Baby Steps in Programming

Once I had chosen which density estimator I was going to use the next step consisted in writing a computer program, a cluster-recognition-program (CRP) which would implement those calculations on any field. I chose Interactive Data Language (IDL) as the programming language which I would learn and use and, since I had no past experience with IDL (and only slight experience with FORTRAN), large part of the project was this learning experience.

Basing the first studies on the Lada & Lada$^2$ (2003, hereon LL03) cluster list I began downloading maps of those cluster fields from the then just-released 2MASS all sky survey. The goal was to apply the CRP to a known cluster so has to be capable of comparing and adjusting it until the results matched. The cluster sample is one consisting of nearby (less than 1kpc away) embedded clusters. All the clusters had already been previously studied by diverse authors, this was fundamental because I needed other data with which I could compare my results. In practice the process was harder than it first seemed since many parameters needed to be adjusted in the program.

The first versions of the CRP were not very efficient because they too a long time (15 minutes) to work with a small array of sources (1000 stars). This made progress very difficult since I could not apply the program neither to densely populated clusters nor to those large in size. This inefficiency turned out to be rooted to an important topic in programming and, by discussing my algorithm online at the IDL newsgroup, I obtained several very helpfull suggestions which increased the data processing speed and allowed me to look at larger regions (for more details on this ordeal see: $http://www.dfanning.com/code\_tips/slowloops.html$).

With the program working more rapidly I applied it to a few simple clusters and began comparing the data I obtained with the values in the literature.