Home | Introduction | Newsletters | Programs | Feedback
|
Home | Introduction | Newsletters | Programs | Feedback
|
|
WHAT'S NEWFebruary 4, 2010:
Word squares are square letter arrays with rows containing words
which match words formed by the corresponding columns. Here is a
Word Square Search program which
searches a 62,000 word dictionary to create word squares based on an initial
word provided by the user. At left is one of 149 solutions found for
the word "WORD". February 1, 2010: A couple of times each year I get a request for help from a student requesting help in some project using my Reaction Times programs. This year it was a graduate student in at a Medical School in Ecuador whose thesis topic is to test the effect of sleep deprivation on interns as result of the long hours they put in. Her problem turned out to be decimal separator confusion because the PC she was using happened to have the Window's Locale set for a European country. Changing back to Ecuador set it from comma back to dot and allowed her to analyze her data. While investigating, I discovered and fixed a problem with the ResponseStats statistical analysis program that kept a newly loaded response statistics file from being made available. Reaction Times programs including ResponseStats Version 3.1 was uploaded today to fix the problem. January 17, 2010: Consider the sequence formed if, starting with any positive integer, we form a sequence where for the sequence ending with N, then next entry is N/2 if N is even and 3N+1 if N is odd. Starting with 3 for example, the sequence is [3, 10, 5, 16, 8, 4, 2, 1]. The unproven conjecture (the "Collatz" conjecture) is that any such sequence will eventually degenerate to 1. Here's a beginner's level program, ThreeNPlus1, that evaluates all sequences within a given range and reports the longest sequence found. January 12, 2010: DFF Newsletter #55 was sent yesterday using the latest version of MailList King, a mailing list management program from Xequte Software. The program is written in Delphi and contains all the features needed to simplify newsletter sending such as automatic subscribing/unsubscribing, limiting the sending rate to comply with host server limits, processing undeliverable email responses to clean the list, etc. The $99 personal edition does everything I need and is recommended if you ever have the need.
A university student recently took on a Java version of the
MasterMind game as a school project. He wrote asking about my "Smarter
than you" intelligence level which can be applied when the program is doing the
guessing about the secret pattern. The program has a Verbose mode which
was supposed to help debugging and explaining the search steps, but the messages
were being cleared after each computer guess. MasterMind
Version 2.1 corrects this and hopefully will help him and others understand
the mini-max guessing procedure used at the smartest level. January 9, 2010:
"Instant Insanity" is a variation of an older cube arrangement puzzle and one of a large family of similar puzzles. In this one, we have four cubes with one of four colors on each face of each cube. The objective is to stack the cubes so each column of faces has all four colors. A viewer recently asked for help in developing a Delphi solver
version.
Most of the online literature describes a graph search algorithm which can be
applied to find solutions
with pencil and paper. With a faster but dumber computer, it's easiest
just to check all 41,472 possible arrangements looking for
solutions. Instant Insanity
Version 1.0 posted today does just that with a few sample or user defined
cube sets.
While playing the "Wii Play" version of the Four-In-A-Row game
with a grandson over
the holidays, I used my Delphi version of the game to advise me on moves
to make. I was embarrassed to find that the program's random
moves "enhancement" frequently allowed Chris to win by placing 4 adjacent token in his first
4 moves - the initial random moves feature assumed that there would be no winner that soon.
Four In A Row Version 2.2 posted today
fixes the problem by limiting initial random moves to 2 or 3 turns. January 2, 2010: Happy New Year! We'll kick off the new year with a program that I wrote a year or so ago but never got around to posting. Accordion Solitaire is a simple solitaire card game that is difficult to win; perhaps one win in 100 games. This version allows user and program play and implements a "sweeper" strategy of play which can increase the odds of winning to perhaps 10% December 28, 2009: Here's wishing everyone a happy holiday season. Christmas here was not the best ever here along the Blue Ridge. Ice and wind on Christmas eve knocked out our electricity for 3 days. It was a small reminder of what it was like in the olden days without power. Believe me, with power is better, (especially for a computer programmer J). It's back today, so I'll get one more program posted before the year ends. Circle From 3 Points, Version 2 is a extension of our program which explores the math defining the circle passing through any 3 given points (unless they happen to fall on the same straight line). The original version used mouse clicks to define the points. Version 2 allows users to input three floating point values of arbitrary scale. I needed that to help solve a specific puzzle from one of the puzzle books received as a Christmas gift. December 18, 2009: Safely back home from our first (and last) repositioning cruise. The first week, visiting Barcelona, Cannes, Florence, Rome, Corsica, and Malaga were great. The second week, crossing the Atlantic, not so much. We're not into casinos, disco, game shows, rock wall climbing, or spa treatments, so that left plenty of time for reading. Of course, being retired, I have plenty of time for reading any week I choose. Seas were rough a couple of days with 30+ foot waves and quartering head winds of 90 to 100 mph. I wish I had had an accelerometer to measure the G forces imparted when a 93,000 ton ship descending collides 180 degrees out of phase with with a 30 foot wave. Not enough to shake one out of bed, but enough to prevent sleep for sure. No sea sickness for us, but I was unlucky enough to pick up an unpleasant intestinal virus two days from port from which I'm just recovering. So, we've chalked it up to "live and learn" and move on. There was only one programming request that seemed to require action when we returned home. The Traveling Salesman Problem program posted in 2002 allows users to trace paths visiting all of a random set of cities and compare their distance to the shortest distance calculated by the program. There are options for "closed" or "open" routes depending on whether we end up at the city we started from. A student using the program for a project has a requirement to begin the open route at a specified city but my original version always started at the western-most city. Traveling Salesman Problem, Version 2 posted today corrects that oversight and starts open routes at the first city selected by the user. November 24, 2009: We're off to spend Thanksgiving with family and then taking a two week repositioning cruise hitting a few western Mediterranean sites in Italy, France and Spain and the across the big waters to New York City. I'm not sure about Internet availability on board, so if email responses are delayed, be patient; I'll catch up when we get back in mid-December. November 19, 2009: While working on some upcoming changes to our Sound generator program, I discovered that our Metronome program posted several years ago, does not like to play the "click" wav files when run under Vista or Windows 7. Metronome Version 2 adds two additional "click" sound files to address the problem. Click1Vista and Click2Vista files extend the duration of each click to be audible. Vista made major changes to audio processing including moving many functions to run as a service. For some reason, the minimum played sound now seems to be 31 milliseconds. November 15, 2009: This may be a project first, coding KenKen Version 2, with the "Solve it" button, turned out not to take as long as I though it would! I included one additional sample test puzzle also; a 9x9 hard version that I wouldn't care to solve by hand. November 13, 2009:
Our AARP Bulletin this month contains this puzzle, shown here
partially solved. It's an example of a puzzle game called KenKen; similar
to Sudoku except that the board is divided into "cages" with an operation and a
target value associated with each cage. The digits 1 to 5 must be placed
on the board so that they can be connected with the given operator to form the
given target value in every cage. In addition, no digit can be repeated in any column or
row. My KenKen Version 1 program is a
simple playable test bed to help investigate algorithms for automatically
solving this type of puzzle "just for fun".
At the puzzle owner's request I am informing you
that KenKen® is a registered trademark of Nextoy, LLC and that this site is not
affiliated with Nextoy, LLC or KenKen Puzzle LLC. November 2, 2009: This card probability question was posted, and partially answered, three years ago. Given two card rank values, what is the probability that there are one or more occurrences of the two values being adjacent or with only one intervening card in a well shuffled standard 52 card deck? The original version used generated random card decks to answer the question experimentally by checking lots of decks and counting what proportion of them met the card placement conditions. I had asked then about an analytical solution - it took three years but programmer and fellow puzzleist Mark Rickert came with the key concept (and sample C code) that allows exact solutions. Card Probability Version 2 has my implementation of the algorithm and even throws in Mark's C code translated to Delphi as a bonus!
The WHAT'S NEW Archives(Click to expand)
|
[Feedback] [Newsletters (subscribe/view)] [About me]Copyright © 2000-2009, Gary Darby
All rights reserved.
|
