Another month, another mod_virgule release. With this release, I’ve removed most of the remaining hard-coded, site-specific stuff. It’s now possible to build and install mod_virgule without having to edit the source code to insert the website’s name or the admin’s email address. I’ve moved all that to the site config file where it can be easily changed. There’s still one little clump of HTML used as the header on some internally generated pages that’s hard-coded. It’s getting really close to being usable for real-world websites but there’s still a lot left to do.
When I wasn’t working on mod_virgule, I noticed an interesting new blog. Anyone interested in reading about random Texas nature topics might find The Nature Writers of Texas blog fun. It contains nature writing from dozens of Texas authors including pieces originally published in newspapers, magazines, books.
And when I wasn’t working on mod_virgule or reading blogs this month, I was working on robot stuff. I’m still struggling to get the example files to compile properly for the New Micros Tini 2131 board. It seems the examples were developed by someone using a Windows variant of the GNU gcc tool chain and never tried out on anything else. They have a number of upper/lower case problems where a source file says #include “foo.h” but there is no file named foo.h. There is a file name Foo.h or FOO.H. I thought even Windows could distinguish between upper and lower case letters by now but apparently not. While correcting the case problems, I noticed some of the files had what appeared to be illegal characters in them; probably those old timey IBM graphic characters used as lines and boxes around comments or something. I’ve excised the weird characters and have now gotten down to what appear to be actual coding errors. I hope it won’t be too long now before I have some actual working code.
Robots.net is now happily running on CentOS 4.1 and Apache 2. I’ve posted the source and changelog to the new mod_virgule code. Hmmm… what’s next?.
I ran into trouble with the TiniArm 2131 board. The recommended flash programmer is proprietary software from Philips that only runs on Windows. After a little searching online, I found lpc21isp written by Martin Maurer that compiles on Linux and might work, however it’s not licensed under a Free Software or Open Source license. I don’t understand people who release copyrighted source code without some sort of a clear license defining how you’re allow to use their code. In this case, I skipped it and moved on. Then I discovered that Paul Stoffregen had written a GTK+ based flash programmer for the Philips line of ARM processors called lpc2k_pgm. It’s licensed under the GNU GPL and it compiles and runs just fine on my Fedora-based laptop. Unfortunately, it doesn’t yet support the LCP2131 chip. I don’t have the time to add support so I emailed Paul and New Micros. They’re going to loan him a board in exchange for adding support to his software. That’s good news even if it means my TiniARM experiments will be on hold for a little while.
I’ve been busy the last couple of weeks helping with a major renovation of the DPRG website. Susan created some new graphics and we came up with a nice, simple layout. We’re using 100% XHTML 1.0 and CSS for the new site. So far it works great in both Mozilla and IE. Opera is a bit flaky when it comes to rendering W3C standard CSS but it’s maybe 90% there. The hard part is just the sheer number of pages and images. There is around 150MB of content on the site and it’s still very much a work in progress.
All the work on the DPRG website has sucked me back into the hardware side of robotics. I’ve got two new toys to play with. I went by New Micros, Inc. last week and picked up an IsoPod development kit and an MCORE board. The IsoPod is an based on a Motorola DSP5680x DSP that provides 40MIPS of processing power
and a ton of I/O on a tiny little board. To fully utilize it requires the use of IsoMax, a new FSM-based realtime programming language that New Micros developed. Since I’m a diehard C programmer, I also picked up the MCORE. With a 16bit Motorola MMC2107 M*CORE processer
running at 33MHz, it’s not nearly as powerful as the IsoPod but gives me the advantage of working in a familiar language. Well, I should say it will give
me that advantage once I prod GCC into compiling code for it. New Micros only provides Forth for the MCORE. But GCC 3.2 supports MCORE as a target, so
I’m trying to get a Linux-to-MCORE cross-compiler working. Should be interesting. Motorol created a Windows-to-MCORE cross compiler using a much older version of GCC, so I know it should be possible. And both New Micros and Motorola have offered technical assistance. I’ll provide more details on it as I progress.
Susan has a new claim to fame. She has for a while maintained what is probably the best online discography of Robert Shaw’s work to be found. As it turns out, the Atlanta Symphony Orchestra Chorus seems to agree. Their discography Link now points to Susan’s page. I think this means she now maintains the official Robert Shaw Discography.