Personal info

Here's a list of my current projects. If you want more information, let me know, I might actually put the source and/or executables for these up for ftp or http. These projects are roughly most recent project first.

This list is likely to be out of date, as the last project on this list I haven't had time for in more than a year. But I like to think that some day I might get back to them. Every now and then (once a year?), I reorder the list and update things so that most recent work is at the top.

• X On a Floppy I have a single floppy boot disk that can turn a PC into an X terminal. The disk works extremely well, and is in use by several departments. If you are interested, send me e-mail, and I'll give you details on how to get a copy. The MMAE department uses it to access a unix based cad system from pc's in the lab. The disk autoconfigures based on the hardware it finds and additional information pulled from various configuration servers on the local network. Currently, there are two versions--one is very small, but somewhat outdated. This works very well on old machines that are tight on memory. A newer version is required for newer video cards, but requires more memory, and is not yet quite as generic (i.e., has hardcoded config options) as I would like it.

  Nobody [non-local] has been clammoring to get this, so I probably won't ever get around to distributing it. The disk now has a command line driven configuration procedure that is fairly easy and includes minimal internal documentation. Local printer support exists, but has not gotten wide use or testing.

• Database assisted network management I've been working more or less continuously for almost a year on a project to totally capture all building network configuration information into an Oracle database. The database will take care of ip allocation, DNS, bootp management, and already uses SNMP (with patch cable maps) to deduce the physical location of machines not in the database. Oracle has been more fun than I expected, although I find that their stand alone tools are a bit lacking. (I've had to write what I need, where I feel they should have existed already.)
• asynchronous multicast/multinetwork file transfer Protocol I've recently come to the conclusion that ftp, http, and smtp are somewhat obsolete in that they no longer do a good job at fufilling the file transfer needs of the net. I've begun designing something that fills in the holes. ftp was designed for authenticated file transfers, but is being used for anonymous downloads. http does this better, but doesn't support anonymous uploads well. stmp doesn't handle binary files efficently, but does anonymous "uploads" quite well. Http makes mirroring somewhat easy, but doesn't really support it directly. Design goals for my protocol try to cover all these problems, including anonymous and authenticated transfers in both directions, automatic cooperative mirroring with minimal administrator intervention, global distribution, multicast, bandwidth use, etc... with the intent to allow interfacing with existing protocols as well.
• Robot controllers I have a working prototype for a modular networked robot controller based on the Microchip 16c65 cpu. So far, I have assembly code that keeps track of a rotary encoder and responds to commands via the network protocol, and C code for unix that speaks that protocol and continuously displays the positions of all encoders on the robot's network. Next step will be to add motor actuators and implement speed control using the chip's built in PWM. I'm looking for someone who would be interested in helping to design an actuator that can handle 6-8A @ 24V. So far, nobody's been able to come up with anything that does better than 2A @ 12V without emitting magic smoke.
• I have written a state based chat language ("cschat"). The chat language is very similar to uucp's chat language, except that rather than just failing or retrying when the desired output is not found, you can make decisions based on what comes back and go to one of several states. This is primarily directed at modern modems that can do data, fax, voice, touch tones, etc... I already have working a primitive answering machine script based on this language. Eventually, when I have enough of the pieces written, I hope to be able to turn this into a getty replacement ("csgetty").
• I have written a command line based audio CD player for Linux. (It might work on Suns too.) It supports cd changers, shuffle & lots of other stuff. You can read the announcement for it. As it is still in alpha testing, you have to send me e-mail if you want a copy. Unfortunately, I more or less have this workinng as well as I want it to, and haven't gotten enough interest in it from the net to work on it more. If you think you'd like to try it, let me know! As it is unlikely that I'll develop it much further, I should probably release it as is and let a wider audience play with it.
• I wrote a cute little program called httpget.c that allows downloading of files via http from the commandline with a high degree of feedback to the user. Please send me feedback if you use this! I consider this project complete unless someone asks for an expanded feature. Every now and then, I add something to httpget. Recently, I discovered that httpget was the smallest file transfer program I could find, so I incorporated it into the X-on-a-floppy project too. :)
• I have some ideas for a distributed chat protocol that would not have the problems that IRC has. Let me know if you'd like to work on a project like this. Some of the ideas would integrate nicely into multiple other projects listed on this page, so we wouldn't have to start writing from scratch...
• I wrote a www counter for those who can't use server includes. You can read the counter page for documentation and rules of use.
• I am working on a program tentatively named OmniClientProgram (ocp). OmniClient is able to interface to both muds and irc. It is still under heavy development, and not yet ready for source code distribution. A sparc binary, a linux binary, documentation, and sample scripts are available. The client's primary feature is a very strong event driven scripting language and a very simple terminal I/O interface model that should be effective on any system. The sample scripts archive includes a script for irc that is about 80% complete. Note that the linux and sparc binaries might not be the same version (although they usually will be). Don't hesitate to mail me suggestions, feature requests, proof reading of docs, critiques, bug reports, new sample scripts, additions to existing scripts, etc.

  Actually, I started rewriting this from scratch in C++. The old version is still available as binaries, and if anyone asks, I might consider releasing the source. I will eventually release the source code to the C++ version. The new version will be significantly easier to expand, and therefore more powerful. Planned new features include multisession handling, advanced screen control (including highlighting), and multiple alternate complex protocols, including possibly talk/ntalk and my own ideas for a serverless distributed irc.

  Actually, unfortunately, due to lack of interest, I've abandoned this project. OPC sits half written, smc is stable, but stagnant. I may eventually get back to this project.

• I've written a number of Mostly Harmless screen hacks for X11. Some day I might put them up for ftp. If you really wanna know, and you happen to be on an X terminal, and you see me online, just xhost + and let me know. >8-} Most people have been at least amused by my screen hacks. When I'm bored, I use them for backgrounds on my desktop.
• My favorite screen hack uses a special random number generator to completely fill your screen's root background with random color pixels. The result looks sortof like xgranite (vintage X11R3?) except that it draws it more gradually and doesn't freeze your screen while it does it. Also, it doesn't allocate a display pixmap, which means that when you move windows around it has to redraw the pattern...but that can cause interesting effects too, and it doesn't slow down your server by chomping all your memory.
• I am currently working on the pieces to a multiplayer 3-d game of some sort. So far, I've written a wire frame library, including an object previewer called 'spin'. You can play with some binaries of spin, along with a few objects I've stolen from cbzone and rend386 here. Unfortunately, IMHO my wire library is not good enough for the final game. I'm looking for something better, but I haven't seen it yet. Lack of a good library has kinda discouraged me on this project. (Several have been released, I need to look for them & try them out...)
• I have a really cool mandelbrot set viewer for X11. Unfortunately it desperately needs a control panel. I have written some minimal documentation for it. Also, I have made available several binaries for Linux. (The binaries are probably stale by now. I'll make new ones upon request.) Solaris 2 binaries available on request. I've worked on this project on and off over the last 3 years. This would be a perfect program to convert to threads, as it really would work much better if the control, calculation, and screen updates were all done in separate threads. Also, I've had limited success with using distributed computation to speed it up; as is, network glitches can cause longer delays than the speed up gained from additional cpus if the machines are not all on the same physical net. A more robust network protocol would help. (Does PVM cover this?)
• I've designed a file backup and retrieval system that is similar to dump, but with added functionality and machine/operating system independent. (Archives should be portable.) Unfortunately, I haven't yet gotten around to actually writing it. Let me know if you'd like details, or if you'd like to help me write it. If someone would actually like to use this, I'd be willing to work on it for a small fee...ala GNU. Contact me for details. If you want to write your own backup system and use all or parts of my specs, I'd be happy to give them out, as long as I get a copy of the result. :) There are currently quite a few backup systems out there, so I probably won't work on this unless I get around to doing a survey of all of them and don't find one I like. (I use dump at home now.)
• I wrote an object oriented paint program for X in C. (Ya, it was fun. Ya, it was a mistake. Ya, I shoulda done it in C++. Sigh. Perhaps I'll translate it some day.) Translation to C++ would be semi-trivial; if I ever do that, I'll put the source up for ftp. In the mean time, you can read the minimal documentation and try out the binary for linux or the binary for Sparc/SunOS 4.1.x.
• A few years ago ('92?), I wrote a postscript to ascii translator. The translator requires Ghostscript and Perl to do its work. I suspect my translator is more robust than the one that came with ghostscript when I wrote it, but I haven't looked at theirs in a while. Unfortunately, I wrote this as a proof of concept (before the one with ghostscript was distributed widely) more than an actual project, so it has some problems, and I never really finished it. If you are interested, download it and take a look. Send me mail if you have comments... As I wrote this as a proof of concept, I have no plans for further development. I think someone else on the net made this more useful, but I've lost the reference.
• I've considered writing a raster to vector image converter for quite a while. A possible output format would be Postscript, which of course, supports both raster and vector images, so could potentially still describe the original exactly even after vectorization. This could be a fairly simple or fairly difficult project depending on how complete a job is attempted. It would be trivial to do things like rectangles, lines, and possibly even circles. OCR of course would be tricky, but others have done this well, and many of the fonts are already available if the original image is a screen capture. Converting splines correctly would be tricky. Converting amorphous objects would be difficult (and possibly not worth it). Detecting pattern fills is definately needed, as a large part of the usefulness of this would be to remove ugly patterns and/or rerasterize them at a higher resolution. Likewise, removing dithering would be difficult, but possible, especially if the algorithms are tuned to unrasterize images from specific applications.

I work on projects when I have time, and the top project on this list always gets the largest time slice. If you express interest in one of the other projects, I might be persuaded to work on them instead...

And when I'm not doing any of these, sometimes I go to the beach.