Ascription is an Anathema to any Enthusiasm

The book I was looking for was not on the shelf, but this was a few books down. The original copyright is 1904. Prudence Hyde renewed the copyright in 1938 by Prudence Hyde. The five great philosophies?

• Epicurean Pursuit of Pleasure
• Stoic Self-Control by Law
• Platonic Subordination of Lower to Higher
• Aristotelian Sense of Proportion
• The Christian Spirt of Love

But then I’m sure you knew that. The author was William De Witt Hyde, a president of Bowdoin College. (Amazon)

Chris Anderson points out an interesting power point presentation from a company called DVD stations ptt/pdf Their business is DVD rental kiosks. The kiosks print DVDs on demand and have really fast connection back to the mother ship. Like all attacks on a distribution channel the upside comes from unlocking value that couldn’t get thru the channel beforehand. In this case the long tail of content. There are some nice charts showing how much value there is out there. One curiousity, they have a little budge in their revenue for movies that are around 9-10 years old, what’s up with that?

I was interested in two charts. The first one shows a portion of their sales pipeline, it shows that a few customers account for a large chunk of their revenue. You gotta love the labels marketing people put on the folks in various stages of their pipeline.

The second chart shows which channels distribute thru and how much revenue comes out of each. Since I don’t watch cable TV I was surprised how much is premium cable and video-on-demand. This is pie is of course a slice of yet larger pies, i.e. the entertainment pie. In the future the big slice is going to be neighborhood puppet theaters - you heard it here first!

Of course I suspect that both these pie charts are just power-law curves, but ironically they don’t display that. They want to be an elite. For example, in the customer catagory space that leads to inevitably into pricing discussions and then into the new dark ages of DRM content.

Fresh version at http://www.cozy.org/silverspoon/.

If you compile up the test program big then you can use it to glean out a list of items that occur (aprox) a certain number of times. For example if “yesterday -ip” dumps the log of yesterday’s web server traffic with just the IP addresses, then: “yesterday -ip | big -c 50 > bl50” will show you the IP addresses that dropped by 50 or more times. Or you can then do “yesterday -ip | big -c 40 -b bl50” to see the ones that dropped by from 40 to 50 times. The “-b” switch stands for black list and all items in that set are discarded.

htp-0.0.1.tgz is a tar file containing some code I hacked together this afternoon implementing bloom counters. Minimal doc and source views.

It has got lots of bugs. While is also an implementatiof of bloom filters it’s totally untested. There is one test executable named big. It is unix pipe fitting. Given a stream of lines it prints out all the lines appear more N times.

The code depends on APR.

The mnemonic HTP was for Hot Spot, but only late in the day did I realize that’s too much like HTTP - oops. That’s change if I get back to this.

No promises, no warrenty, lousy license.

It sure looks to me like there is good, fast, linear, bounded storage, way to encode a stream of data on the fly that uses the varient on bloom filters outlined yesterday to select the compression codes on the fly.