When it comes to writing up my analyses, I use LaTeX. It's a wonderful and efficient tool for preparing documents with lots of math and it's straight-forward to automate the construction and inclusion of tables, figures, and bibliographies. I've also found that as I push LaTeX harder and harder, it becomes more and more like programming. Even in its simplest incarnation, you still have to compile your document and worry about syntax errors! As is the case for Stata, there are lots of resources on the web for learning LaTeX. I've put together a few of my favorite LaTeX links to help people get the most out of this powerful document markup language.
In the course of my research, I sometimes write code that other people might find useful. I've packaged up a few of those pieces here:
nptools is a C library and a set of command-line programs that compute multivariate and univariate nonparametric densities and do various things with them. One of the most interesting things is computation of counterfactual densities of the sort reported in McKee and Todd (2007) and in fact, this software was originally written for that paper.
This code parallelizes the GNU Scientific Library implementation of the simulated annealling optimization algorithm. Simulated annealing is a global optimization algorithm that works well when the objective function is poorly behaved and may have multiple local minima. It's not the fastest algorithm in the world, but this code lets you exploit a cluster of computers to speed up the annealing process. Using this C code as a starting point, you should be able to plug in your own objective function and go.
summout is a collection of four new Stata commands that are used to generate tables of summary statistics. summ2 acts just like summarize, but it allows the use of summstore to save means, medians, and standard deviations of variables in a given sample. summout combines saved results to produce tables in Latex, tab-separated value, and fixed (human-readable) format. misumm does exactly the same thing as summ2, except it works with multiply imputed data sets created with Patrick Royston's ice command. All four commands come with Stata help files. This package was inspired by Ben Jann's outstanding estout package.