Premier essai avec Parallel Desktop 4. L'installation se passe bien mais les Parallels tools sont incompatibles ce qui est pénible. En plus Parallel ne communique pratiquement pas. Ils sont nuls.
Je veux tenter vmware mais découvre VirtualBox 2.2.2 gratuit. Je tente et ça marche plutôt pas mal !
Voici un résumé:
Chargement de la version 32bits de Ubuntu 9.04
Création d'un disque virtuel dans virtualbox et définir le lecteur optique comme étant sur l'image disque iso précédemment chargée
Lancement de l'installation et redémarrage
Lors de la configuration de la machine virtuelle, le seul point important est de ne pas activer les effets 3D sur les paramètres d'affichage.
Installer les additions via le menu Périphérique -> Installer les additions invités
Voir ici:
http://www.commentcamarche.net/faq/sujet-7576-virtualbox-installer-les-additions-client-dans-ubuntu
Ensuite j'ai voulu mettre à jour Openoffice en 3.1
Désintaller la 3.0:
sudo apt-get remove openoffice*
sudo aptitude purge openoffice.org-core
(Je ne comprends pas ce que je fais mais ça a marché)
Charger la distribution Debian de OpenOffice
La décompacter
ouvrir un terminal et aller dans le dossier de la distribution par cd et faire un glisser-déposé du dossier de la distribution:
Taper:
sudo dpkg -i DEBS/*.deb
sudo dpkg -i DEBS/desktop-integration/*.deb
Il y a un problème car l'environnement java JRE n'est pas présent ce qui empêche d'installer des extensions.
sudo apt-get install sun-java6-jre sun-java6-plugin sun-java6-fonts
Pour que le sudo ne demande plus de mot de passe:
sudo visudo
Mettre sans commentaire
%sudo ALL=NOPASSWD: ALL
ctrl X pour sauver
sudo adduser toto sudo
avec toto l'utilisateur
après je fais sudo -u toto
(ca ne marche pas toujours, je ne comprends pas pourquoi)
Pour monter un dossier partagé mgi dans le dossier '/home/mgi/Bureau/Mount'
Il faut d'abord le déclarer dans la configuration de VirtualBox
puis :
sudo mount -t vboxsf mgi '/home/mgi/Bureau/Mount'
C'est là où c'est intéressant que le sudo ne demande pas le mot de passe car je mets cette commande dans /etc/rc.local et au démarrage, le dossier est directement accessible.
mardi 26 mai 2009
lundi 11 mai 2009
Deep sleep
Activate: sudo pmset -a hibernatemode 1
Desactivate: sudo pmset -a hibernatemode 0
The content of RAM is here:/var/vm/sleepimage
To erase it:
sudo rm /var/vm/sleepimage
Desactivate: sudo pmset -a hibernatemode 0
The content of RAM is here:/var/vm/sleepimage
To erase it:
sudo rm /var/vm/sleepimage
mardi 5 mai 2009
Nombre aléatoire normal sur Excel
How to Calculate a Random Number from a Normal Distribution
Remember that the NORMINV function returns a value given a probability:
NORMINV(probability, mean, standard_dev)
Also, remember that RAND() function returns a random number between 0 and 1. That is, RAND() generates random probabilities. Therefore, it seems logical that you could use the NORMINV function to calculate a random number from a normal distribution, using this formula:
=NORMINV(RAND(), mean, standard_dev)
However, Jerry W. Lewis -- a former Excel MVP and a professional statistician -- offers a stern comment about this approach. "Using NORMINV(RAND(),...) to generate Normal variates is totally unacceptable prior to Excel XP, and is only marginal in XP. This is because of inadequacies in NORMINV and in the tails of NORMDIST itself."
"NORMINV prior to Excel XP produced a very un-normal fraction of values around 6 million standard deviations from the mean," Jerry wrote. "This is due to inaccuracies in the implementation of NORMDIST and NORMINV. Excel XP brought those values into a less obviously wrong location, but otherwise did little to improve the situation. NORMINV in Excel 2003 is a decent implementation."
Instead, Jerry recommends the Box-Muller method described here: http://mathworld.wolfram.com/Box-MullerTransformation.html
This method, he wrote, is limited only by the inadequacies of the RAND() function prior to Excel 2003, which had unacceptable autocorrelation.
The Box-Muller approach suggests that Excel users should use this formula to calculate a random number from a normal distribution:
=SQRT(-2*LN(RAND()))*SIN(2*PI()*RAND())
The Box-Muller method is mathematically exact, Jerry writes, if implemented with a perfect uniform random number generator and infinite precision.
Remember that the NORMINV function returns a value given a probability:
NORMINV(probability, mean, standard_dev)
Also, remember that RAND() function returns a random number between 0 and 1. That is, RAND() generates random probabilities. Therefore, it seems logical that you could use the NORMINV function to calculate a random number from a normal distribution, using this formula:
=NORMINV(RAND(), mean, standard_dev)
However, Jerry W. Lewis -- a former Excel MVP and a professional statistician -- offers a stern comment about this approach. "Using NORMINV(RAND(),...) to generate Normal variates is totally unacceptable prior to Excel XP, and is only marginal in XP. This is because of inadequacies in NORMINV and in the tails of NORMDIST itself."
"NORMINV prior to Excel XP produced a very un-normal fraction of values around 6 million standard deviations from the mean," Jerry wrote. "This is due to inaccuracies in the implementation of NORMDIST and NORMINV. Excel XP brought those values into a less obviously wrong location, but otherwise did little to improve the situation. NORMINV in Excel 2003 is a decent implementation."
Instead, Jerry recommends the Box-Muller method described here: http://mathworld.wolfram.com/Box-MullerTransformation.html
This method, he wrote, is limited only by the inadequacies of the RAND() function prior to Excel 2003, which had unacceptable autocorrelation.
The Box-Muller approach suggests that Excel users should use this formula to calculate a random number from a normal distribution:
=SQRT(-2*LN(RAND()))*SIN(2*PI()*RAND())
The Box-Muller method is mathematically exact, Jerry writes, if implemented with a perfect uniform random number generator and infinite precision.
Inscription à :
Articles (Atom)