|
Doing the experiments is fundamental to good science.
November 01, 2006
URL:http://www.drdobbs.com/embedded-systems/do-the-experiment/193402891
Ed is an EE, an inactive PE, and author in Poughkeepsie, NY. Contact him at ed [email protected] with "Dr Dobbs" in the subject to avoid spam filters.
SAS, which in this context is the Society for Amateur Scientists, held its Fourth Annual National Citizen Science Conference in late August. Unlike huge industry conferences, this type of gathering doesn't attract a large crowd and, in fact, essentially all of the attendees also presented papers.
Although this SAS isn't a programmer's organization and the convention had nothing to say about hammering out code, you'll find a link between today's students and tomorrow's tech industry, plus a missing link in the desktop Linux revolution.
Dr. Shawn Carlson wrote the "Amateur Scientist" column in Scientific American from 1995 through 2001, until the magazine discontinued it. That column's 73-year run inspired generations of amateur scientists with examples of meaningful scientific experiments on a minimal budget. The magazine now devotes a similar number of pages to humor, light commentary, and superficial descriptions of tech gadgets.
Dr. Carlson received a MacArthur Foundation "genius" Fellowship in 1999, an indication that he's not to be taken lightly despite his engaging manner.
He founded the Society for Amateur Scientists in 1994 with the intent of "Helping ordinary people do extraordinary science." SAS members tend to be folks who find out how the world works through experiments using equipment ranging from cardboard and tape to state-of-the-art gear, but with dedication second to none.
For example, Bill Appleby decided to find out how fabric similar to that on the Hindenburg zeppelin burns when ignited. Much to his surprise, he found that uncoated cotton burned faster than any of several coated versions, including fabric doped with highly flammable cellulose nitrate. None of the samples burned fast enough to account for the Hindenburg's rapid demise, which disproves the flammable-fabric explanation of the disaster.
Oddly, it seems he was the first person to actually perform such an experiment, rather than simply proclaiming that the fabric must have fueled the fire. His equipment consisted largely of a charcoal grill to support the samples, so cost wasn't keeping anybody else from testing the claims.
Matthew Templeton described amateur astronomers' observations of variable star intensities, noting how many variable stars there are and how few professional telescopes have time to watch them. It seems amateurs provide many of the continuing measurements, as well as the occasional first-sighting newsflash.
On the other end of the electromagnetic spectrum, Joseph DiVerdi gave us a tour of the radio telescope installation at Colorado's Table Mountain, which lay abandoned for several decades before a group of amateurs began rehabilitating it. Generations of owls deposited cubic yards of pellets in one of the telescope mounts, so science took a back seat to simple cleanup for quite a while.
Forrest Mims III showed the many observations he and his family have made, ranging from atmospheric ozone concentration through detecting microbes carried by forest-fire smoke. Aleta Karstad presented her gorgeous nature drawings in a talk showing how we should add illustrations to our field notes; as I can rarely trace a straightedge without getting my thumb in the picture, her example was somewhat discouraging.
Although this was a family affair with several youngsters in attendance, most of us were old enough to reminisce about the Good Old Days when we nearly killed ourselves while "experimenting" with chemistry of one sort or another. That, unfortunately, highlighted another problem covered by David Wheeler in his talk on Home Science Projects and Homeland Security. It seems that the effort to stamp out crystal meth labs has made it illegal for Texans to own anything shaped like an Erlenmeyer flask, let alone buy many useful chemicals. Of course, generations of moonshiners have done high-quality chemistry without any glassware at all.
The fact that budding scientists can't perform many simple experiments for lack of equipment and materials is just part of the problem.
Anyone following the progress of science, technology, engineering, and math (STEM, not to be confused with "stem cells") education in the U.S. has little to cheer about. Basically, we're not doing very well, as measured by global achievement or our own past accomplishments. You've read of the sheer numbers of Indian and Chinese engineering graduates, which leave little doubt that we're in trouble, and some of my columns have explored the challenges of STEM education.
The standard remedies, which boil down to pouring more money into schools, seem to have produced the problems we're now facing. Dr. Carlson, having worked both within and without the educational establishment for several decades, proposes bringing interested middle- and high-school students together with mentors in a setting dedicated to teaching science by doing science. His LABRats (the odd capitalization evidently distinguishes it from unrelated groups) organization has run a year of prototype sessions and seems well on the way toward receiving sufficient funding to clear the gantry.
LABRats has a structure based loosely on traditional Scouting, minus the uniforms and rigid hierarchy, plus an emphasis on scientific thinking and investigation. Although the program will be inclusive in all regards, the fact that participants elect to join should help maintain both interest and order.
LABRat's "Do the Experiment!" motto focuses attention on the basic premise of STEM subjects: observed facts trump received dogma. As Admiral Grace Hopper (may have) once said, "A single accurate measurement is worth a thousand expert opinions," but it's painfully obvious that popular culture denigrates such reasoning. Dr. Carlson plans to train a nationwide student cadre to think clearly and reach logical conclusions, which strikes me as a worthwhile goal in and of itself.
However, I foresee trouble ahead that has nothing to do with LABRats itself. While it's entirely possible to encourage technically adept students to excel in STEM subjects, it may be difficult to maintain that enthusiasm as they see their parents and mentors laid off in favor of overseas outsourcing. What's the point of working hard for a technical education when the company you eventually work for decides you're just too expensive?
In somewhat related news, the College Board folks who do the SAT tests recently noted that
[f]orty-three percent of 2006 college-bound seniors reported grade averages of A+, A, or A-. Ten years ago, the figure was 36 percent, and in 1987...27 percent. This year's average grade point average was 3.33, compared with an average GPA of 3.21 in 1996 and 3.07 in 1987.
I find it difficult to believe that nearly half of all freshmen are working at A levels, given the abundance of remedial math and writing classes. Indeed, the recent drop in SAT scores shows that either grade inflation is a clear and present danger or that easy courses rule. Perhaps both.
Dr. Carlson's LABRats may be the best hope we've got for changing things at a national level. On the personal level, encouraging your children to think logically and act ethically will start them in the right direction.
You've read the debates over why Linux isn't a player in the desktop market. I hold that Linux won't make any inroads until the hardware driver situation gets ironed out. Comparing your Sunday sale circulars with linuxprinting.org, sane-project.org, and linmodems.org will verify that, to a good first approximation, no cheap (that is, on sale this weekend) printers, scanners, or internal modems work with Linux.
The situation is even worse for tech folks, because essentially no unusual products have Linux drivers. My token Windows laptop runs a battery tester, a vector network analyzer, an ancient Crosspad tablet, various radio interfaces, a few UPSes, and some audio gear. None of those manufacturers evince the least interest in Linux and there are no Linux equivalents on the market.
Back at the SAS conference, Rich Marvin, Onset Computer's Program Manager for Education, described their line of HOBO data loggers. These small, battery-powered gadgets solve a problem that doesn't crop up in the software field: autonomously recording time-stamped, real-world measurements.
For example, one HOBO unit records internal and external temperature, light level, relative humidity, and an external voltage (which can come from a temperature, pH, moisture, or other sensor). A sealed logger, suitable for dunking in nasty aqueous gunk, samples temperature and light. They recently introduced a three-axis 3-G accelerometer suitable for logging the thwacks your experiment endures.
The loggers connect to a PC through a serial or USB cable. Onset provides a comprehensive program that configures the loggers: selecting which channels to record, how often to sample, and so forth. After initialization, you detach ("launch") the logger, put it into the experiment, activate it, and walk off. When the experiment finishes, you retrieve the logger, connect it to the PC, extract the data, and save it in a proprietary file format. You can also export the data to files in formats from Excel to plain old CSV.
Just hearing about these things made attendees want to run out and measure stuff. Onset has a nice series of educational projects to get novices started, a loaner program for teachers to get them started, and plenty of background material showing how it all works.
The configuration software runs in Windows (of course) and Mac OS X. Both versions require the Java Runtime Environment, but I can't tell if they are full-throttle Java apps or not. In any event, the program provides comprehensive data display and analysis, making it more complex than you might expect.
Not surprisingly, Onset doesn't provide a Linux version: "At this time we are not planning support for a Linux environment." Mac and Linux desktop market shares may both huddle in the 2 percent range, but "Linux" subsumes myriad distros, none individually rising above noise level. It makes good business sense to ignore such a fragmented market, even if I really want a native version on my SUSE desktop.
Incidentally, it's entirely legitimate to ship a proprietary program for Linux, as long as it's not linked with code released under, say, the GPL. Some folks hold that simply running a program on a Linux kernel requires GPL licensing, but that's simply not the case.
I hoped that Onset would document the interface sufficiently well so that I (or, more likely, someone better qualified) could hammer out a simple, command-line program, but "The software protocols are proprietary to Onset Computer Corporation and not available to the public."
If I were building a logger, I'd design a small language that did data logging and nothing else, then implement an interpreter in the logger's firmware. The PC program would convert all the configuration information into a program in that language, squirt it over the link, and hang up. The logger then interprets that program, records the external data, and waits for pickup.
Small interpreted languages, which DDJ covers quite well, simplify the interface and decouple the underlying hardware. Unlike the command-and-response interface of old lab-bench instruments, an interpreted language allows the logger to run autonomously.
Regardless of whether Onset loggers use a tiny interpreter, accept a binary slug of machine-language firmware, or parse a structure of configuration values, the company evidently regards that information as a key part of their business. By not documenting it, they retain the ability to change it as needed and, perhaps, hope to stay one step ahead of their competition.
As you've seen in previous columns, however, security through obscurity simply doesn't work. Sniffing a serial- or USB-data link doesn't pose much of a challenge and the interface probably isn't encrypted, so not documenting the interface doesn't provide much security at all. Anybody who wants to extract that information faces maybe a week or two of tedious reverse engineering and it's a done deal. Customers may not have the motivation, but competitors certainly do.
Nonetheless, I plan to buy a few of those loggers for some long-term data-collection projects. It's just one more reason I must maintain that last token Windows laptop.
Speaking of Windows, attentive reader Evan Anderson pointed out a subtle problem with the drive-recovery techniques I explained in the July column. It seems those old-school DOS-style 8.3 filenames still appear in the Windows Registry entries of (wait for it) Microsoft programs, among others, and simply copying the files back to a drive may not reproduce the correct names. No file copying utility can update file and directory names stored in the Registry, that steaming lump of binary information at the heart of Windows, and any mismatch between Registry entries and actual file names spells catastrophe.
Evan gave a simple example. If you install Microsoft Office before Microsoft MapPoint, the two program directories would be C:\PROGRA~1\MICROS~1 and C:\PROGRA~1\MICROS~2. However, if the file copy program you use to restore the drive proceeds through directories in alphabetic order, it plunks MapPoint in MICROS~1 and Office in MICROS~2. You can imagine the ensuing carnage.
He suggests using the freeware XXCOPY command-line utility (from www.xxcopy .com) to copy FAT filesystems while maintaining their 8.3-format short names. Obviously, this works only for functional filesystems, so it's not clear how it would fare after my slash-and-burn repairs, but it's certainly better than cp.
Based on that, here's the plan for my next Windows box repair. As before, I will dd the entire FAT filesystem to the server, open that file as a loopback drive, and wield the Big Hammer as needed. Then, instead of using cp to copy the files back, I'll dd the filesystem intact onto a blank USB hard drive, jack it into the failed system, boot a Windows rescue disk, and then use XXCOPY to transfer the files.
As nearly as I can tell, the only reason that problem hasn't bitten me yet is sheer, dumb luck. Perhaps you can avoid embarrassment with a quick download, too.
The Society for Amateur Scientists is at www.sas.org, from which you'll find the LABRats program at www.sas.org/labrats/ index.html. The executive overview is at www.sas.org/labrats/LABRats_ExecSum .htm and, if you're willing to put up with full-frontal Flash, try http://labrats.org. The references in the various presentations will guide you to more data on the educational problems we face. Those GPA numbers are from www.collegeboard.com/press/releases/ 150054.html.
Wikipedia's Hindenburg entry at http:// en.wikipedia.org/wiki/Hindenburg_%28airship%29 cites Appleby's research, which appears at www.sas.org/tcs/weeklyIssues/ 2004-12-17/project1/ index.html.
The American Association of Variable Star Observers is at www.aavso.org. An old view of the Table Mountain radio telescope site is at www.its.bldrdoc.gov/home/programs/tm_quiet_zone/1966/ chap3.html.
Forrest Mims describes his research at www.forrestmims.org and some of Aleta Karstad's work is at pinicola.ca/aleta.htm.
A look at the effect DHS has on amateur science is at www.wired.com/wired/archive/ 14.06/chemistry_pr.html.
Onset Computer is at www.onsetcomp .com, with its Education program at www .iscienceproject.com. Despite my frustration at their lack of Linux support, they seem to have good stuff.
Terms of Service | Privacy Statement | Copyright © 2024 UBM Tech, All rights reserved.