In *nix parlance, random and urandom are virtual devices providing sources of random data. While random makes the user wait for high quality data, urandom just hands over whatever data is at hand -- even if the quality is poor -- in the hopes that it will be good enough to be useful as long as it is timely.

In the spirit of urandom, I went spelunking through my personal source code repository for things that never got to release quality (or even completed, in some cases) or just didn't seem general enough to publicize, but which may be of use to someone nevertheless. Most of them were used for a short period and then forgotten, though a few I use regularly for personal projects but never put in the extra effort to spit shine and release. Very few have traditional test libraries or install scripts or similar helpful niceties. Heck, I had to add copyright notices to most of them before uploading them here . . . .

Scripts

micro-perf 0.1.1: .tar.gz (2 KB)

An embarrassingly simple script I use as a wrapper for the Benchmark module; basically a refactoring of benchmark boilerplate so that my test files could become dead simple. The tarball includes the runtests script, and a few samples testing ways to speed up various operations I found to be bottlenecks at one time or other, both in the Perl core and in the Time::HiRes and PDL modules.

study 0.1.1: source (6 KB) sample data (.txt, 53 KB)

study is a command line flash card app I used to study for an "anything goes" math test. (Thankfully, the test was easier than I had feared, and I got lucky a couple times, so I passed.) The sample data file is the flash card set I used; the cards gave me flashbacks when I looked them over just now. study was clearly designed to be extended for different types of questions than merely doing definition flash cards, but it turned out I didn't need anything more than that. YAGNI strikes again.

Run study and enter ? to see available commands.

perlbox 0.1.10: source (13 KB)

Once upon a time, I thought the concepts behind busybox and Perl Power Tools were pretty cool, so I thought I would combine them -- and hopefully get the economies of scale busybox enjoys in a happy Perl-based toolkit (besides, Perl Power Tools got stalled and it was looking like noone would pick it up). That would also give me a chance to learn quite a lot more about obscure *nix commands and unusual options and parameters. A few commands in, and Real Life took over all my time. I still plan to get back to it some day, because I still think it's cool, but I'm not sure when that will be. Also, even though Perl Power Tools got stalled, I noticed recently that a couple other projects have popped up with similar goals, so hopefully the "market" will be well-served in the mean time.

Use perlbox -? to see what works so far.

Modules

Config::Multilevel 0.04: tar.gz (7 KB)

Config::Multilevel is Yet Another Take on reading and writing config files, created several years ago to scratch an itch. It's been used by a number of my random scripts over the years, but its most important users right now are the Status module (below) and PIGGE, my latest 3D engine. This module is here on the urandom page rather than up in the main Perl section mostly because it lacks a decent test library, and hasn't been converted to modern module packaging standards. That's on my short list, though, so expect this to move up in a later version.

Status 0.03: tar.gz (6 KB)

Status is one of my many iterations of a combination logging / status / profiling module, with lots of configurability. It requires the Config::Multilevel module above, and itself is a prerequisite of the PIGGE 3D engine. Like Config::Multilevel, this module is here rather than in the main Perl section mostly because it lacks a decent test library, and hasn't been converted to modern module packaging standards. Updating both modules is on my short list, though, so expect them to move up to the big time in a later version.

Boilerplate 0.1.1: source (11 KB)

Some time ago, I got frustrated at how many times I had written similar code to do all the "proper" stuff that good *nix admin scripts, daemons, servers, and so forth were supposed to do. So I threw YAGNI completely out the window and started writing a module that would have all of the boilerplate in it I hoped I would ever need. The user would merely have to use the module and provide the core functionality for their program as a set of methods on a magic program object. I got a few KB of code written and ran out of tuits. Since then, amazingly, I haven't actually needed to write a completely new program of the type that could use Boilerplate, so development seems stalled for at least a while yet and it never made it out of my "prototypes" tree. (Define 'irony'.)