We can find these historic details because links have at least a provisional permanence to them. They are, literally, paths to locations. Thanks to those, we can document the history we make, and learn from it as well.

- Doc Searls on Why Durable Links Matter

The Internet in general, and the web specifically, have given people an unprecedented power to both record and erase history.

Last week I mentioned that Google Fiber was coming to Provo. This process would be accelerated because they were purchasing the existing fiber network, iProvo, instead of having to built out a completely new one.

Turns out the sale price of iProvo fiber network to Google is $1. It should come as no surprise that some people are less than thrilled about this price.

The original cost to build out the iProvo fiber network is reported at $39 million. To fund this the city of Provo issued bonds. Those bonds have not yet been paid in full. The $1 in revenue from the sale isn’t going to pay off the bonds. In an effort to help pay off the bonds in 2011 a surcharge was added to the utility bills of every household in Provo.

While most people seem happy to have Google coming in to upgrade and run the fiber network in Provo, others are concerned that the sale of the network does nothing to address the remaining millions of dollars in bond obligations.

Both sides are benefiting from this, but the $1 sale does feel a bit odd. I’d assume the existing fiber network has some value, less than the original $39 million and more than $1. Perhaps something on the order of $5 to $10 million would have been reasonable. Google would still have been given a great deal on existing infrastructure, and the city of Provo would have some money to pay down the bonds faster.

In the negotiations for this I’m sure Google held all the cards. The city of Provo was likely happy to bend over backwards and into a pretzel to make sure Google didn’t walk away from the deal.

The Salt Lake City WordPress Meetup is moving locations, back to Draper.

The next meeting is tomorrow night ( April 24th ), where Matt Jones will be talking about e-commerce with WordPress.

The point is, some days you have to drop down into tcpdump and pcap files to get the truth.

- Steve Souders while tracking down how much HTML5 video data iOS actually receives.

Turns out server logs were giving an incorrect picture of the amount of data that was being transmitted. It took watching individual packets to track down what was really happening. Go read the post for the odd twist at the end.

Reading this reminded me of using Wireshark to track down multiple packet issues in HTTP POST requests from XMLHttpRequest.

The big announcement in Utah yesterday was that Google Fiber is coming to Provo. Instead of building out completely new infrastructure, Google will be purchasing the existing fiber network in Provo. From there they will be upgrading it to be in line with other Google Fiber installs.

I live about a 30 minute drive north of Provo, so this won’t make it to my home. Though Comcast recently upped my connection to 50Mbps down, so my current connection isn’t horrible. I wonder if Comcast knew this announcement from Google was coming and wanted to preempt some of the complaints from users.

The other thing that I wonder about is the NSA datacenter a short hop north of Provo. Was this a motivating factor on the either side? No idea. Would be nice to know if Google Fiber in Provo ends up peering with the NSA data center.

Speaking of peering, there are a few others in the area that would probably love to have a network peering arrangement with Google Fiber in Provo. C7 has a data center in Bluffdale ( north of Provo and not far from the NSA data center ) that apparently has a hefty number Twitter servers. Bluehost has a data center in Provo. Those are the two I could think of off the top of my head, there are data centers in the area who would be interested as well.

This is just the beginning of Google Fiber in Provo, and even though I won’t have access to it, I’ll be watching closely to see how it develops.

Last week I mentioned Percona MySQL was switching to jemalloc. This week mysqlperformanceblog.com published benchmarks comparing MySQL 5.5.30, Percona MySQL 5.5.30, and MySQL 5.6.10.

As far as I can tell the Dell & HP websites have been largely designed to be horrible, in hopes of routing you to a sales person. I tried to fight this, but it was pointless. Phone someone at one of those companies and save yourself several hours.

- Paul Reinheimer, on buying a new server

In nearly all cases I’d rather complete a purchase with an easy to use website instead of calling a sales person.

I’ve previously mentioned that the last time the US federal debt went down instead of up was 1957. This came up in a conversation again recently and I started thinking about the percentage change in the federal debt for each year. I looked at the period from 1958 to 2012, what I would call the modern era of federal debt.

I went back to the federal debt records, they go from 1790 to today. Dumped that data into a spreadsheet, then calculated the percentage change for each year from 1958 – 2012. I used the all actual numbers, I didn’t attempt to normalize them to 2012 dollars since each year is just a comparison of the year before.

Copying that into Google Charts Javascript code generates this graph:

Percentage Change in Federal Debt, 1958-2012

google.load('visualization', '1', {packages: ['corechart']});
function drawVisualization() { // Create and populate the data table. var data = google.visualization.arrayToDataTable([ ['Year', '% Change'], ['1958-06-30', 2.15], ['1959-06-30', 3.03], ['1960-06-30', 0.57], ['1961-06-30', 0.92], ['1962-06-30', 3.19], ['1963-06-30', 2.57], ['1964-06-30', 1.91], ['1965-06-30', 1.78], ['1966-06-30', 0.83], ['1967-06-30', 1.97], ['1968-06-30', 6.55], ['1969-06-30', 1.77], ['1970-06-30', 4.86], ['1971-06-30', 7.34], ['1972-06-30', 7.32], ['1973-06-30', 7.23], ['1974-06-30', 3.69], ['1975-06-30', 12.24], ['1976-06-30', 16.36], ['1977-09-30', 12.64], ['1978-09-30', 10.40], ['1979-09-30', 7.13], ['1980-09-30', 9.82], ['1981-09-30', 9.93], ['1982-09-30', 14.45], ['1983-09-30', 20.59], ['1984-09-30', 14.16], ['1985-09-30', 15.95], ['1986-09-30', 16.58], ['1987-09-30', 10.59], ['1988-09-30', 10.72], ['1989-09-29', 9.80], ['1990-09-28', 13.15], ['1991-09-30', 13.36], ['1992-09-30', 10.89], ['1993-09-30', 8.53], ['1994-09-30', 6.38], ['1995-09-29', 5.99], ['1996-09-30', 5.04], ['1997-09-30', 3.60], ['1998-09-30', 2.09], ['1999-09-30', 2.35], ['2000-09-30', 0.32], ['2001-09-30', 2.35], ['2002-09-30', 7.25], ['2003-09-30', 8.91], ['2004-09-30', 8.78], ['2005-09-30', 7.50], ['2006-09-30', 7.24], ['2007-09-30', 5.89], ['2008-09-30', 11.29], ['2009-09-30', 18.80], ['2010-09-30', 13.87], ['2011-09-30', 9.06], ['2012-09-30', 8.63] ]); // Create and draw the visualization. new google.visualization.LineChart(document.getElementById('visualization')).draw(data, { width: 750, height: 400 } ); } google.setOnLoadCallback(drawVisualization);

You can mouse over each point to see the exact value for any given year.

A few observations on this graph. First, these are all increases, no negative numbers. Next, you’ll easily notice that the year with the largest percentage increase was 1983, at 20.5%. The next highest increase was 2009, at 18.8%.

The highs weren’t what most surprised me in this data, it was the lows. In 2000 the percentage increase in federal debt from 1999 was only 0.32%. That is the smallest percentage increase since 1957. I knew that we’d reduced the yearly budget deficits during that period, but I had no idea how close we were to actually reducing the total federal debt.

For 2000 the dollar increase of the federal debt was $17,907,308,271.43, or roughly $18 billion ( yes, billion with a b ). That sounds like a significant amount, but compared to other years is very small. To give you some idea of how small, in 2001 the percentage increase was 2.35%, the dollar amount was $133,285,202,313.20, more than 7 times the amount in 2000.

I wanted to know more about what it would have taken to actually reduce the total debt in 2000. The 2000 budget request expenditures were listed as $1,765 billion. To keep things simple I’m looking at this from the point of view of how much in spending cuts we would have needed in the 2000 budget to clear $18 billion that could have been used to reduce the total federal debt.

A little bit of math shows that $18 billion was 1.02% of the $1,765 billion in spending listed in the proposed 2000 federal budget.

Lets say we could find 1.1% in reductions in the budget to go towards paying down the debt. That would get us $19.4 billion. The first $18 billion would go to cover the increase that happened in the debt, leaving a $1.4 billion reduction in the total federal debt. That is only a tiny fraction of the $5,674 billion total debt in 2000, but it would have been the first time in more than 40 years that it went down instead of up.

All totally hypothetical and vastly over simplified, but this gave me a much clearer picture of the history of the federal debt.