If you see something like that:

$ git fsck --lost-found
dangling commit 9dbd85d427b32773aee016efc7912f9d1c39aac4
dangling commit 604948146be1bdc2181962543cca039c18cc06b4

You might want to get rid of those dangling commits, in order to accomplish it, do following:

git reflog expire --expire=now --all #Dangerous command, it will be very difficult to recover stuff later on, if you execute it
git gc --prune=now

In order to hide a folder in Finder, open Terminal and write following:

setfile -a V /path/to/folder

Or

chflags hidden /path/to/folder

To unhide:

setfile -a v /path/to/folder

Or

chflags nohidden /path/to/folder

1) install homebrew

$ ruby -e "$(/usr/bin/curl -fksSL https://raw.github.com/mxcl/homebrew/master/Library/Contributions/install_homebrew.rb)"

2) Install Xcode from App Store (4.3.2)

3) Install Command Line Tools for Xcode from https://developer.apple.com/downloads/index.action

4) Install MySQL

$ brew install mysql
$ unset TMPDIR
$ mysql_install_db --verbose --user="$(whoami)" --basedir="$(brew --prefix mysql)" --datadir=/usr/local/var/mysql --tmpdir=/tmp

By default, data files and the error log file ([HOSTNAME].err) will be located at:
/usr/local/var/mysql/

To start o stop the server:
$ mysql.server start
$ mysql.server stop

If you get «InnoDB: Operating system error number 13 in a file operation.» error when starting the server you forgot to run mysql_install_db.
«FATAL ERROR: Could not find ./bin/my_print_defaults» error means that you are trying to run mysql_install_db without all the --basedir option.
Optionally, we can place MySQL configuration file to /etc/my.cnf

5) Install git
$ brew install git

6) Install RVM

$ bash -s master < <(curl -s https://raw.github.com/wayneeseguin/rvm/master/binscripts/rvm-installer)

7) Install Ruby

rvm install ruby-1.9.3-p125

To use this version of ruby:
$ rvm use ruby-1.9.3-p125

To set it as default version:
$ rvm --default ruby-1.9.3-p125

We can verify that ruby-1.9.3-p125 was installed by running:
$ ruby -v

8) Install Rails:

gem install rails

Hello all!

Recently I've purchased Macbook Air mid 2011, I very like it so far.
However, I've noticed when I'm watching movies or YouTube videos, the fan is spinning really fast, iStats shows me something like about 6400 (!) RPM.
I've tried to use fan controlling apps like smcFanSpeed or Fan Control.
However, both of those apps doesn't allow you to decrease fan speed (probably it's designed that way in order not to burn the laptop in inexperienced user hands)

I'm kinda experienced user, so I took a risk and wrote simple script that allows you to set fan speed manually.
The instructions are pretty easy:
1) Download it setFanSpeed2011.tgz
2) Open the Terminal, go to folder where the file was downloaded
3) Extract it: tar xf setFanSpeed2011.tgz
4) Go to inside extracted folder: cd setFanSpeed2011
5) Run it: sudo ./setFanSpeed XXXX (Where XXXX should be needed RPM)

P.S. When I'm gaming or watching movies, I found that 3600 RPM work for me pretty well, the CPU stays on ~174F (~79C) temperature.
The same temperature as when fan was spinning at 6200 (!) RPM, but now it's not such noisy!
P.P.S Created and tested on OSX Lion 10.7.1

Disclaimer,
Note, I'm not taking any responsibilities if you will burn your laptop by using my tool! Decreasing the fan speed is very dangerous!
Since, if you don't watch the CPU temperature it might be overheated!

In this article we will discuss a solution to a software puzzle “Find Sophie”. The puzzle is a modification of the famous NP-hard “Traveling Salesman” problem.

Unfortunately many job interviews require software engineers to solve hard programing puzzles. The puzzles do not test abstract thinking, understanding, communication skills, reasoning, learning, planning, and problem solving. In short, it does not test any of those things that define intelligence (http://en.wikipedia.org/wiki/Intelligence).

Luckily, Microsoft is no longer the front runner of the software industry and majority of the companies stopped imitating Microsoft and do not as “Four People on a Bridge” type puzzles. It is fair to say that Google, Facebook, and Apple have taken Microsoft’s place as software industry leaders. So now every company wants to be the next Google or Facebook and, as a result, they ask the same questions that Google, Facebook, and Apple ask – algorithmic puzzles.
It is almost impossible to solve those puzzles without the knowledge of basic (and not so basic) algorithms. “Find Sophie” solution is based on graph theory algorithms so before we get into details we have to review some of them.

First, some useful terms:

Vertices are the circles in a graph and edges are the lines connecting the vertices. Big-O notation is used to quantify the running time of each algorithm. Graphs can also be conveniently described using a two-dimensional matrix that contains 1 if there is an edge connecting the two vertices and 0 otherwise. Weighted graphs have a number assigned to each edge. For weighted graphs, ones can be substituted with the non-zero weight of the edge. Trees are a subset of graphs – directed graphs. For trees, vertices are called nodes. Each node can have zero or more children (nodes that can be navigated to from the parent node).

Depth First Search
DFS Mostly used to traverse a tree, and it does so by going as deep as it can until it hits a node that has no children. Then it backtracks. Non-recursive implementations use a stack to keep track of the nodes traversed.
O(V + E)

Breadth First Search
BFS Instead of going into depth, like in DFS, we first explore all the branches (children) before backtracking. O(V + E)
Dijkstra's Used to find shortest paths in non-negative weighted graphs between a pair of vertices. A queue is frequently used to implement the algorithm. O(V^2) (basic implementation)

Floyd–Warshall
Used to find shortest paths in weighted graphs between all pairs of vertices. Unlike Dijkstra’s algorithm will also work for negative weights. Fails to calculate the shortest path if negative cycles exist.
O(V^3)

Prim's algorithm
Used to find a minimum spanning tree (a tree that connects all vertices together and has the minimum total weight) in a weighted undirected graph.
O(V^2)(basic implementation)

Now we are finally ready to explore the THE PUZZLE.

After a long day of coding, you love to head home and relax with a loved one. Since that whole relationship thing hasn't been working out for you recently that loved one will have to be your cat, Sophie. Unfortunately you find yourself spending considerable time after you arrive home just trying to find her. Being a perfectionist and unable to let anything suboptimal be a part of your daily life, you decide to devise the most efficient possible method for finding Sophie.
Luckily for you, Sophie is a creature of habit. You know where all of her hiding places are, as well as the probability of her hiding in each one. You also know how long it takes you to walk from hiding place to hiding place. Write a program to determine the minimum expected time it will take to find Sophie in your apartment. It is sufficient to simply visit a location to check if Sophie is hiding there; no time must be spent looking for her at a location. Sophie is hiding when you enter your apartment, and then will not leave that hiding place until you find her. Your program must take the name of an input file as an argument on the command line.

Input Specifications:
The input file starts with a single number, m, followed by a newline. m is the number of locations available for Sophie to hide in your apartment. This line is followed by m lines, each containing information for a single location of the form (brackets for clarity): where probability is the probability that Sophie is hiding in the location indicated. The sum of all the probabilities is always 1. The contents of these lines are separated by whitespace. Names will only contain alphanumeric characters and underscores ('_'), and there will be no duplicate names. All input is guaranteed to be well-formed. Your starting point is the first location to be listed, and in effect it costs you no time to check if Sophie is there.
The file continues with a single number, c, followed by a newline. c is the number of connections that exist between the various locations. This line is followed by c lines, each of the form: where first entry is the name of location and the second one is the number of seconds it takes you to walk between them. Again these lines are whitespace-delimited. Note that the locations are unordered; you can walk between them in either direction and it will take the same amount of time. No duplicate pairs will be included in the input file, and all location names will match one described earlier in the file.

Output Specifications
Your output must consist of a single number followed by a newline, printed to standard out. The number is the minimum expected time in seconds it takes to find Sophie, rounded to the nearest hundredth. Make sure that the number printed has exactly two digits after the decimal point (even if they are zeroes). If it is impossible to guarantee that you will find Sophie, print "-1.00" followed by a newline instead.

EXAMPLE:
It definitely looks like a graph problem but it is almost impossible to solve a problem like that without a good example:


In our example, Vertex A is the entrance to the apartment (the first possible hiding place of the cat). The probability that the cat is hiding there is 0.2. It takes 2 second for the guy to walk from A to B (another hiding place), and the probability of the cat being there is 0.3.
By definition, expectation, in a discrete case, is x1*p1 + x2*p2 + x3*p3 + … where x is the value with probability p. For example, expected value of finding Sophie if we go to location C and on the way pass location D is probability of cat being at location D multiplied by time it takes to reach location D plus probability of cat being at location C multiplied by time it takes to reach location C –> 0.4 * 5 + (5+9)*0.1 = 3.4.
We have to visit all vertices to find the expected value. In our example there are 6 possible paths.
ABCD, ABDC, ADCB, ADBC, ACBD, ACDB
Mathematically speaking, we want to rearrange 3 (BCD) different objects into a sequence. In discrete math, this is called n-factorial (3! = 3*2*1 = 6).
Now it might seem obvious that ADCB is a better choice than ABDC (via A) because going from B to D we have to come back to A but this is not correct. While for ADCB we have 5*0.4 + 0.1*14 + 20*0.3 = 9.4, For ABDC (via A), we have 2*0.3 + 9*0.4 + 18 *0.1 = 6. It is easier to represent all 6 paths as a tree.


If we calculate all six paths we will see that ABDC (via A) gives us the best expected value. So the algorithm that we used to solve this example does the following:
1) Find all permutations (ABCD, ABDC, ADCB, ADBC, ACBD, and ACDB).
2) For each permutation find shortest path for each pair of vertices on the path (for ABCD, find shortest path between AB, BC, and CD)
3) For each permutation calculate the expected value using the shortest paths.
4) Choose the minimum expected value
That is a good solution but it has a problem. Its time complexity is at least O(N!). So solving the problem for 20 nodes will be unrealistic (20! = 2432902008176640000). So we really need to find some optimization. And by looking at the tree we can see that we can easily disregard at least some of the paths without calculating the entire path. So if we start from the left, and calculate ABCD and then ABDC we know that current minimum expected value is 6 and there is no need to calculate ACDB because ACD already sums up to 6.1. While in case of 4 nodes the optimization is minimal, for 20 nodes this simple optimization will help us to disregard entire branches, transforming the solution from to O(N^3). Of course, the worst case is still O(N!) but let’s hope for the best.

SOLUTION:

Once we understand the example, it is easy to see the solution. For N hiding places, we construct a tree that has (N-1)! paths (each permutation is a path). For each path we calculate the shortest (in terms of expectation) path between every two vertices on the path and sum them up. This way we calculate the expected value of each path, and all we have to do now is to find the minimum one. And yes, as discussed above, we have to add some optimization to achieve a better average running time then O(N!).



( Read more )

T-SQL has SELECT TOP xx PERCENT syntax to return xx percent of rows.

MySQL does not have this functionality but it can be replicated using the following script.

set @rowcnt = 0;

SELECT @rowcnt := (count(*) / 4)
FROM person;
PREPARE STMT FROM "SELECT * FROM person LIMIT ?";
EXECUTE STMT USING @rowcnt;

Here we selected 1/4 of records from the person table.

mysqldump is commonly used to back-up MySQL databases. It can easily back-up entire tables. But what if you want to back-up only certain columns from a table?

There are many ways to export data but here are two of them:

1)	mysql –h<server_name> -u<user_name> -p<password> -e "SELECT col1, col2 FROM <table_name>;"  > C:\test.csv

2)	SELECT col1, col2 FROM <table_name> INTO OUTFILE 'C:\\test.csv' FIELDS TERMINATED BY ',' OPTIONALLY ENCLOSED BY '"'   LINES TERMINATED BY '\n' ;

While the first command is executed from the command prompt, the second one is executed from within SQL server. I personally prefer the second statement because it runs faster than the first one.

To import the data back into a database, we will first need to create a table and then run the statement below:

LOAD DATA INFILE 'c:\\ test.csv' INTO TABLE <table_name> FIELDS TERMINATED BY ',' OPTIONALLY ENCLOSED BY '"'   LINES TERMINATED BY '\n' ;

IT may be useful when you need telnet.exe on win7 machine and for some reason want install/enable it using .bat files

pkgmgr /iu:"TelnetClient"

Statement-based replications vs. row-based replications

Statement-based replications record actual SQL statement that made changes to the database (update, insert etc.). One of the disadvantages of statement-based replications is that not all changes can be adequately passed to the slave. For instance if a timestamp is generated using SYSDATE, and since we log the actual query, the dates will differ on the master and the slave.

Row-based replications record each change made to a row (i.e. value of column B in table A has changed from 2 to 5). One obvious issue with row-based replications is that they require much more space than statement-based replications. Row-based replications were introduced in MySQL 5.1 to overcome limitations of statement-based replications.

A side note:

SHOW BINLOG EVENTS \G

Will show the very first binlog file!
To display a specific binlog file type:

SHOW BINLOG EVENTS IN 'mysql-bin.000057'\G

50. «Programming today is a race between software engineers striving to build bigger and better idiot-proof programs, and the universe trying to build bigger and better idiots. So far, the universe is winning.»
— Rick Cook

49. «Lisp isn't a language, it's a building material.»
— Alan Kay.

48. «Walking on water and developing software from a specification are easy if both are frozen.»
— Edward V Berard

47. «They don't make bugs like Bunny anymore.»
— Olav Mjelde.

46. «A programming language is low level when its programs require attention to the irrelevant.»
— Alan J. Perlis.

45. «A C program is like a fast dance on a newly waxed dance floor by people carrying razors.»
— Waldi Ravens.

44. «I have always wished for my computer to be as easy to use as my telephone; my wish has come true because I can no longer figure out how to use my telephone.»
— Bjarne Stroustrup

43. “Computer science education cannot make anybody an expert programmer any more than studying brushes and pigment can make somebody an expert painter.”
— Eric S. Raymond

42. “Don’t worry if it doesn’t work right. If everything did, you’d be out of a job.”
— Mosher’s Law of Software Engineering

41. “I think Microsoft named .Net so it wouldn’t show up in a Unix directory listing.”
— Oktal


( Read more )