Slashdot Mirror

Photographing documents is always terrible with phones because they have no variable focus. So they use a pinhole camera for infinite focus, but it fails at less than about 3 meters.

This is not the case with all devices. The HTC Dream/Android G1 has an autofocus lens. You can hear it focus. It beeps and uses a red/green framing indicator to show whether focus worked. It seems to focus down to about 3".

Here's a sample of as close as I could focus on an LCD screen:
Reduced from full resolution: http://twitgoo.com/4ffwh
the 1.5MB file: http://graflex.org/klotz/2009/10/g1cam/cam.jpg

I don't think the calculator is a particularly good example; the person who wrote it encoded the state machine of the keys by switching entire bindings in and out.

Recently I wrote an entire webmail application with a small PHP back end that outputs the mailbox list, the message list, and individual messages as XML (each addressed by a URL using the REST methodology -- XForms works very well with REST). The UI, written entirely in XHTML+XForms, is 300 lines.

There's a little bit of JavaScript used to supplement the basic markup, using Mozilla's XBL to make it output dates in a human-readable format.
(Mozilla XForms ought to do this anyway, and when it does, it will likely ship as XBL with JavaScript and CSS to bind it).

See http://graflex.org/klotz/2006/11

The best way to improve your pictures is to learn something about composition and lighting. If the subject matter is good, you have a good picture.

You want better data? Get a better camera. Ditch that point-and-shoot for a DSLR, or even (gasp!) a film camera. My 50 year old Crown Graphic takes pictures that the very best DSLRs can only dream about.

...laura

At graflex.org we have used Microtek and Epson flat scanners. You do not need a drum scanner for your task.
You might try posting your question at there or on photo.net, or just reading the existing postings there.

And yes, you can still get 4x5 film. Try it out sometime!

Try AJAX with markup only, no hand-written JavaScript, by using the W3C's XForms standard. There are a number of implementations such as the Servlet-based Chiba (which fronts for Dojo and other packages), FormFaces written entirely in JavaScript (no server-side component), entire server-side pipeline systems such as Orbeon, and full client implementations such as the Mozilla/Firefox XForms Extension (just type "xforms" into your Firefox extensions finder), FormsPlayer IE Plugin, or cell-phone capable implementations such as PicoForms and SolidForms.

I wrote an entire webmail reader using PHP for the back end and XForms for the client. It runs in the Mozilla XForms implementation but could easily be made to work in any of the above, which differ mostly in how the CSS works. xmlmail

And for completness, I was an editor of an earlier version of the W3C XForms recommendation.

Resolution goes as the wavelength times the f-number; you get better resolving power with smaller f-numbers. You don't need meters-sized optics because your resolution requirements are driven by the f-number. For the same f-number, the resolving power of a Hubble-sized optic and one on your cell phone camera is the same (same blur spot on the focal plane). The Hubble-sized optic has the advantage of being able to see much fainter objects because it is a much bigger light bucket.

For this sensor, it has something like 10.5k pixels per side. Spread that across 4 inches and you end up with a pixel size of about 9.5 microns. The size of the blur spot is notionally 2.44*wavelength*F/N, so if you want to put two pixels across the blur spot, you'd need an f-number (F/N) shorter than about f/15. Anything greatly shorter than that is overkill, i.e., you don't need anything like f/2 or even f/4. Finding quality lenses for a Speed Graphic that meets this is easy.

It may be at its limit for the number of megapixels but, there's still a lot of things to improve like the maximum color range a digital camera can record. With 16 bits color channel, we would be able to record a lot more informations so we wouldn't be limited as much when we try to capture a high dynamic range picture. There's tools like in Photoshop CS2 to give you the abilities to have high dynamic range but it would be a lot better to have it directly in the camera.

The CCD cameras used by astronomers routinely produce 16 bits per pixel. Most of these are monochrome devices: to shoot a colour picture you must shoot pictures through red, green and blue filters, then combine them.

The key advantages for astronomy are zero reciprocity failure (film loses sensitivity in long exposures; CCDs don't), high quantum efficiency (almost all the photons intercepted by the sensor are noticed) and excellent linearity (you can digitally subtract extraneous light, like city lights).

However, even in astronomy, there is a hard core who still do film. There are many reasons: some people just like the look, others enjoy the craft of wet darkroom work, and so on.

My favourite camera is a 4x5 press camera, a Crown Graphic. It takes perfect 1950s newspaper photographer pictures. And I develop and print them myself.

...laura

I think that many computer applications, and to some extent certain kind of programming, are a little too much like watching TV, and harm your brain rather than enhancing it. Of what's going on today, I think the Make-magazine stuff is probably the most exciting and most likely to provoke actual thought... Kids doing robotics is pretty close to what kids doing ham radio was when I was young. Below is a meandering story of how I got from a 5 year old ham to today, back into ham radio, and reading Slashdot too.

In kindergarten, I remember bringing electrician's hot-side testing screwdrivers to show-and-tell ("Now you just stick this screwdriver into the electric socket and the neon bulb will light if it's the hot side"), and rigging up telephone networks with old handsets and batteries. After having learned morse code at age 5 and gotten on the air under my father's call (he got his license in response to my interest), I finally learned enough to read the whole test and got my license at age 7. Now my kids are about the same age, and found learning morse code to be fun; they talk to each other, and recently had a poster accepted at a peer-reviewed conference, comparing speed and errors in Morse code and typing! (Ok, it was the 2nd grade science fair.)

Soon I got interested in computers, but there weren't any actual ones to distract me; well, there was one in town, and it used punched cards. It was a Honeywell Special 200, the first IBM Clone, though it was a clone of an IBM 1401... Then there were the PDP-8's that were connected to Stanford via phone line for one of the first "computer-aided instruction" projects. I met the guys who maintained the Model 28 teletypes for them and they got their ham licenses after my father and I got ours...

When two-meter FM became popular, I helped establish the first local repeater, probably the only one within 100 miles. We had to do HAAT testing and I learned about altimeters, topographic maps, and government forms... By the time I graduated from high school and went to MIT, I found other pursuits -- PDP-10's, Lisp, classes... I pretty much got off the air. But ham radio gave me an entre into an entire world that wasn't available when I was growing up.

After a few years spent exploring 4x5" photography, I started doing some wireless mobile device work, and poor signal strength led me to get up on the roof and install a 1.9Ghz repeater. I felt a strange familiar feeling, and when my wife said, "I don't care how many antennas you put on the roof," I filed the fact away. When a co-worker shows up with a Yaesu VX-2 two-meter and 70cm handitalki that receives DC-to-daylight and said it was $120, I went ahead and bought it. I'd kept my ham license renewed, and used it once or twice in the intervening 20 years, but I had to re-learn lots of stuff. I wore the HT on my belt (along with two calculators and a slide rule, a hiptop, and a blinking LED pen) for the Halloween party at PARC and won what can best be described as the five-sigma prize...

A bit of web surfing led me to QRZ.com, EHam.net, and of course ARRL, and I found out about a local club meeting taking place that night. So I went with the co-worker, and found a bunch of pleasant nerds, schoolteachers and librarians, firefighters, electronics designers, computer scientists, and other random people.

At the club meeting, a satellite communications engineer told me about recent developments in DSP-based communications that used a PC sound card to modulate and demodulate; my extensive 20-year stint in programming made me think this might be interesting, so I bought a

I think that many computer applications, and to some extent certain kind of programming, are a little too much like watching TV, and harm your brain rather than enhancing it. Of what's going on today, I think the Make-magazine stuff is probably the most exciting and most likely to provoke actual thought... Kids doing robotics is pretty close to what kids doing ham radio was when I was young. Below is a meandering story of how I got from a 5 year old ham to today, back into ham radio, and reading Slashdot too.

In kindergarten, I remember bringing electrician's hot-side testing screwdrivers to show-and-tell ("Now you just stick this screwdriver into the electric socket and the neon bulb will light if it's the hot side"), and rigging up telephone networks with old handsets and batteries. After having learned morse code at age 5 and gotten on the air under my father's call (he got his license in response to my interest), I finally learned enough to read the whole test and got my license at age 7. Now my kids are about the same age, and found learning morse code to be fun; they talk to each other, and recently had a poster accepted at a peer-reviewed conference, comparing speed and errors in Morse code and typing! (Ok, it was the 2nd grade science fair.)

Soon I got interested in computers, but there weren't any actual ones to distract me; well, there was one in town, and it used punched cards. It was a Honeywell Special 200, the first IBM Clone, though it was a clone of an IBM 1401... Then there were the PDP-8's that were connected to Stanford via phone line for one of the first "computer-aided instruction" projects. I met the guys who maintained the Model 28 teletypes for them and they got their ham licenses after my father and I got ours...

When two-meter FM became popular, I helped establish the first local repeater, probably the only one within 100 miles. We had to do HAAT testing and I learned about altimeters, topographic maps, and government forms... By the time I graduated from high school and went to MIT, I found other pursuits -- PDP-10's, Lisp, classes... I pretty much got off the air. But ham radio gave me an entre into an entire world that wasn't available when I was growing up.

After a few years spent exploring 4x5" photography, I started doing some wireless mobile device work, and poor signal strength led me to get up on the roof and install a 1.9Ghz repeater. I felt a strange familiar feeling, and when my wife said, "I don't care how many antennas you put on the roof," I filed the fact away. When a co-worker shows up with a Yaesu VX-2 two-meter and 70cm handitalki that receives DC-to-daylight and said it was $120, I went ahead and bought it. I'd kept my ham license renewed, and used it once or twice in the intervening 20 years, but I had to re-learn lots of stuff. I wore the HT on my belt (along with two calculators and a slide rule, a hiptop, and a blinking LED pen) for the Halloween party at PARC and won what can best be described as the five-sigma prize...

A bit of web surfing led me to QRZ.com, EHam.net, and of course ARRL, and I found out about a local club meeting taking place that night. So I went with the co-worker, and found a bunch of pleasant nerds, schoolteachers and librarians, firefighters, electronics designers, computer scientists, and other random people.

At the club meeting, a satellite communications engineer told me about recent developments in DSP-based communications that used a PC sound card to modulate and demodulate; my extensive 20-year stint in programming made me think this might be interesting, so I bought a

Bring an old Graflex but make sure you chain it to your ankle so you don't lose it in a flood. You need to use chain because it's kind of heavy, and make sure you lock it.

I built an analog remote control computer display meter in April 2001 and demonstrated it publicly: here.
The file date on the oldest version of the index.html.old file is April 21, 2001.

I built an analog remote control computer display meter in April 2001 and demonstrated it publicly: here.
The file date on the oldest version of the index.html.old file is April 21, 2001.

See here. These were the keyboards that provided Control and Meta in EMACS, as you can see in the four keys at the bottom.

SAIL (Stanford AI Lab) and the MIT AI Lab both had them, and the displays were 512x512 pixel green-screen displays that were bitmapped frame buffers on a PDP-11, built out of the first Intel 1Kbit semiconductor memory, attached to MIT and SAIL's respective PDP-10s.

I believe the MIT keyboards were slightly different but I can't find a picture of them, just the SAIL version. The keys felt really luxurious. I'm pretty sure the space bar was bigger. The ESC (27) key looked like a little diamond, and the big roundish key in the top left was labeled "ESCAPE" but was a key that went straight to the PDP-11 -- at MIT you could call the elevator to the 8th or 9th floor with ESCAPE E, buzz the 9th-floor door with ESCAPE D, look at a FINGER display of MIT AI with ESCAPE F, or look at someone else's buffer (there were 16, minus one for finger) with ESCAPE n F. The CALL button in the upper right was also for the PDP-11 and you pressed that to attempt to get a frame buffer. Often during the day we couldn't get one, so we had to work at night. The PDP-10 was faster at night anyway...

http://graflex.org/klotz/hiptop/bot

The CPIA driver works fine with it for Linux, for viewing. Unfortunately, the code to turn the lights off and on has been commented out of the driver due to a buffer overflow.

I've got it running with the RedHat 7.3. I had to merge the driver from http://webcam.sourceforge.net with the drivers in the kernel source and recompile. I tried compiling the driver outside the tree but got bit by a bug in USB link ordering. Once you apply the updates you'll be able to use simple command-line statements to turn the lights off and on.

If you're going to try these patches on RedHat 7.2 or on some other Linux distribution, you'll have to merge the driver on http://webcam.sourceforge.net with your distributions's driver yourself.

For my code and images, see http://graflex.org/klotz/qx3.

Of course, the best thing would be for the webcam.sourceforge.net people and the kernel people to resolve their differences and get the write code for /proc/cpia enabled. Until then, turning the lights on under Linux will be a DIY project.

I made a wallet-sized PDP-11 (see photo) using these tools.

I put the simh PDP-11 emulator and unix_v7_rl.dsk along with the following script onto a CF card formatted as a DOS FAT partition.

set cpu 18b
set rl0 RL02
att rl0 unix_v7_rl.dsk
boot rl0
#boot
#rl(0,0)rl2unix

You have to type those last two lines manually to the PDP-11's boot prompt.

I'm ready to roll with a PDP-11 in my wallet (or, if you include the $9.95 CF-USB (Linux driver) card, in my Penguin Mints container, which matches the black and yellow 48MB Lexar card I got on sale at Fry's for $19.95).

Total cost for a PDP-11 running Unix: $29.90, mints not included.

BTW, the default V7 "root" password is "root" (I ran John the Ripper and it took 0.00002 seconds).