Sometimes you just have to redirect one of the standard streams, for example, if you want to log all output to a file, but don't have any good way of doing that for whatever reason. This technique can also be used to redirect the standard stream and create additional streams for child processes, and so on.

An issue that I've run into a couple of times now is that when using a template class I've wanted it to encapsulate another class, using it for a number of operations. This isn't really a big deal, because any class that actually implements functions named what I call will work.

I'm finally starting to get back into real programming, which is delightful.  For about a month or so I was working almost exclusively on PHP modifications to Drupal, support scripts, and database maintenence.  I almost forgot how to code, or at least how much I love it.

I'm back on the Congo coding track, but there are a few other things that are starting to come to mind now that real coding has taken the fore again.

I actually wrote this offline a bit ago, it's pretty technical, but it's a cool format for basically having many numerically indexed, block allocated streams in one.

Apparently when I said updated regularly, I meant updated regularly when I actually have any time to think about any of my projects. Lately we've been working an a pay project which is a great project, and fulfilling, but not very interesting from a programming problem point of view. What we have been doing is over at http://coloradolocalfirst.com and is starting to come together.

Currently there are few options for organizing and displaying a "document" in a consistent, flexible, visually appealing way.  There's always HTML/CSS which is renound for inconsistancy and unexpected results, docbook addresses the need for a documentation meta-language, but focuses on producing many different incompatible formats of output from a single source file.  There are a number of new systems that take xml and use it to produce postscript or pdf output, some of which are looking pretty cool, but fall short of being useful to more than a very narrow niche market.

For a long time I've been fascinated with algorithmically producing fractal landscapes.  The main application of this that interests me is creating fantasy worlds for gaming (video, pencil and paper, etc.).  This is a tricky problem, but I realized recently that random landscaps won't really work when you already have some preconcieved notions of what the world should look like.  Then I set out to create an interactive landscape editor.

I've been struggling with the concept of having simple, easy to use packages for pretty much everyone out there, and at the same time have the flexibility to do crazy advanced things like have four different versions of postgres or mysql on a system at once. For most people, they want a simple desktop or simple server machine and want upgrades to be as easy and seemless as possible. If there's a safe way to upgrade postgres 7.2 to 8.1 then that's what I want to happen. But in some cases you need both.