The National Safety Council (NSC) estimates that 25% of all crashes are caused by cell phones. The National Highway Traffic Safety Administration (NHTSA) estimates that driver distraction from all sources (including cell phones) contributes to 19 percent of all crashes (2008). Even then, this is not necessarily the cause of the crash.

Both cannot be true, someone here is wrong. The NSC report is merely a statistical extrapolation of what might have happened, heavily based on averages. More real world data to corroborate their numbers is needed. The NHTSA has the opposite problem. Their numbers are based on real world reports. These suffer from reporting errors, often because those involved in a crash might not report being distracted while driving.

After playing with IPv6 extensively over the weekend, I discovered a number of oddities about the protocol.

First, autoconfiguration only works with /64 subnets. Anything more or less, even if the router advertises the prefix, clients will not add the prefix to its own interface. This seems like it can problematic as users cannot easily break down these prefixes into smaller subnets. This could be a key area for using DHCPv6 as a replacement.

Second, autoconfiguration will not work if the computer is configured as a router for IP forwarding. This means that additional work for defining routes needs to be done even if you only plan on using the default gateway on a network. This can either be in the form of static routes or a full blown routing protocol.

Third, 2002::/16 (IPv4 transition addresses) doesn't count as a real IPv6 address when source address selection occurs. Instead it is considered a separate scope, similar to an IPv4 address. This means when websites have both a 2001::/16 IPv6 address and a IPv4 address, the IPv4 address is used by default. If the website has a 2002::/16 IPv6 address and a IPv4 address the 2002::/16 IPv6 address is used to connect. This seems very odd and inconsistent and can lead to confusion.

If Rep. Susan Davis' (D-CA) office can't plan a town hall correctly I don't know how we expect her to do any better with larger government programs like health care. The parking lot wasn't large enough to accommodate over 1,000 citizens, some constituents ended up walking several blocks into order to park their cars only to get denied entry. The facility was full with standing room only for nearly 1,000 concerned citizens.

In her defense the question and answer session was handled rather well. Every got to put their name "in a hat" and they were randomly drawn through the Q&A session. Each person had one minute and thirty seconds to make a statement or ask a question. Susan Davis would then respond, usually with a shorter answer than the question given. Also to her credit, even though the event was only scheduled from 1 to 2 PM she took questions well past 2 until about 2:30.

Some highlights:

To answer the question before it was asked, Rep. Davis stated, "I just want you to know that, yes, I have read the health bill, it is over 1,000 pages..." (14:35)

"We rank 45th in infant mortality rates, behind Cuba," said Davis (18:10) before the crowd uproared.

In response to a question about hate crime legislation (HR 1913) only protecting certain groups, she stated, "I think that when you are discriminating against one group, you're really discriminating against all groups." (34:50) That sounds like an oxymoron to me, if you discriminate against all groups then its not discrimination is it? That's just "the way it is."

One person asked, "I know that you are supposed to represent your constituents and I happened to look at the poll you have on your website and its says that 'the health care reform should include a government managed public option,' 85% of respondents said no." (1:10:05) After an uproar, Davis responded, "I would like to see a public option...I think you have to stick to your principles." (1:13:30)

[UPDATE: 09-03-2009]

Rep. Susan Davis' office also provided handouts with some "facts" about health care. The exact straight line increase chart called "The Cost of Doing Nothing" struck me as odd so I did some research. I spent a few hours looking at the Kaiser Family Foundation's website trying to find this $1,800 increase every year through 2023 since that is what the slide cited as the source. I couldn't find it. I couldn't find any Kaiser Family Foundation research that projects costs into the future. I could find a figure close to the $12,500 starting point for 2008, so at least that checks out. So I called Susan Davis' office and inquired about what publication I could find this statistic in. After a couple of days I got a response that mentioned a few sources.

[UPDATE: 09-23-2009]

I got another call from Susan Davis' office specifically citing a New York Times Economics blog as the source for the $1,800/year figure. This is sort of true since it suggests $18,000 over ten years starting in 2010, but this is more likely to be distributed unevenly, less than $1,800 in the early years and more in the later part of the decade due to inflation. I do not know why the lower KFF estimates were used except to mislead. When using the starting point of $12,500, $1,800 is a 14.4% increase, compared to 10% when starting at $18,000.

Analysis of blog post shows that this number it an expert opinion and not a fact, it has not been peer reviewed and assumes the current defunding trend for existing government health programs continues. In fact when considering an inflation rate on average of 3% this is exactly canceled by 3% wage increases as stated on the blog. That only leaves a true cost increase of about 4% per year. Much of this 4% is not due to an increased cost of care, it is because government programs are paying out less than in the past and service providers need to recover those costs. A quick table of costs at a 7% annual growth rate.

2010 - $18,000
2011 - $19,260
2012 - $20,608
2013 - $22,050
2014 - $23,594
2015 - $25,245
2016 - $27,013
2017 - $28,904
2018 - $30,927
2019 - $33,092
2020 - $35,408

Another government program is NOT a solution based on testimony by John M. Pickering of Milliman, Inc., to the House Committee on Ways and Means. Analysis of this testimony is UNBELIEVABLE. What it says is that in 2007, $1,788 (10.7%) of the total health care cost for a family of four went to COVERING THE COSTS OF THOSE ON MEDICAID AND MEDICARE and the uninsured. So because the government does not pay out market rate for services rendered, its costs the average family $1,788 every year. If the existing government programs paid out at market rates, costs for these families would drop by $1,788. The additional cost for government programs would likely be spread across the tax base because this is a zero sum game.

Full Town Hall Audio (MP3)