3. May 2008 by emeris.

It makes me sad when people abuse science in the name of ripping people off. It makes me sadder when people claim to explain the science being used and muddle it even more.

For example: http://bellasugar.com/1073289 (which I am not linking to because I don’t want to help their google ranking at all (I know google regularly crawls this page)).  First they define an ion, while linking to wikipedia, and still get it wrong. Typically “atomic particles” is referring to the constituents of atoms. Second they claim that ionic hair driers are faster by “…breaking down water molecules into smaller bits…” which strikes me as nonsense. If this were a major factor in drying the hair, it should also cause a fair amount of damage to your skin as well.

I briefly contemplated signing up for the site to make these comments, but then I thought of this xkcd and decided to write a blog post instead.

To see other people pointing out people who are wrong on the internet see here. Teh stoopid, it burns!

Posted in bad science, bug, physics | 2 Comments »

30. April 2008 by emeris.

I work primarily in matlab(2007b 64bit) on a group computational server, hence my main mode of access to the server is through ssh with X forwarding. The problem is that on my laptop I am running the ubuntu 8.04 (32bit) with the compiz eye candy turned on and the java jre does not get along with compiz (from what I gather it has to do with something about repainting frames and swing waiting for the window manager to do that, but compiz never does this so java based programs sit there with grey screens forever, this is a bug in the JRE not a bug in linux and has been fixed in newer releases). The simple solution would be to turn off the eye candy, but some of it does in fact add to productivity (particularly when going from matlab window <-> emacs for code <-> firefox for matlab documentation given the limited desktop space on a laptop screen some of the window manager features are very nice) so after poking around for a bit I found this:
http://www.fedoraforum.org/forum/showthread.php?p=1003847#post1003847

The summary is that matlab (version 2007b) uses java1.6.0, but installing java1.6.0u6 (java 6 update 6 (the forum claims that update 5 also works which is a claim I saw else where)) and the running
$export MATLAB_JAVA=(location of java1.6.0_6)

everything works wonderfully.  This needs to be done on the computational server as that is the jre that is running matlab, not my local version.

Posted in bug, computer, linux | 1 Comment »