Tuesday, May 15, 2012

Drug Testing Aid Recipients - A Novel Idea

Seems like a good idea, right?  Let's make sure we're not doling out cash to people so they can buy meth and cocaine.  Seems to make sense to me.  That is until you do some actual thinking, rather than knee-jerk emotional reaction logic.

Let's start with cost.  It has repeatedly been shown that drug testing is expensive.  On average, a single test costs $42 per sample.  That is exclusive of the personnel cost to administer, records management, verifying positives (to reduce false positives), and cross-referencing to "legitimate" prescriptions and legal substances (poppy seeds) on an individual, case-by-case basis.  Detecting illegal drug use is an expensive operation, no doubt.  There were 2.5 million welfare recipients in Michigan in 2011.  A little math and this is insanely expensive (and consequently lucrative to test providers).

But look on the bright side, it would certainly create a hell of a lot of jobs.

Oh, but we'll save a ton of money by cutting off all those druggies, right?  Surely that should pay for the administration and related costs of testing.

So how much would we save in welfare payments?  The average welfare cash assistance is around $900 for a family of 4 plus another $500 in food stamps per month.  So figure (high) an average of $1,500 per month per potential druggie.  Studies have shown that there is no statistically significant difference in drug use between welfare recipients and the general population, which runs around 3%.  Taking our Michigan example, that means we could cut off 75,000 druggies from their welfare payments (presumably so they can be forced to beg or steal in order to eat).  That would save a whopping $112,500,000 per month.

Monthly testing should only cost about $1,500,000,000 per month (based on the lower "per incident" cost of two studies, one in the private sector, one by a congressional committee)

What a deal!  Government at it's finest.  Spend $1.5b to save $0.112b.

Shall I go on?

Sunday, October 9, 2011

iPad2 - Smashed!

It finally happened.

About 11:00pm last night, my lovely daughter managed to arrange a high-speed liaison between my iPad 2 and the tile floor. Oh joy.

Smashed beyond all reapair, I took my sorry excuse for an iPad back to the Apple store. Tail between my legs, the associate politely informed me of the benefits of Apple Care warranty and the rugged Otter Box cases available for my precious gadget. 10 minutes later, I'm sitting at the Genius Bar, awaiting my fate.

In classic Apple style, my appointed Genius gave me an awesome deal on a replacement, and had it in my hands within 30 minutes of me showing up at the store. No, it wasn't free, but it's a bit much to expect a free replacement after declining the warranty and then dropping it on the floor.

Just one more reason I love Apple products (yeah, call me a fan-boy or whatever). Not only are the products outstanding, but when it comes to customer service, they are second-to-none.

Thanks again, Apple. This time for dulling the pain of a rather unpleasant event.

Kudos to the Apple store in Boulder and my Genius, Max.

Thursday, October 6, 2011

Steve Jobs - 1955-2011

Few people have changed the course of human history as Steve Jobs has. May he forever be remembered as the most visionary technology innovator of our time.

Much as the work of Thomas Edison and Henry Ford ushered in the industrial revolution, Steve Jobs and his remarkable vision and leadership has largely defined the information revolution.

Henry Ford put power and promise of the automobile within reach of the common man. Steve Jobs put the power of a nearly infinite universe of information in our pockets.

Thomas Edison made the technological breakthroughs that enable modern life accessible to the masses. Steve Jobs envisioned and created products that enable unprecedented communications, access to information, and in a very real way he "redefined" modern life on Earth.

The world has suffered a great loss.

Monday, September 12, 2011

Magento on Rackspace Cloud - Update

At the beginning of the year, I posted about my experience running Magento on the Rackspace cloud infrastructure.  The gist of that post was that Magento does in fact run quite well, as long as you use a minimum of the 2GB instance provided by Rackspace.  (Original Article Here)

Since then, I have been migrating several sites to this infrastructure.  Initially, I figured it would be easy to put 5 or even 10 sites on one instance, and scale it up to 8GB or 16GB or whatever was needed.  Vertical scale with a single cloud server appeared to be a quick and simple deployment plan.  Unfortunately, this all came crashing down on me when I went to add the sixth site.

Rackspace kindly informed me that they could only assign a total of 5 IP addresses to a single CloudServer instance.  This really makes no sense to me, since I can get 5 more just by adding another server, but that's their policy anyway.

In case you didn't know, every site that has its own SSL certificate requires its own dedicated IP address.  This includes all of the Magento sites that we currently manage at 4word systems.

Forced to deal with the reality of spinning up several 2GB+ instances to support a larger number of sites, I got to thinking about my original strategy.  After all, to keep on the planned path would turn out to be expensive as hell.

What if I offloaded the database activity to a separate server?  How might that impact the performance / hardware requirements for the web server?  How big would the DB server need to be?

Through a good amount of trial-and-error I have discovered the following (so far).

1:  The database server can be surprisingly small if properly tuned.  MySQL database tuning and maintenance can be a tricky thing, but currently I am supporting around 10 Magento databases on a 1GB CloudServer instance with no performance problems whatsoever.  I will again mention that this absolutely requires a well-tuned MySQL installation.

2:  The web server size requirement also drops considerably.  Previously, I had a single Magento site that was so busy most of the time, I needed to upgrade from 2GB to 4GB to prevent heavy swapping to disk.  After offloading the database to the new 1GB instance, I was able to reduce the web instance from 4GB to 1GB.  This is a remarkable improvement.  2 servers at 1GB is half the cost of a single 4GB instance, and the 1GB database server is able to service at least 9 more websites.

The long and short of it is this:  If you plan to host Magento on CloudServers, do yourself (and your budget) a favour.  Split the database on to a different server.  Even if you are only managing a single site, you will get more performance for less money any way you cut it.

Friday, September 9, 2011

Mass Transit in America - Time for a New Dawn

The history of this country is rich with ambitious, nationwide transportation projects.  One might argue that transportation has enabled much of the growth that has fuelled the rise of the US as a global superpower.  Transporting materials, supplies, finished goods and people across the nation with speed and efficiency is a hallmark of American accomplishment.

Sadly, with the completion of the interstate highway system in 1992 (at a cost of 4x the original estimate and 3x longer than planned) transportation has taken a vacation when it comes to public policy.

The World Has Changed
A lot has changed since 1956, when the Federal Aid Highway Act was written.  Smog, congestion, oil-dependency and highway fatalities are a fact of life now for virtually every American.  While the highway system accomplished it's grand objective, the consequences can not be ignored.  Each day, tens of millions of people wake up, get in their cars, and waste an average of an hour each day driving to and from their place of employment.  Risking life and limb, the gainfully employed venture out onto the interstate highway system so they can afford to pay for the vehicles and fuel that make it possible for them to do the same thing the next day.

Count the Cost
With the average price of a new vehicle hovering around $30,000, few Americans can afford to directly purchase a new car.  Instead they turn to the benevolent finance companies (aka banks) who will ensure that the actual cost will nearly double.  Given that a car typically will be out of warranty within 3 years and nearly worthless after 10, it's not unusual for the capital cost of a vehicle to come to $7,000 - $8,000 per year.

Add on the compulsory insurance coverage for another $750 or so per year.

With the average round-trip commute at 32 miles per day, an average fuel economy of 25 MPG, we add another $1,200 - $1,300 per year in fuel alone.

Add to that the cost of regular scheduled maintenance, tires, brakes and repairs, it's easy to see that operating a car solely for the purpose of getting to work can easily run $9,000 - $11,000 per year.

News Flash: For nearly 1/2 of American households, that represents 20% or more of total pre-tax income.  When you look at after-tax income, vehicle expenses can easily exceed 30% of the average family's take-home pay. 

I'm convinced most people don't see that.  We've been conditioned not to.  Advertising works, people.

Oh, and that's before we even consider the more than 250 hours each year lost by the average commuter on their way to and from work.  At an average rate of $25 / hour, there's another $6250 down the drain.

It's time for change.  I thought that's why we hired this guy for President. . .