Wednesday, April 20, 2011
THE ALGORHYTHMICS
I told you that I signed up for the subscription to the NYTimes, right?
Well, I did.
Good thing.
They tell me that I have read 249 articles in the past month. Is that March? I guess so because the number doesn't change every time I read one in April.
The new Times deal that I "bought" is 15 dollars per month. Not bad. I spend a lot more than that on books and other reading materials.
Here is the nice thing though. They made a deal with heavy readers. Lincoln cars, I think it was Lincoln, would pay for the first months, all of 2011, if I signed up that day or that week. I did and they did. So I am home free for nine months.
Along with the new deal on the Times, there are some new features.
One is that they have put a "recommended" list on some pages. They have "most emailed" and, on some, "most blogged" and, now, on some pages "recommended for you"
Here we are in algorithm land again. They have such a thing on Amazon for books and Netflix for movies. The Netflix even gave a million dollar award for anyone who could come up with a better "recommender".
Both of them are for the shit.
The worse is Amazon. They list books I would never consider and sometimes, just to be cranky, will list books that I have already ordered! Black and white.
Netflix, I must admit, is a bit better. But not much. Many of the recommended films would not please me at all. I have heard about them, read a review or just got the plot and I don't want to see them.
The NYTimes recommender is better though. Quite surprising. Many of the articles are ones that I have not yet seen in the paper. And, for the most part, they do, in fact, interest me. They have some obvious hooks. Gay issues for example. But others are quite mysterious. When I click through and see the piece they have recommended, it is surprising how many hit the spot. The r-spot. Reading.
Some are lemons. Maureen Dowd for example. Not preferred but usually on the list.
The other thing is that their list does not repeat. There are no articles that I have already read. None.
At first I was quite skeptical. Change. No. But now, I am drawn to it and curious. Very good. The list is singing to me.
OH. Dijkstra's algorithm has nothing to do with newspapers. Clickon to see.
It is just pretty. I was looking for an algorithm illustration and voila!
Dijkstra's Algorithm solves the single-source shortest path problem in weighted graphs. Here we show it running on a planar graph whose edge weights are proportional to the distance between the vertices in the drawing -- thus the weight of an edge is equal to its visible length.
I don't think that is going to get on my recommended list of articles.
