Good Video

I have been watching some of the TED talks, some of them are really good. But for some reason I have seen this one a dozen times already.

Science Posters. (Made on a Mac)

Don't forget to check out the
2007 WWDC science posters.

2008 A New Year

2008 is here. And it has been here for a while now. I'm just getting started on a new job so I haven't published anything new for a while. With the job change there is also a small change (not a big one in this case) on my research area. I will start doing some Density Functional Calculations (DFT) so I might be talking a little bit about it on here from time to time. In the mean while I'm using the MIT open courseware to start learning the techniques used to do this kind of calculations. If you are interested you can take a look at Atomistic Computer Modeling of Materials.

Best Regards
Chuy

Google Chart API : part II

I solve my mistake, if you want to use charts into blogger just don't break the src=".." lines into pieces like I did in the example in the last post. The whole src have to be in one line.

Google Chart API

This chart was made using Google Chart API.

After Playing for a while trying to generate the previous plot I couldn't use google API properly. I had to generate the plot on my personal website save it to disk and post it. This API is a great way to make real-time charts, but for some reason it doesn't work within blogger it must disabled for security issues or something.

There is a couple of things I don't like about the API, and some serious limitations so I will stick to gnuplot. but it's nice to have different options with different capabilities.

The biggest problem I find with this API is that you must format your data to make it compatible with the API. Lets say you are running a simulation, and your output data ranges from 1 Tesla to 7.24 Teslas dependent on Temperature, you will need to map the data from 1 to 100, label the chart appropriately so it reads 7.24 on the highest data point and make sure your Temperature interval is uniform. It's not a difficult transformation but it's annoying.

On the other hand, if you are running your simulation on a server, you can easily modify your program so it will write a preview output. Lets say 1 of every 100 points calculated goes to this preview output. And make a web page you can access from any computer with firefox. The web page will then plot in real time what this preview output file looks like. It's also possible to make a web page that will take track of several simulations running on different servers (computers).

The code used to generate the chart is:

<img src="
http://chart.apis.google.com/chart?
cht=lc
&chco=ff0000,00ff00,0000ff
&chs=200x125&chd=t:0,0,0.3,0.7,1.2|0,0.3,1.4,3.1,5.3|0,0.8,3.9,8.7,15.2
&chxt=x,y
&chxl=0:|0|1.0|2.0|1:||20||60||100"
>

Reasons not to publish.

I recently run into this short story:
Reasons not to publish. by Gregory Benford.

Multivalued hell

Hello All.
I have been unable to post stuff thanks to my new friends called multivalued funtions and brach which have been keeping from doing any progress on my research. So next time you use any of these multivalued function (i.e. the arcsin(x)) stop to think about it for a second, I will.

I'm working on the second part of the linked list post, that should come out soon. Also this month on the cise publication the article "On the evaluation of Finite Hilbert Transforms" proposes the double exponential method as a good numeric method to find Hilbert Transforms. I talked about Hilbert Transforms before, and I use them in my research so I will implement the method and publish the code on here.

One more thing. This is just a mathematical curiosity. I was trying to calculate the derivate of x^x, as it turns out you can take the natural log on both sides of the equation before taking the derivative. I get ln(x^x) = ln(x*x*x*...) = x*ln(x) but this bugs me since x doesn't have to be an integer. Any ideas?