Every once in a while I get stuck working trying to get something
to work. The man pages don't help, the online
documentation is incomplete anyway. I type a concise description into
Google and hope that
someone else out there had the same problem and bothered to
write their solution down. Most of the time someone has, but
not nearly always. This is my attempt to write down small
issues I have had getting things to work, a few tricks and
tips.
I recently had to set up a SMTP server in Windows for testing an application in development. Turns out Microsoft is nice enough to ship Windows XP with a bunch of built in servers. They call it Internet Information Services. [full article]
We created an interesting data stucture for our load balancing distributed
execution system. This data structure provides FIFO queu and priority
functionality simultaneously.
[full article]
gnuplot 4.0 does not do histograms out of the box. But its solid box
drawing capabilities can be hacked into drawing some pretty nice
histograms. I wrote a Python script to automate the process of making
clustered histograms.
[script]
[full article]
LAM/MPI is really nice for writing parallel code. Yet, any decent size C/C++ application is bound to require some debugging. For my Senior Conference final project I am working on writing parallel terrain surface interpolation algorithms. I'll be putting all the LAM/MPI debugging lore I had to learn the hard way here.
In an ideal world, URI's simply indicate how a resource on the Internet can be accessed. To this extent URI's should only contain information relevant to the content sought by the user. Server side implementaton should be transparent to users: specifically users shouldn't be seeing implementation specific extenstions (php, aspx and whatnot). URI's should also be easily guessable and follow a logical hierarchical organization of content.
mod_rewrite is a powerful tool for handling incoming request URI's in Apache. However, there are several issues that arise from using it in a context where content is dynamically generated and there are seperate production and development servers. I've cleaned up my own URI's, I'll posting information about how I tackled these issues soon.