Thesis Paper

Last week I started my Master thesis. I didn't start writing anything yet of course, but I am doing a research on the relevant theory. What I am going to write about is Hash tables.

My supervisor, Rasmus Pagh, has already published his research and his proposal on better hashing algorithms. An easy to read version of his idea and proposal could be downloaded from here.

We will now go further and investigate other benefits and advantages of his proposals. In particular, we are going to use his algorithmic idea to make comparisons of two sets for duplicated words faster if possible.

The problem could be described as the following. Suppose that we have 2 sets with words. We want to find words in the first set that appear in the second set as well.

The two main competitive algorithms are the Merge algorithm and the use of hash tables. The Merge algorithm is the same one that merge sort utilizes. The use of hash tables, allows us to use one hash table for accessing words and the other one for searching for existence of words.
Both algorithms run in linear time. The proposed algorithm also runs in linear time, but I am trying to make it faster than the others by avoiding branch mispredictions(take advantage of branch predictions) and decreasing cache faults.

Wish me good luck :)

Fotos is back!

Welcome back to the blog community Fotos.

We are excited waiting to see your new style and read all you have to say about your new life :)

(Fotos used to write about big constructions and edge technology and architecture buildings. From now on he is going to write mostly about himself and his life.. That's as far as I can understand from his blog)

Command & Conquer Generals VS Warcraft III

Lately I have been playing CC Generals Zero:Hour and I was trying to compare it with Warcraft III. After spending some hours at lan parties with my friends playing CC, I concluded that there is no comparison at all!!

* In CC you have unlimited money!!! (Come ooooon...)
* In CC there is no army limit!!!
* In CC you don't need micro(and it's not easy to do it), since you have HUGE army and you can just "attack ground".. And, you always have the money to build more :)
* In CC you don't have hit points(in points), attack speed(in any form), attack damage(in any form), armor, movement speed.. Come ooooon..

What kind of strategy game you play if you cannot calculate some basic things?
And you can play forever, because you have have unlimited money...

Even more,
You don't breath with your ghoul..
Your heart doesn't beat with your ghoul's...
You don't believe in Mitsos..
You don't "cry" for losing Mitsos..
You don't shout ("Oxi to mitso re gamwtoooo!!")

Maybe on day I will meet a guy that is a pro gamer at Zero:Hour and I will recall everything I said :)

( I will update this post if I remember more disadvantages of CC )

Tennis Scoring

Every time I was watching a tenis game I was wondering why they are not counting like all the other sports that I know..
In case you didn't know or didn't notice, in tennis, the counting is: love, 15, 30, 40, game. And then you start the next set.
My friend Google, helped me to find some answers.

Probably the counting is based on a clock face, and 45 was abbreviated to 40, probably because it is easier to pronounce. About "love" which means 0, there are 2 scenarios. It comes from "l'oeuf" French for 'egg' (The use of 'goose-egg' for a zero score is common in baseball), or it just means that the game is for love and peace in the world and stuff like that.

This counting system seems reasonable for that time, when the number 60 was considered "good" and "complete" number, in about the same way we consider 100 a "good" and "round" number. I still don't understand why they keep it and they don't make it 0, 1, 2, 3, 4. Actually, currently, you can use both counting systems, but the official is the old one.

6/6/6

A friend remind me something that I was trying to forget for years now..

Today is the day that antichrist is visiting our world..
I want to ask for forgiveness from anyone I hurt and anyone I made suffer with any of my actions or words.

See you hopefully in paradise, but I will probably be in hell :(
(if it is not this year, it might be in 10 years and so on.. )

Ας ξυπνήσει κάποιος τους επιστήμονες επιτέλους!

Αν και όλοι ξέρουμε οτι ο θεός έπλασε τον άνθρωπο απο λάσπη και νερό, ο θεός έφτιαξε τον κόσμο κτλ, οι επιστήμονες συνεχίζουν τις έρευνες τους..

Μα γιατί; Σε αυτούς δεν το είπε κανείς οτι έχει βρει η θρησκεία τις λύσεις για όλες τις απορίες του ανθρώπου;
Συμβαδίζουν μωρέ η θρησκεία και η επιστήμη(ή αλληλοκαλύπτονται), αλλά η θρησκεία τα λέει απλά για να τα καταλαβαίνουμε, ενώ η επιστήμη κοιτάει το πρόβλημα στα μάτια..

Τί σχέση έχει όμως η κυριαρχία του ανθρώπου ώς είδος απο την μέρα της δημιουργίας του και η αγάπη του θεού για το τέλειο δημιούργημα του, όταν είμαστε ζώα και έχουμε κοινό πρόγονο με τον πίθηκο; Ήταν πιο εύκολο για τον άνθρωπο τότε να συλλάβει οτι είναι λάσπη και νερό, απο ότι να καταλάβει ότι ήταν κάποτε κάτι σαν πίθηκος το οποίο στον κατακλησμό του νώε μισοπέθανε και ότι έμεινε έγινε σιγά σιγά άνθρωπος;

Όταν λέω ότι κάποιος πρέπει να ξυπνήσει τους επιστήμονες, μιλάω για τα παιδιά που ασχολούνται με την εξέλιξη και για αυτους που ασχολούνται με την αρχή του σύμπαντος..

Παλιότερες μελέτες πάνω στα ίδια θέματα εδώ και εδώ

Search Engine Project

Last week I started my last project before my thesis. It's about implementing a search engine and consider the complexity of searching algorithms. So far, what I have realized is that it is more about choosing the right data structure to store your data, than optimizing the search algorithm.. At least up to now, the only thing I was doing to optimize the search was to change the data structures.. This week I will focus on the search algorithm and some extra features..

My personal motivation is to make the search faster than the one's implemented by other students. I have a friend that did this project a year ago and I have some measurements from some other students that did it 2 years ago.. Let's see.. I have 3 more weeks..

Station Ripper, part 3

I guess that is the final update for this topic.

I just fixed most bugs I could find and I made it nice and easy.
Also included the possibility to delete automatically all the folders that have one mp3, so that you don't get "empty" folders.. That could be done either when you create the folders(and delete only the currently created folders with one mp3) or to the existing folders(and delete even previously created)

In that way, you can get the most played artists and keep only these( keeping in mind that at least in chill out(the one I am ripping at least) most popular might possibly mean best, since there is not much business involved(compared to house/dance music))

Currently I have about 10GB mp3s of chill out and flamengo :)
and still in the beginning..

Oh, and very important, ripping stations is legal!
So, you can create your legal mp3 collection!!

"Fortunately, recording Internet radio still remains in legal limbo, and there's some precedent for it being legal, since you can record regular broadcast radio." <source>

Station Ripper, part 2

As I wrote at my last post, I was planning to do something to organize the mp3s that you can rip from StationRipper. So, I made an attempt..

I made a very simple program, at which you define the folder that contains the mp3s from a specific station(could even be other mp3s that have the same naming format), and it automatically(of course, that's the point of doing it..) creates the necessary folders and puts the files in them.. So, if a station broadcasted 3 songs of Muse, you will have them in the same directory, which will be Muse :)

I have uploaded the source files (in Visual Studio 2005) and an standalone .exe file.

Don't expect any extreme intelligent programming..
I just finished programming and it is 6.30 in the morning..
But still, it is working fine.. I have tried it at about 2000 mp3's :)

Station Ripper

You have found an online radio station that you like very much and plays great music?
I have great news for you!!
(sounds a bit like telemarketing, right? :D)

Last week a friend talked to me about StationRipper and I found it interesting enough to try it..
It is program that can rip/make mp3s from the online radio stations. It can "listen" to many stations simultaneously and rip. It automatically separates the songs and puts them in different mp3s and separates mp3s from different stations in different folders. It is very nice.. I have already 1 GB of new music and I have it only one night ripping from 3 stations :)

You can select your favorite stations from shoutcast of course and from a preselected list suggested by StationRipper.. Maybe more.. I didn't explore everything, but still I like it :)

I recommend to try it at least..

I am planning to write a program to go into the folder of the station, and group all songs from the same artists in separate folders. Then I will have artists collections of specific genres (since the stations usually broadcast music of one or two genres).

Windows Update

I was amazed/impressed/happy when I learned how to bypass windows cdkey authentication to download windows updates from the official windows site!!

And especially because it is sooooo easy :)

Description here
Thanks to Poul

Statistics

If you have ever been at casino, you have probably seen the table of the last occurred numbers at roulette..

Also, at Lotto and other games where the player selects numbers, some people keep track of the latest results..

There are 2 totally contradicting points of view.
There are the people that believe that it is meaningfully to keep track of the results
and there are these people that say that it is stupid..

Let's use a perfect dice to explain the two ways of thinking..
(by perfect I mean that you cannot foresee the result of rolling the dice.)

The defenders of the first opinion, say that if you roll the dice A times(A should be BIG number), the number of times you get the "1" should be about A/6. And as A increases(we roll the dice more and more times), the “probability of getting "1" A/6 times gets higher.
So, If we have rolled a dice 20 times and we see that all the times we got "1", it is very likely that "1" will not come the next time, because the other numbers have to occur as well with about the same probability if we were to roll the dice 100 times for example.. And if we get "1" again, then the probability to get any other number, increases comparing to the probability to get "1" again.

But, others believe that all the above analysis makes no sense since the experiments are independent from each other. That means, that the rolling of a dice one more time, has nothing to do with the result of the previous rolling..

I hope I made myself clear.
I guess that the second view is the correct one..