm</guid>
        <link>http://fluff.info/blog/arch/00000235.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>4 June 2008 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Unsolicited financial advice</title>
	<content:encoded>
I dunno, I've just really disliked my writing lately. I wrote this a month ago, and every time I was going to post it just wound up tweaking something. Meanwhile, here's an amusing post about the 
&lt;a href = "http://njrereport.com/index.php/2008/04/09/tracking-realtor-spin/"&gt;National Association of Realtors (R)&lt;/a&gt;.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000234.htm</guid>
        <link>http://fluff.info/blog/arch/00000234.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>24 April 2008 00:00:00 +0000</pubDate>
</item>


<item>
        <title>Best of SXSW 2008</title>
	<content:encoded>
The entry is just liner notes. Here's a link to 
&lt;a href = "http://www.eatlime.com/download.lc?sid=223DF741-56ED-8D48-637B-74847E6FBDF8"&gt;2 hours of fun music&lt;/a&gt;.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000233.htm</guid>
        <link>http://fluff.info/blog/arch/00000233.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>22 March 2008 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Hyperrational is boring</title>
	<content:encoded>
I used to be the Risk Control Analyst at a brokerage firm on Lasalle Street in Chicago. I was just north of the Federal Reserve---had a great view of their loading dock. After a year, I quit and moved to Spain.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000232.htm</guid>
        <link>http://fluff.info/blog/arch/00000232.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>20 February 2008 00:00:00 +0000</pubDate>
</item>

<item>
        <title>On Senator Obama</title>
	<content:encoded>
Barak Obama came to a lunch at the think tank I like to hang out at
a few years ago, before the speech that made him famous, when he was
merely the Senator from Illinois.
I remember two things from the meeting: first, I recall that he wowed everybody, and that everybody left thinking `this guy is gonna go far!' And for the sake of comprehensively listing my recollection from the meeting, I recall 
that he's missing a joint on one of his fingers.  You'll notice that he very much leads with 
his left hand, and only rarely will you get a clear shot of his imperfect hand:
&lt;p&gt; 
&lt;img src="http://files.meetup.com/373491/obama_hands_rally.jpg"&gt;
&lt;p&gt; 
&lt;em&gt;And that's all I remember.&lt;/em&gt; 

The guy spoke to a room of about two dozen policy analysts for about an hour and a half, about actual policy topics, and all I can recall is a generally positive affect. That makes him the perfect politician.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000231.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>13 February 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>NIH Contracting: a how-to</title>
	<content:encoded>
First, I deserve some kind of award for how few expletives I used here. Second, there are still a few points on the philosophy of organization that should still be interesting to those of you who don't care about the details.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000231.htm</guid>
        <link>http://fluff.info/blog/arch/00000231.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>4 February 2008 00:00:00 +0000</pubDate>
</item>

<item>
        <title>The tyranny of the majority: design edition</title>
	<content:encoded>
I've had this one on hold since October, debating whether I should post it or not. Well, here it is. A post clearly explaining Jeff Koonz's entire artistic career.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000230.htm</guid>
        <link>http://fluff.info/blog/arch/00000230.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>18 January 2008 00:00:00 +0000</pubDate>
</item>

<item>
        <title>On writing</title>
	<content:encoded>
From San Diego to Zurich, I got comments that my last entry was a total bummer, so here's the perkiest entry I could muster to make up for it. I could have taken this in several depressing directions---primarily regarding whether our fellow people are a resource or a constraint---but for you, dear reader, I elided all that.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000229.htm</guid>
        <link>http://fluff.info/blog/arch/00000229.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>10 January 2008 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Musical reverse</title>
	<content:encoded>

I know you've heard Darling Nikki, because it's on the soundtrack to
Purple Rain, which sold 8.7 billion copies in the early 80s. Prince was
a contrast to the hair rock/D&amp;D metal bands that had tracks 
that allegedly said Hail Satan! when you played them backward.
&lt;p&gt;
So the end of Darling Nikki has a part that is clearly reversed. I found
a reel-to-reel tape recorder at a yard sale one day, and it came with a
box of about forty Spanish lessons, so I actually had reason to pick it
up. I think I got to about lesson three before sort of tapering off, but
now that I had this behemoth in my room, I could record Darling Nikki
onto a blank tape, flip it over, and hear the ending in reverse. It
says: `Hello. How are you? I am fine, `cause I know the Lord is coming
soon.' 
&lt;p&gt;
Shortly thereafter, I had an unpleasant breakup with a
girlfriend, who wound up with all my pants and the reel-to-reel player,
which she used to add a bit of height to an endtable.  
&lt;p&gt;
Now that we live in a digital world, it's not so difficult anymore. Just
open the track in Audacity (which is free and open and installs via your
&lt;a href="http://fluff.info/blog/arch/00000137.htm"&gt;package manager&lt;/a&gt;),
select all, and click on Reverse in the Effects menu.  Bam! What Prince
hid below the noses of millions of listeners is in plain view.
&lt;p&gt;
I tried it with Philip Glass. I'm sure you know his music: it's 
minimalist. Like minimalist art, it produces a large expanse of
identical landscape, so minor changes stand out. 
I've listened to the second movement to his Concerto for Violin and Orchestra
several hundred times, so I thought it'd be interesting to see what it
sounds like from a new perspective. And ya know what, it sounds very
much the same.  Most of the instruments in the orchestra are playing something so
cyclical that it sounds exactly the same played backward.
But over this identical backdrop, the lead violin sounds
somewhat different (and is playing a reversed melody). Suddenly, I have
twice as many Philip Glass albums.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000228.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>5 January 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Social technology</title>
	<content:encoded>
You may have noticed that I'm very interested in systematic failures and problems that have no solution. Maybe I need a perkier hobby.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000228.htm</guid>
        <link>http://fluff.info/blog/arch/00000228.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>24 December 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Academia doesn't scale</title>
	<content:encoded>
Much of academic custom hasn't changed for about three hundred years. In other words, the peer review system predates American Democracy.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000227.htm</guid>
        <link>http://fluff.info/blog/arch/00000227.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>14 December 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Single-payer health care and transparency</title>
	<content:encoded>
The U.S. health care system is in many ways the perfect example of how complexity constraints can cause a free market to fail.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000226.htm</guid>
        <link>http://fluff.info/blog/arch/00000226.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>26 November 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Installing Windows</title>
	<content:encoded>
I finally broke down and installed windows. Of course, I didn't do it
myself; I'm just a loser writing a computational stats textbook, so it's
well beyond my technical abilities. So I should say that two guys came
over and put in a shift from 8 AM to 4 PM to do the installation.  And, as
is always the case with these things, at about 3 they told me that
they'd be back the next day to finish up, since the whole installation
process is taking a total of about three person-days. It's all very 
high tech, and you can really tell that a lot of engineers put a lot
of effort into the product, even if the process of installaing into an
existing setup is a complete mess.&lt;p&gt;

But now that everything is installed, it's pretty nice to lean back here
in my room, staring at (through?) my three big screens side-by-side, 
feeling generally warm and cozy. Really, I should've done this a long time ago.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000225.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>20 November 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Properly leaving Yahoo!</title>
	<content:encoded>
If you're reading my lil' blog, you know I'm boycotting Yahoo!, because of its enthusiastic cooperation with the Chinese government's torture and imprisonment of pro-Democracy activists. &lt;p&gt;
But I forgot one detail: I'm still handing over all my content to Yahoo! for their use in improving their search engine.
Yahoo! depends on us, the content providers, for its business, so we have a large say in how they go about making money.&lt;p&gt;
 

So, I've added this to my 
&lt;tt&gt;robots.txt&lt;/tt&gt; file in reference to Yahoo!'s web spider (Slurp), and you, dear reader, are encouraged to do the same. &lt;p&gt;

&lt;tt&gt;
User-agent: Slurp &lt;br&gt;
Disallow: / &lt;br&gt;
&lt;/tt&gt;&lt;p&gt;

You'll still have no problem finding me on search engines that don't endorse torture, and if Yahoo! puts in place a policy of not handing people over for torture in the future, I'd be happy to let them make ad revenue from my content again.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000225.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>1 November 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Suggestions for the ISO/IEC C committee</title>
	<content:encoded>
In which I open a sentence with the phrase
`Because you can send an anonymous inline struct whose values are set via designated initializers...'. If that doesn't make sense to you, don't even bother clicking through. If it does, well, don't just sit there in suspense....
&lt;p&gt;
Oh, and if you feel like you know C but that still doesn't make sense, then you're eight years behind; have a look at
&lt;a href="http://home.datacomm.ch/t_wolf/tw/c/c9x_changes.html"&gt;what's new in C99&lt;/a&gt;---notably numbers 16 and 17 in that list.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000225.htm</guid>
        <link>http://fluff.info/blog/arch/00000225.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>30 October 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Definition of a crackpot</title>
	<content:encoded>
Actual dialog I had the other day (that did not inspire this essay, which I wrote about four months ago):&lt;br&gt;
Me: I'm the author of a textbook on statistical computing, forthcoming from [well-regarded academic press], and I have a nifty idea for aggregating information from some genetic tests that I wanna run by you.&lt;br&gt;
Him: But you don't know anything about the chemistry of allele calling mechanisms, so how can your statistics possibly work?
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000224.htm</guid>
        <link>http://fluff.info/blog/arch/00000224.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>20 October 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Why your drugs are not vegetarian</title>
	<content:encoded>
Yes, I do believe there should be some medical exemption, but
no, I have no idea of where I would draw the line. But it's not `anything sold in pill form'.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000223.htm</guid>
        <link>http://fluff.info/blog/arch/00000223.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>14 October 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Talk like a statistician day</title>
	<content:encoded>
Well, it's Talk Like a Pirate day, when computer geeks everywhere
discuss BitTorrent, sailing technique, and the R programming language.
&lt;p&gt;
I had a brief discussion with a linguist pal or two the other day about
the Sapir-Whorf hypothesis, which broadly says that choice of language
influences thought. Ms AO of Philidelphia, PA, explains that linguists
toss around two versions of the S-W hypothesis: the one that's so literal
that it can't possibly be true (that if there's no word for blue then
people can't concieve the concept of blue) and the version
that's so vague that it can't possibly be disproven (that if it's easy
to say something, then it's easy for your little back-of-mind
voice to think it). 
&lt;p&gt;
Working linguists will give you much more subtlety than that, and there
are many grey shades in between the extremes to be explored. But
computer scientists, oh, they are in &lt;i&gt;love&lt;/i&gt; with Sapir and Whorf,
and will spend all day bickering about how important the choice of
language is.
&lt;p&gt;
Bradley Efron,  a famous statistician (yes, I know that's an oxymoron),
points out in a PDF regarding &lt;a
href="stat.stanford.edu/~brad/talks/future.pdf"&gt;the future of
statistics&lt;/a&gt; that we currently have a dozen analysis paradigms
all being used at once:
there's Bayes and Classical and machine learning and all sorts of
weirdness that people are dreaming up every day.
&lt;p&gt;
The future will go to those who can synthesize those several threads
together.
&lt;p&gt;
Statistical languages like R typically have no easy means to do this.
This is me putting words to why &lt;a
href="http://apophenia.info"&gt;Apophenia&lt;/a&gt; has me so very enthused right
now---Apophenia allows the arbitrary combination of models to do
synthesis like a linear model estimating metaparameters based on data
from clusters of Bayesian-estimated subparameters. If that reads like a
gibberish to you in English, imagine trying to code it. The closest I
know of to making this work is a thing called BUGS, but its syntax
is not superpleasant.
&lt;p&gt;
Why? Because the system has no means of using a model as a noun. A model
in every stats package I know of is a set of verbs to take the input
data (a noun) and filter it in various ways (verbs). The models live
entirely at the verb level.
&lt;p&gt;
But the model itself can be bundled into an object and filtered: it can
be updated, constrained, joined with other models, and otherwise
tweaked, just as a data set can be filtered. Taking a base model and
modifying its details allows for more realistic models to be written
more quickly.
&lt;p&gt;
And that's why I'm enthusiastic about Apophenia, even if nobody else
cares about it: the syntax for a model is a standard object, not a
separate model-specifying grammar that invariably describes only a 
limited set of verbs. 
&lt;p&gt;
It would be silly for me to claim that this
is original, because nothing in this world is original; let me know if
you know stats packages that accommodate specifying and then
transforming models, so I can learn from them and steal their ideas. 
In the mean time, the textbook is now officially forthcoming from
[Academic press] so this time next year, maybe a few people will have
this idea in mind, and pirates next year will be able to talk about
models as nouns.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000222.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>19 September 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>How to write about being organized</title>
	<content:encoded>
Is it just me, or is the Web filled to the gills with this stuff?
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000222.htm</guid>
        <link>http://fluff.info/blog/arch/00000222.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>18 September 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>My immense disappointment with the software industry</title>
	<content:encoded>
This concludes my five-part series of heavy-handed moralizing. One reader says that I'm being unfair to Jane Austen in comparing her to the modern software industry; if so, I apologize.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000221.htm</guid>
        <link>http://fluff.info/blog/arch/00000221.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>14 September 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Neil Diamond, "America"</title>
	<content:encoded>
The only version I've heard is the live version (Hot August Night is a concert album), which includes a rousing opening by a string section.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000220.htm</guid>
        <link>http://fluff.info/blog/arch/00000220.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>30 August 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>My family</title>
	<content:encoded>
And now I await the inevitable onslaught of comments politicizing my family history. I may revise this entry as more anecdotes come to me.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000219.htm</guid>
        <link>http://fluff.info/blog/arch/00000219.htm</link>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>22 August 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Maps are fun.</title>
	<content:encoded>
Today's web site recommendation is
&lt;a  href="http://strangemaps.wordpress.com/"&gt;Strange Maps&lt;/a&gt;, a collection of maps about the real world viewed askew, countries that don't quite exist, 
and other arcana about the world, history, and culture. The imaginary line dividing two countries is often the basis of major, long-running conflicts---what your high school teachers called History---so the writeups range from light overviews of the map to full history lessons. Also, being a map site, there are lots of pictures.
&lt;p&gt;
[You could also read this as a lead-in to my forthcoming next entry---any day now---about my history with an ambiguous country.]
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/30000218.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>17 August 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Yahoo!: a brief follow-up</title>
	<content:encoded>
The head of the 
&lt;a  href="http://arstechnica.com/news.ars/post/20070806-congressman-slams-yahoos-despicable-practice-in-china-vows-to-investigate.html"
&gt;House Foreign Affairs Committee&lt;/a&gt; is also investigating Yahoo! over a senior VP's evidently false claim that the company did not know why the Chinese government was requesting information that the government would then use to put a journalist in prison for a decade.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/20000218.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>04 August 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Like a coffee shop, but without the coffee</title>
	<content:encoded>
I'm trying to work in museums more. My normal reflex reaction is to just
work at the coffee shop, but why not a museum? Many of them are free
every day of the week, the quiet atmosphere is conducive to working, you 
get wireless if you're lucky (which, evidently, I am), you don't 
feel obligated to drink lots of iron-absorption-blocking coffee, and the artwork
is a step or ten up from the art school student stuff you get at
most cafès.
&lt;p&gt;
The idea partly comes from the &lt;a 
href="http://www.gardnermuseum.org/index.asp"&gt;Isabella Gardner Stewart museum&lt;/a&gt; in Boston. The little
map/flyer explains that the museum's proprietors see the intent of a
museum as not aimed toward admiring artwork over there on the wall, but
of experiencing inspiration through a sort of immersion. Each room was
loosely arranged around a theme, and the artwork included much art
beyond paintings and sculptures, including a large central garden and
lots of interesting furniture. Unfortunately, because the furniture was
a part of the artwork, there was nowhere to sit.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000218.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>01 August 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Yahoo: a person lacking cultivation or sensibility</title>
	<content:encoded>
Me: The next blog is gonna be about Yahoo!. &lt;br&gt;
Pal: Is it going to be about why they're evil? &lt;br&gt;
Me: Uh, yeah. Am I getting predictable? &lt;p&gt;
Next time will not be about an evil government agency or company, I swear. It'll be about why I write about evil government agencies and companies.
</content:encoded>
        <link>http://fluff.info/blog/arch/00000218.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000218.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>26 July 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>The US Trade Representative: Not working for you</title>
	<content:encoded>
I have made serious efforts to come up with non-corrupt explanations for the USTR's various positions, and there is seriously nothing forthcoming.
There are a number of such government offices that do work so technical and dull that the popular media ignore them, and therefore they are free to provide their services to the highest bidder without public oversight. I think it's the greatest flaw in the theory that the media can regulate a Democratic government.
</content:encoded>
        <link>http://fluff.info/blog/arch/00000217.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000217.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>2 July 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>How to make spam</title>
	<content:encoded>
I learned how to make spam. It is fun.
&lt;p&gt;The algorithm is from the 1980s.
I learned from &lt;em&gt;The Practice of Programming&lt;/em&gt;, a book written in 1999 by two coding lumniaries, Kernighan and Pike. As an example of fun data structures, they proposed the following Markov chain algorithm:

&lt;p&gt;
--Open up your favorite text.
--Read in three words. Write down the first two in the prefix table,
and the third as a suffix for that prefix. &lt;br&gt;
--Step one word forward, so the text's words two and three are now the
prefix, and word four is the suffix. &lt;br&gt;
--Repeat through the whole text. Every word will have a turn as the first, second, and third word in the chain. At the end of this, you have a list of every two-word prefix, and under each such pair, every word that follows that pair.&lt;p&gt;

--Now produce a text. Start with any two word prefix. Go to that prefix's list, and randomly select one successor word. &lt;br&gt;--Step forward: use the second word of the prefix and the successor to 
form a new two-word prefix. Go to that point in the list, and pull the
next successor from that list.&lt;br&gt;
--Repeat until you have an email-length amount of text.&lt;p&gt;

The end result is a text where every three-word sequence appeared in the original, but the order is otherwise a mess.
Kernighan and Pike implemented this in twenty lines of Perl, and my breezy C version is 45 lines (26 semicolons). But the results read semi-coherently:
&lt;p&gt; 
365 Even if a fool brings suffering on himself. He does not know the true teaching is difficult, and the achievement of Buddhahood is difficult.&lt;br&gt; 182 To abstain from all bonds - that is what one has grasped this as it is. A disciplined mind leads to death.  Those who have understood the truth. &lt;br&gt;70 Like fresh milk a bad rebirth.  &lt;br&gt;316 Seeing danger where there is, by holding to right views people go to hell, the good does travel against the wind, and a brahmin lose his temper. Shame on him who talks a lot. They even criticise him who strikes a brahmin, and nor should a brahmin lose his temper. Shame on him who loses his temper because of their evil deeds. It is a delightful spot. &lt;br&gt; 98 Delightful for them are the well a nd truly restrained.
&lt;p&gt;
You can see that some is legible (Shame on him who talks a lot.) and some semilegible fun (Like fresh milk a bad rebirth).
&lt;p&gt;
The variety comes in the training text. You can tell that I used a text by the Buddha: &lt;a href=http://www.gutenberg.org/etext/2017&gt;the Dhammapada&lt;/a&gt;. But using another Project Gutenberg text to build the tree of prefixes and their suffixes will create an entirely different-feeling text. So next time you get a spam email like the above, you can try to guess what the training text was.
&lt;p&gt;
By the way, notice that this evil and annoying technique for generating spam combines an algorithm from a well-regarded textbook with free texts from Project Gutenberg. That is, our spammers combined two entirely wholesome and well-meaning projects to produce a product that pisses everybody off.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/40000216.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>28 June 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>More Open Source office politics</title>
	<content:encoded>
The next full entry (next week, I expect) is about the politics of software in a manner that
really means something, but for now, here's some amusing turnabout
office politics for ya.
&lt;p&gt;
It's to be expected, now that there's money to be made from
open source, that people would attempt to grab power. Today's event: The
Open Source Institute, unable to trademark the term Open Source, asks
everybody to please
&lt;a href=
http://www.linux-watch.com/news/NS9666856083.html
&gt;
just only use the term when the OSI approves
&lt;/a&gt;.
&lt;p&gt;
They make a fair point that there are a whole lotta licenses out there,
and some are a little spurious. And I'd normally think that requests 
like this are kinda cute. But the request that OSI be the
central arbiter of what is Open Source produces just as much
potential for abuse as the "problem" of unregulated licenses. The examples they give of abuse are regarding companies that
require users to post a link to their website as a condition of use.
This more-or-less came up before, with a
&lt;a href=
http://www.gnu.org/philosophy/bsd.html
&gt;
clause in the BSD license
&lt;/a&gt;
that required users to put a line in the documentation in the way of `portions Copyright
Regents of the U of California'. The link to the Free Software
Foundation in the last sentence acknowledges that this requirement is
"obnoxious" but is still a valid type of free software license.
&lt;p&gt;
And ya know, the FSF is not known for being reasonable. I already ranted
about how much of an anal-retentive power grab the FSF's GPL v3 is. But
their rules for what makes an open source software license are pretty
darn simple. There are
&lt;a href=
http://www.gnu.org/philosophy/free-sw.html
&gt;
only four rules
&lt;/a&gt; (numbered using offset numbering rather than index
numbering. Uh, haha.), and they're all pretty basic. By my count
(and for BSD, the FSF's opinion too), both the BSD and "badgeware" licenses
fit those rules. I mean, adding an acknowledgement or copyright notice
is only so onerous.
&lt;p&gt;
In this case, the stereotypically hard-ass fanatics at the FSF are being &lt;em&gt;more
&lt;/em&gt; open and accepting than the open source people.
If the OSI power grab were
somehow enforceable, it would mean that many thousands of clearly open
source licenses could not claim themselves as open source. For example,
the &lt;a href=
http://www.affero.org/oagpl.html
&gt;
Affero version of the GPL
&lt;/a&gt; 
adds a single clause to the standard GPLv2 (section 2d), and they got
permission from the FSF to do so. But it ain't approved by the OSI, so
it ain't open source. Similarly, anybody else who adds a line
to the GPL---even one lifting restrictions---is no longer putting out
OSI-approved software.
&lt;p&gt;
Anyway, since the GPLv3, I myself have been writing software released
under the GPLv2, modified to give explicit permission granted to combine 
my work with GPLv3 software. So that means I'm no longer writing open
source software, just free software, which is OK by me.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/30000216.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>23 June 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Stats text ad copy</title>
	<content:encoded>
Academic press asked me to come up with 250 words for the back cover of my textbook 
on statistical computing. Here's what I came up with while waiting for a simulation to run.
The first and last two lines are lifted from elsewhere.
&lt;p&gt;
&lt;p&gt;
&lt;p&gt;
I saw the best minds of my generation destroyed by
              madness, starving hysterical naked,&lt;br&gt;
       angelheaded hipsters burning for the ancient heavenly
              connection to the starry dynamo in the machinery of night,&lt;br&gt;
        but the dynamo was not forthcoming because the machinery was
        not---despite its straight-faced best promises---could not bear its
        flesh to suffer the difficulties of arrows and vectors of
        immeasurable length,&lt;br&gt;
    who, when painfully waiting a week for the dynamo to spin its final
        spin and leave a single number neatly on the doorstep only to
        realize that it is the wrong number and the wrong doorstep and
        probably the wrong dynamo&lt;br&gt;
    swore that this time, finally, would be the last, the ultimate, the
        end of the numeric suffering and pulsating of waiting and
        flagellating, repeat,&lt;br&gt;
    but nevertheless succumbed to the computing beast once again,
        because, after all, the only other well-worn option was to resign,
        give up, throw oneself into the crowd where there is no one one.&lt;br&gt;
    Reader, while you are not well, I am not well, and that A follows B
        logic means that I am not well and feel the dynamo's slow churn drag
        a hole through my breast where I can no longer feel the animal heat
        of speed and knowledge and inexorable truth inexorably coming
        forward and finding itself lodged inside of me.&lt;br&gt;
    Instead, I am left with a simple longing for truth.&lt;br&gt;
    Truth is all.&lt;br&gt;
    We must love it.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/20000216.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>22 June 2007 00:00:00 +0001</pubDate>
</item>

<item>
        <title>Micronumerosity</title>
	<content:encoded>
Given that I refer to an article in today's newspaper, this one even counts as topical. Fortunately, having a
statistician in-joke as a title should keep my hit-count down.
</content:encoded>
        <link>http://fluff.info/blog/arch/00000216.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000216.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>14 June 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Your genetic information</title>
	<content:encoded>
In which I explain that I have no idea what the future will bring, even though I've had some hand in building 
some of the 
&lt;a href="http://avocado.econ.jhu.edu/modeling/"&gt;the tools that may get us there.&lt;/a&gt;. 
I hope it all works out OK.
</content:encoded>
        <link>http://fluff.info/blog/arch/00000215.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000215.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>12 June 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Here comes the ocean, and the global climate change</title>
	<content:encoded>
Here you go, Ms ZK of Canberra, Australia. A reminder that that ocean surrounding you is living and breathing.
</content:encoded>
        <link>http://fluff.info/blog/arch/00000214.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000214.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>6 June 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>GPL v3, Microsoft, Patents, and Bloat</title>
	<content:encoded>
Normally, when I write an entry on the politics of software, I tell you here why it's still interesting to those who are not involved in the issue. But this one is all about politics, sorry. Also, it's a little sad.
</content:encoded>
        <link>http://fluff.info/blog/arch/00000213.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000213.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>20 May 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Pricing information</title>
	<content:encoded>
Just make up a number and present it with a straight face. 
</content:encoded>
        <link>http://fluff.info/blog/arch/00000212.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000212.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Mon, 30 Apr 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Notes on health insurance</title>
	<content:encoded>
So I got private Blue Cross/Blue Shield health insurance. Some notes on
the process.

&lt;p&gt;
* First, let me state how much I hate the language used by insurance
providers. The OED tells us that a "Premium" is "A reward given for a
specific act or as an incentive; a prize." That is, a premium is a good
thing, and you want lots of it. But in insurance-land, a premium is a
payment you make, and you want less. On your taxes, you want all the
deductions you can score, but on insurance forms, a deductible is a
bad thing---a payment you have to make before the insurance kicks in. In
short, the language of insurance is from the point of view of the
provider, not the consumer. Because it is positive-sounding, it persists
in consumer-oriented advertising even though in that context it is
doublespeak.
&lt;p&gt;
Even the name, Blue Cross/Blue Shield, is redundant. Herein, I will
refer to them as "The Smurfs".
&lt;p&gt;

* I used
&lt;a href="https://www.carefirst.com/eSales/index.jsp"&gt;this rate calculator&lt;/a&gt;
to find my rate via various plans.
I narrowed it down to two choices: $78/month, with a $1,200 deductible,
and $293.00/month with a minimal deductible. By the first plan, I will
pay between $936 and $2,136/year for health care, depending on whether I run through
the full deductible. By the second plan, I pay $3,516/year no matter
what. As far as I can tell, the benefits are otherwise identical between
the two plans, but see below.
&lt;p&gt;
That is, consumers using the no-deductible plan pay $1,400/year more for
their insurance. Two possibilities: the pricing is based on an ex-ante
irrational fear by consumers regarding that $1,200 deductible, and they
just flinch and pay extra to avoid it, thus making pure profit for the insurer. The
other is that users are ex-post unable to commit to paying that extra
$1,200 and so pass up on $2,600 worth of services that they would have
taken up otherwise. So option one, the insurance company is making major
bank; option two, the insurance company is relying on a need to save
money sufficiently dire that people pass on medical care. I'm not
particularly perked about either option.
&lt;p&gt;

* I went with the off-the-shelf high-deductible plan. It comes with vision
coverage, and I can't imagine that they did anything special for me.
It's via a contractor, so the Smurfs tell me to call the vision contractor to
find out my benefits. Contractor tells me that they didn't get the
benefit info from the Smurfs, so I should call the Blue folks back to
find out what
my vision benefits are.  After 25 minutes on hold, the service smurf tells
me that she has no idea what my benefits are. Let me stress this:
&lt;em&gt;Nobody at either insurance provider could work out what my benefits
comprised.&lt;/em&gt; It's enough that consumers are unable to work out
their benefits under a given plan, but when even the insurers themselves have trouble
with it, you know the system is fatally complex.
&lt;p&gt;

* Two weeks after I sent my payment for my unspecified service, I got a
packet of information in the mail about what I had agreed to. I'd
estimate that it's maybe 90pp, but it's hard to tell because the
document begins with a 35 page exposition, dated July 1995, and then
the remaining is a series of, oh, 40 one or two page addenda.
[One of them points out that domestic partners of any sex are covered. Neat.]
&lt;p&gt;

So let's say I want to know if I can get allergy shots, now that the
plants are blooming. I would first check the base specification
from 1995, which is not too painful, since I'm used to legalese and it's
short enough that the lack of a table of contents is forgivable.  Cool:
just $5 per visit, no deductible applies---as of 1995. Now I have to
check the next 60 pages to find out whether anything has changed since
then. Have the smurf lawyers really been so busy that they've been unable to
give the basic agreement a revision in the last decade?  &lt;p&gt;

Take from this what you will, but I'm just amazed at how many barriers
there are to making a rational choice about insurance. It's one thing
that the pricing scheme seems to depend on consumer irrationality and
that the web site doesn't give full information, but now that I've
signed up for the contract and have full-time service smurfs looking at
my unexceptional case, I &lt;em&gt;still&lt;/em&gt; don't know what my coverage entails.&lt;p&gt;

I thought
that I was a fan of single-payer medicine just because I didn't have
insurance, but now that I am insured I want it even more.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000210.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>19 Apr 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>A tour of DC for political scientists</title>
	<content:encoded>
In high school, I lived in an apartment complex at the North end of Silver Spring (in Montgomery County), right behind the NSWC, which sometimes stood for Naval Surface Weapons Center and sometimes Naval Surface Warfare Center. It had a large outer perimiter that was primarily a golf course, and then inside of that a small campus of buildings. It always annoyed me that half of the directions I could go from my apartment were off limits. One holiday, as I was walking somewhere on the opposite side of the thing, I just gave up and walked through. Now and then, in the middle of the golf course, there'd be a little lagoon which was oddly beeping. On the way back, I just went right through the center of the campus, which was a handful of buildings along a main road, and felt like something out of the Avengers. Anyway, that's pretty far off the Capitol-to-Embassies axis that today's meanderings is about.
</content:encoded>
        <link>http://fluff.info/blog/arch/00000211.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000211.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Sat, 14 Apr 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Philosophizing from the bench</title>
	<content:encoded>
Believe it or not, this one is a survey of mathematical philosophy from 500 BCE to present.
</content:encoded>
        <link>http://fluff.info/blog/arch/00000210.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000210.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Wed, 4 Apr 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Imagined communities</title>
	<content:encoded>
Essential to your view of a product is your view of who else is using
the product. You aren't just using Coca-cola, you're joining a worldwide
network of people who are all Coca-cola users. Long-distance callers
make long-distance calls. Et cetera.  &lt;p&gt;
I could think of no better name than that for this sort of thing than
Benedict Anderson's book title, &lt;em&gt; Imagined Communities &lt;/em&gt;.
[He himself was writing about how the modern nation-state was
made possible by the printing press, which made it possible for a person
on one side of Italy, for example, to imagine a person on the other.]
It goes without saying that advertisers understand this and work hard to
make you think that Coca-cola drinkers are the sexiest, most wonderful
people alive.  &lt;p&gt;
That said, the worst possible thing you could do for your product is set
up a Web 2.0-style interactive page for it. After being amused by &lt;a href 
="http://mike.newsvine.com/_news/2007/03/26/633799-hacking-john-mccain"
&gt;this guy's pseudohack of John McCain's MySpace page&lt;/a&gt;, I thought I'd give the page itself a quick look-see.  &lt;p&gt;
i OnLY DanCe WitH Mya writes: Yeahhhhhhhhhhhh McCain rules.  &lt;p&gt;
steve writes: you should of won in 2000, i supported u then and i support u now
&lt;p&gt; Ħμŋŋ!çμŧŧ  writes: Whats up brother! keep up the good fight... The
AMERICAN fight!!!  &lt;p&gt; 
Before looking at the page, if I had to think of the average McCain
supporter, I'd think middle-America, unexciting, not too super-crazy.
But now my perception is that this guy's supporters are illiterates and
zealots. Sometimes it's better to leave these things to the imagination.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/30000209.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Mon, 2 Apr 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>An Evening with the Minister of Culture</title>
	<content:encoded>
Saw Gilberto Gil, Brazil's Minister of Culture, this weekend. If you
don't know the guy, he became  popular in the early 70s during the
Tropicalia movement, which was Brazil's response to hippie music.
It was all very perky and frequently fell into nonsense syllables. So
he's one of those legends who managed to stay relevant, at least in the
Southern Hemisphere.
&lt;p&gt;
Among those who have vaguely heard of the guy, the show was a
hot ticket. He hasn't toured since
1999, and this tour has only about ten dates. Also, he's 64, and as
folks get older, they seem to tour less, Rolling Stones notwithstanding.
&lt;p&gt;
Though his voice and playing were perfect. He played, with perfect focus  
and technique, for about two and a half hours, during which I kept
drifting of and thinking about multiplying fractions and what I'm going
to have for dinner. All of which is to say that he's still on top of his
game. The only signs of age were that it was a seated performance (An
Evening with Gilberto Gil) and the politician-perfect grey in the
temples---which went wonderfully with his dredlocks.
&lt;p&gt;
So what does a politician play for an audience? Exactly what they want, of
course, with frequent audience participation. The standard
recording of &lt;i&gt;Aquele Abraço&lt;/i&gt; is a live version that starts with a quick
dedication, for example, and he gave exactly that dedication here, which
brought about the audience's `I recognize and like this song' applause.
The recording has an audience whoop at the beginning of the second verse,
which I always found to be odd---I guess it's something that
was happening on stage during the recording. Anyway, much of the audience 
came in right on cue with the whooping.
&lt;p&gt;
As for patter, he made limited reference to politics. He dedicated one
song to "The spirit of sharing," which fits in with his various efforts
to &lt;a href="http://www.lessig.org/blog/archives/002400.shtml"&gt;
keep music sharable and remixable&lt;/a&gt;, and made a mention or two
of how being Minister of Culture is cool `cause you get to meet
interesting people. But for the most part, he seemed to just be there to
have a good time like the rest of us.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/20000209.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Thu, 22 Mar 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Web 2.1</title>
	<content:encoded>
OK, the revolution will not be televised. Will it be webcast?
</content:encoded>
        <link>http://fluff.info/blog/arch/00000209.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000209.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Sun, 18 Mar 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>The protest</title>
	<content:encoded>
What stood out for me were the pro-war counterprotesters, some of whom
explicitly and proudly used the pro-war label. They were a minority, and
homogeneous. They were mostly male, all white, mostly white-hair aged, and all had
on either black or cammo jackets. As a group, they looked more like a
biker gang than a counterprotest.
&lt;p&gt;
By contrast, the protestors were a truly diverse bunch of people. And
not college brochure diverse, but the real thing, with all races, ages,
and classes healthily represented. There were the usual Communists and
dredheads, but they were interspersed through the crowd; there were
probably as many Vets against the war.
&lt;p&gt;
Which makes me wonder all the more why the proposed troop surge. I
mean, the biker gang there probably doesn't have all that much political
clout, and the Stop the War message is no longer something just the
dirty hippies stand behind, but something Middle America votes for 
as well.
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/30000208.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Sat, 17 Mar 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Hello vs Hi</title>
	<content:encoded>I always took Hi to be a short form of Hello, but I finally asked the nice people at 
Oxford about it, and they trace them to entirely different roots. Hi, a variant of Hey (a call to attract attention), traces back to Swedish 
&lt;i&gt;hej&lt;/i&gt;, or as the OED explains: "ME. hei: cf. Du. and Ger. hei, Sw. hej, in sense 1. Cf. also HEIGH."  Meanwhile, Hello is a variant of &lt;i&gt;hallo&lt;/i&gt;, which is old German for &lt;i&gt;to fetch&lt;/i&gt;. The OED's etymology line:
"A later form of HOLLO (hollow, holloa), q.v. Cf. Ger. hallo, halloh, also OHG. halâ, holâ, emphatic imper. of halôn, holôn to fetch, used esp. in hailing a ferryman. Also written hullo(a, hillo(a, hello, from obscurity of the first syllable."
&lt;p&gt;
Policy implication: the short form of a word is typically an informal or otherwise less-preferred version of the long form. Information versus info, picture or photograph versus pic or photo, any contraction (can not versus can't), or if you're in Spain, pelicula verus peli, policia versus poli, and so on. But Hi is not a cutesy contraction of Hello, but a parallel form; both should therefore be on the same level of formality (except when calling a ferryman, when Hello is preferred).
</content:encoded>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000208.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Sat, 10 Mar 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>The best of SXSW</title>
	<content:encoded>This post took a lot longer to write than most, but it was a lot more fun.</content:encoded>
        <link>http://fluff.info/blog/arch/00000208.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000208.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Tue, 6 Mar 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Super productivity enhancer</title>
	<content:encoded>There are all sorts of web sites with tips on how to stay productive and on task, but I have found that 
checking &lt;a href="http://www.die.net/earth/peters.html"&gt;this&lt;/a&gt; page, updated every few minutes, gets me working like no other.
The terminus is coming! What have I done with my life!?</content:encoded>
        <link>http://www.die.net/earth/peters.html</link>
        <guid isPermaLink="false">http://fluff.info/blog/2000208.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Sun, 4 Mar 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Royalty check!</title>
	<content:encoded>First, a thought: if you're reading this RSS feed, why is a web page
even necessary? So I'm taking the next step in the Web 2.0 XML-enabled
craze by making RSS posts that don't have a web page attached.
&lt;p&gt;
Logistics aside, here's my news. I got my first royalty check for the
book that I spent most of 2005 writing. Hundreds upon hundreds of copies sold last year, which added up to
a total income of---$1.31. I'll try not to lose it all in one couch.</content:encoded> 
        <link>http://fluff.info/blog/</link>
        <guid isPermaLink="false">http://fluff.info/blog/1000207.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Mon, 19 Feb 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>How strict constructionism can be judicial activism</title>
	<description>If this were a topical headline-chasing blog, this entry would be over a year late.</description> 
	<content:encoded>If this were a topical headline-chasing blog, this entry would be over a year late.</content:encoded> 
        <link>http://fluff.info/blog/arch/00000207.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000207.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Fri, 16 Feb 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>My preface</title>
	<description>This entry is pretty different from the others; I begin a sentence with `and'. And no, I am not making any of it up.</description> 
	<content:encoded>This entry is pretty different from the others; I begin a sentence with `and'. And no, I am not making any of it up.</content:encoded> 
        <link>http://fluff.info/blog/arch/00000206.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000206.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Fri, 2 Feb 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Mafias and bureaucracies</title>
	<description>I don't know why I always put something in the middle of every two-part series. I guess I just like the variety. In this episode, I explain why it's great that the UN is a bloated, do-nothing bureaucracy.</description> 
	<content:encoded>I don't know why I always put something in the middle of every two-part series. I guess I just like the variety. In this episode, I explain why it's great that the UN is a bloated, do-nothing bureaucracy.</content:encoded> 
        <link>http://fluff.info/blog/arch/00000205.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/10000205.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Sat, 20 Jan 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Incremental backup with rsync</title>
	<description>Wait! Don't go yet---at least the first half may be a fun read. There's even a diagram. Next time: Mafias part II.</description> 
	<content:encoded>Wait! Don't go yet---at least the first half may be a fun read. There's even a diagram. Next time: Mafias part II.</content:encoded> 
        <link>http://fluff.info/blog/arch/00000204.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000204.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Wed, 10 Jan 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Mafias and governments</title>
	<description>Even Jimmy Carter supports the two-state solution along (more-or-less) pre-'67 borders. You don't want to piss off Jimmy Carter, do you? That'd be like kicking a puppy.</description> 
	<content:encoded>Even Jimmy Carter supports the two-state solution along (more-or-less) pre-'67 borders. You don't want to piss off Jimmy Carter, do you? That'd be like kicking a puppy.</content:encoded> 
        <link>http://fluff.info/blog/arch/00000203.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000203.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>6 Jan 2007 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Taxing value</title>
	<description>There's a lot of value in identifying what is an unanswerable question. Having identified the hard questions, make sure not to base the world's largest revenue system on them.</description> 
	<content:encoded>There's a lot of value in identifying what is an unanswerable question. Having identified the hard questions, make sure not to base the world's largest revenue system on them.</content:encoded> 
        <link>http://fluff.info/blog/arch/00000202.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000202.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>14 Dec 2006 00:00:00 +0000</pubDate>
</item>

<item>
        <title>The future of energy</title>
	<description>Finally! Haphazard pontification about the distant future.</description> 
	<content:encoded>Finally! Haphazard pontification about the distant future.</content:encoded> 
        <link>http://fluff.info/blog/arch/00000201.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000201.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>6 Dec 2006 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Navel-gazing entry II</title>
	<description>My graphomaniacal gift to the world. Happy holidays.</description> 
	<content:encoded>My graphomaniacal gift to the world. Happy holidays.</content:encoded> 
        <link>http://fluff.info/blog/arch/00000200.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000200.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>28 Nov 2006 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Is IBM evil?</title>
	<description>First, let me allay your worst fears: this is not another essay about computing. I wrote it in May 2003, and am reposting it here because it is still relevant.</description> 
	<content:encoded>First, let me allay your worst fears: this is not another essay about computing. I wrote it in May 2003, and am reposting it here because it is still relevant.</content:encoded> 
        <link>http://fluff.info/blog/arch/00000199.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000199.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>22 Nov 2006 00:00:00 +0000</pubDate>
</item>

<item>
        <title>IP Policy for Organizations</title>
	<description>As a matter of fact, yes, I have had trouble with the IT department at [name of think tank] lately.</description> 
	<content:encoded>As a matter of fact, yes, I have had trouble with the IT department at [name of think tank] lately.</content:encoded> 
        <link>http://fluff.info/blog/arch/00000198.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000198.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>14 Nov 2006 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Crime rates and PR: an ode to Baltimore</title>
	<description>I realize that this may read like an apology for Baltimore, but it's really just about the dangers of overreducing data.</description> 
	<content:encoded>I realize that this may read like an apology for Baltimore, but it's really just about the dangers of overreducing data.</content:encoded> 
        <link>http://fluff.info/blog/arch/00000197.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000197.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Mon, 6 Nov 2006 00:00:00 +0000</pubDate>
</item>

<item>
        <title>Peanut sauce</title>
	<description>This is sort of an apology for the last entry.</description> 
	<content:encoded>This is sort of an apology for the last entry.</content:encoded> 
        <link>http://fluff.info/blog/arch/00000196.htm</link>
        <guid isPermaLink="false">http://fluff.info/blog/arch/00000196.htm</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Sat, 28 Oct 2006 00:00:00 +0000</pubDate>
</item>

<item>
        <title>But wait---there's more!</title>
	<content:encoded>The older RSS entries are avaialable at the feed &lt;a href="http://fluff.info/blog/oldfluff.xml"&gt;here&lt;/a&gt;.</content:encoded> 
        <link>http://fluff.info/blog/oldfluff.xml</link>
        <guid isPermaLink="false">http://fluff.info/blog/oldfluff.xml</guid>
	<dc:creator>Eric Blair</dc:creator>
        <pubDate>Fri, 27 Oct 2006 00:00:00 +0000</pubDate>
</item>
</channel>
</rss>
