Idea Blog
Welcome! The purpose of this web page is to record and develop my creative ideas into things that someone could actually use or turn into a product. Unfortunately, my profession (lawyer) does not permit me to explore all of the creative output that I have, mainly because of time commitments. I suppose one reason I'm doing this is to see if there is any interest from people in the cutting edge of the fields involved, because that could lead to contacts in academia or companies that could lead to some type of side gig. All intellectual property herein is licensed under the Creative Commons license for attribution non-commercial share alike uses. Commercial or other users should contact me at andrew[delete-me]@andrewwatters.com for other uses. With that said, please enjoy.
2009-11-20 10:41 PST - Flashmob Deployment of Mesh Network-Enabled Reactive Government Agents
This is going to be awesome. I wish I could document ideas like these for a living instead of simply writing down the titles and never getting around to writing the full analysis. I am inviting any think tanks or government contractors that might stumble upon my page to send me a message any time...
2009-03-12 10:12 PDT - Hyperjections™: Graphical Hypertext Refutation of Written Statements
One of the challenges faced in the practice of law is intentional vagueness. My experience has been that people purposely make vague, plausible, but untrue statements to create false impressions and/or lie, while giving the appearance of being forthcoming, sincere, and reasonable. The statements are vague so the speaker can later say "oh, I didn't mean that- what I meant was...." In other words, wiggle room to "clarify" (meaning correct) what the speaker said if it is later discovered that the speaker was lying or failed to disclose something important. In this way, statements in law practice (and politics) often have the air of a public relations official trying to deflect negativity from his company. Consider the following apalling quote from an excellent New York Times article (background is provided first):
Bob Graham, a former Democratic senator from Florida, who as the chairman of the Senate Intelligence Committee underwent a briefing similar to Ms. Pelosi's about three weeks after hers, sides with the speaker. He said he recalled a "bland" session.
"I do not have any recollection that day of there being a discussion of something that would have been as neon as waterboarding or other torture techniques," Mr. Graham said.
Got all that? Pretty good example of an intentionally vague, misleading statement that allows Mr. Graham to give false impressions and/or lie without penalty if he is later caught. Here is the hyperjection of his statement:
PDF (208kb).
Perhaps he could have said instead:
"I believe waterboarding was discussed, because it was on the agenda. However, I do not recall specifics because I intentionally don't take notes, so that no one can subpoena my notes and later prove I was told something that makes me look bad."
Had Mr. Graham made the statement above, it would allow the subject organization to bring out people who suddenly had conspicuously clear memories of what was discussed, so the subject organization could fill in the gaps.
I'm inferring that statements like this are planned out, discussed, critiqued, revised, and approved, rather than made contemporaneously and with sincerity. They just slide on by much of the time, perhaps because it takes so much effort to refute what is said.
2009-08-29 7:29 PDT - 3-D Tracing of Financial Transactions
2008-12-25 04:35 PST - Helical Analytics
This idea relates to condensing a massive amount of information into an intuitive spatial paradigm that is easier to work with than a lengthy 2-D chart or linear approach to storing information. Consider the plight of a person using the program analog to look at website traffic. Although analog is a great program, it's still a daunting task to glean meaningful information from a high-traffic website's data logs. Sure, there are web data analytics tools that offer, for a price, a number of different ways of looking at data logs including beautiful charts. But what I'm thinking of is something I don't believe has been done: an information management tool in hardware-accelerated 3-D space that keeps track of a massive amount of data in real time. Kind of like the LGL library only zoomable, flyable, and live-updating. This kind of tool could be for companies with very popular websites like nytimes.com. The graphical representation in the program consists of one or more than one clockwise single helix composed of a cloud of discrete data points rotating counter-clockwise in 3-D space. Almost like a galaxy of stars rotating in three dimensions instead of two. The top of the helix represents new data, which fades in and progresses toward the bottom of the screen and fades out as it ages. Each entity/point is a rollover item that shows thumbnails of accessed webpages or other detail information regarding the web user and subject matter of his accesses. As the items rotate downward, the detail information stays open if desired and the administrator can choose to see what else that particular web user is doing, which results in highlighted points on the helix that can be separated out into an independent diagram as needed.
[Update 2009-03-09 18:01 PDT- I found a solution for this where I least expected it- a screensaver in Linux that shows spiral rotating blocks. I will do a movie of this so you see what I mean, but here is a screenshot:
]
Refinements would include automatic flagging of interesting data points for administrator review; adjustment of speed of rotation to deal with staffing resources or increases in data flow; changes in geometry of the graphic to reflect user preferences; automation of reporting features for various scenarios; a fixed number of points to reflect existing resources rather than an event-based system; assignment of data points or sets to other staff for followup.
This tool is possible with today's technology, namely session management available in most or all existing web application platforms. The tool could be extended to probably any area where massive amounts of real time data are generated, such as military intelligence; financial transactions and tracing, auditing performance of employees by tracking their network accesses by protocol type; organizing document production and written discovery requests in complex civil litigation; tracking telephone calls made to the 911 system; immigration processing; supply chain management; parts management for complex industrial machines or electronics; and any number of applications with massive datasets that can be organized in a 2-D chart but might be better suited to a 3-D representation.
2008-12-30 13:25 PST - 3-D Interface for Search Engine Results to Improve User Information Processing
Much as I like google, I wish I could see more than 5-6 results at the same time. I also wish there were a way to browse the results and still keep all the results in view. I also would like to be able to organize the results in a different form than linear descending by relevance on my computer screen. Enter a 3-D Tetris type interface where the user view is from the bottom of a cylinder looking up, and results fall toward the user while self-organizing into categories specified by the user or system, and attach to the walls of the cylinder. The user can select any of the results at any time and suggested top hits have some indication that they are more important than other results. Less important results are further away.
[Update 2009-03-09 17:59 PST: And here is a great sample- from a screensaver in Linux of all places. Each of these blocks could be an image or search result, and the whole thing could be in motion with adjustable speed and rollovers. This would also work well for the Helical Analytics idea shown above.
]
2009-05-18 20:37 PDT - The Platoon as a Swarm: Peer-to-Peer Biofeedback Tactical Network
[2009-06-26 14:35 PDT - A shout out to recent visitors from the U.S. Army; it would be great to hear from you about this.]
Overview
Rather than going on the radio or having to look at a GPS device to see where comrades are, what if a soldier could feel others' presence in real time? How about feeling the presence of hostile forces and the paths of incoming fire? If each peer in the network knew where his nearest neighbors were, he would only have to adjust his movements to theirs with the confidence that the "swarm" as a whole would take a natural path through any terrain or obstacle, with an optimum amount of space between individual soldiers. In addition, safe paths, enemy positions, directions of incoming fire, and all manner of other information could be "tagged" and then "known" by the whole unit even without visual contact. Because the brain is malleable, sensory inputs such as electrodes in a shirt or bodysuit would be used with training to give the wearer a true sixth sense, supplementing the others. Because the system is designed to recognize biological limitations in sensory input and processing, the system can facilitate a soldier's natural ability rather than forcing him to adhere to a Land Warrior-type system of menus, cursors, bulky appendanges, computer screens, and other elements that are not so useful to the lower brain under extreme stress.
Mechanism
A peer-to-peer frequency agile wireless network detects range and direction to nearest neighbors and is refreshed continuously. Every peer then knows the information possessed by every other peer. Position information is transmitted to the wearer by way of a bodysuit with distinct areas of sensory input that have different meanings. Example, the back is devoted entirely to giving the wearer a picture of the positions of comrades; a feeling of light pressure indicates the presence of a friendly. The further toward the outside the pressure, the further away the friendly is. The center approximates the position of the wearer. A pinching sensation indicates that a friendly has been wounded; a line being traced across a portion or all of the back indicates movement of a friendly. The chest is devoted entirely to providing a picture of enemy positions and incoming fire; the network uses audio or other means (e.g, connection to a Boomerang device) to plot the trajectory of incoming rounds back to the source and makes all peers aware of this. The legs are devoted entirely to providing direction on where to run/walk; the arms are devoted entirely to aiming, when directed by a ssquad leader, at unseen enemy positions visible to one of the other peers.
Achievements Possible
Everyone knows who is in trouble and also who is responding to assist.
When someone calls on the radio, position is indicated.
Real time changes to formation to optimize the distribution of soldiers even when visual contact is lost.
Mapping assistance from drones who also mark enemy positions.
No sensory overload from trying to shoot, talk, and listen to the radio at the same time.
Immediate Applications
Full speed running in total darkness even if night vision goggles are lost or disabled.
Full speed running through minefields and other dangerous areas.
Human wave assaults.
Adjustments to formation by commander with bird's eye view of unit and map overlay.
Commando raids on entire camps of enemy soldiers.
CQB of a known floor plan building.
CQB of an unknown floor plan building.
Search and Rescue.
Professional Football or other team-based sports.
Key Improvements on Existing Technology
The system requires minimal power because there's no display or other battery-intensive technology. Instead of an unwieldy goggle-based system that floods the Land Warrior with visual information he can't necessarily process under fire, this is a way to enhance the soldier's senses in a more natural way. A very modest computer on a chip could perform the necessary calculations. There would be no moving parts to lose or break. The system would be easy to manufacture and repair. With burstable wireless transmissions over short distances, we're talking power requirements akin to a bluetooth wireless headset. In fact, bluetooth would be a great wireless platform for the prototype.
Key Challenges
Jamming. Actually, using Ultra Wideband could solve that and also prevent detection by a hostile force. A peer whose suit is disabled or malfunctions would probably need to drop out of the formation or perhaps bring up the rear.
Inspiration from an article on sensory augmentation. For a visualization of what is possible, check out this cool video:
I'll put some more info up once I have a better idea of what to do with this.
2009-05-15 09:50 PDT - Mobile Litigation Support Unit
Concept: outfit a Dodge Sprinter van as self-contained home away from home for the attorney who needs a mobile work center out at court. With dedicated wireless internet, a VPN back to the office network, printers, scanners, computers, supplies, workspace, and so on, this would be the attorney's version of a mobile command post. One cool feature would be using the van's internet connection as a backbone for laptops actually in the courtroom, which don't get secure internet. In this way, one could have a secure network connection back to the office, make changes to pleadings or other documents, print them out to the van, and have them brought in within a few moments in final form. And many other possibilities. Other names in the running included "Litigation Assault Vehicle," but that offends the set that fancy themselves paragons of civility, so to give the appearance of reasonableness we could call it the "Collaborative Law Workshop."
2009-05-07 13:35 PDT - Semi-Autonomous Dirigible-Based Air-to-Air and Air-to-Ground Missile Platform
I don't know where I come up with this stuff, but I'm not complaining. Check back soon.
[Update 2009-05-10 - Someone in the Air Force came up with a similar idea but this one will solve the issues stated in the Air Force concept :)]
2009-05-05 18:19 PDT - Elimination of Database Server Query Lag in Web Applications Using PHP Output Buffering and Linux Command Line Utilities
This is my best computer project ever, resulting in a 1000% increase in performance of my law school website. This is an alternative to the memcached system. Portions © 2005 University of California.
2009-03-16 06:01 PDT - Use of Facial Morphology/Biometrics of Serial Murder and Kidnapping Victims to Infer Physical Profile of Perpetrator and Future Victims
We've all done it before- we see someone who looks a lot like someone we know, and we accidentally say hello to them before realizing it's the wrong person. Or, perhaps the next significant other looks a lot like the previous one. DARPA is researching ways to recognize wanted individuals based on the distinctiveness of their walking gait. Questions arise: are there basic patterns of human morphology that, given a large enough sample size, reoccur? Does everyone have a "type?" Do victims of a serial killer share a body shape and/or facial morphology that produces a powerful enough physical response in the perpetrator that he seeks out the same distinctive shape and looks without even realizing it? If the answer to the last question is yes, I believe there is a way to not only develop a physical profile of the perpetrator beyond the basics, but to predict which potential victims are most at-risk and warn them in advance of a serial killer working in their areas.
There is apparently a study involving prisoners' responses to what physical factors were most likely to make a woman a victim of violent crime. I'm going to review that and put up comments here. Another article I found out of curiosity (please, not morbid curiosity- more like in an effort to understand this behavior) in the field is National Center for the Analysis of Violent Crime, Criminal Investigative Analysis: Sexual Homicide (1990).
Someone I know worked on a pending study of female facial morphology. The theory was that women with highly feminine "neotenous" faces (who are perceived as more warm, supportive, and sexual but less intelligent) have lower self-confidence but equivalent performance in a series of tasks than do women with less feminine faces (perceived as more dominant, cold, and intelligent), due to socially reinforced self-doubt. Also, the theory was that the husbands of the more neotenous looking women report being more happily married than men married to women with less feminine faces. The study confirmed the first and the second theory. It occurred to me that one could take photos of, say, Fortune 500 women CEOs and predict how happy their spouses were, which could prove the theory. The next step was collecting the husband's biometrics to see if a particular male facial morphology correlates with a particular female facial morphology. If there is a statistically significant correlation, one could predict with some degree of confidence what the husband of a woman CEO might look like. Consider the following photo of Ebay CEO Meg Whitman and her husband:
I wonder how happy her husband is?
Consider a different couple:
My first reaction, of course, is "what could she possibly see in that guy?" But anyway, Ann looks like Barbie, and Bob has profoundly masculine, almost Neandertal features. Maria Shriver looks somewhat like Ann Coulter, does she not? Based on the two sets of photos, I suggest Ann Coulter will marry someone who looks like Arnold Schwarzenegger.
So, what to do with this information? Consider the cases of two women missing from Vermont and New Hampshire:
Brianna Maitland and Maura Murray both went missing in Spring 2004, within 45 days of each other, within 90 miles of each other, while driving their cars alone on rural roads in Vermont and New Hampshire. Their cars were found abandoned in what appeared to be staged accidents. That's a lot of commonality to be coincidence. I think it's safe to assume the same person kidnapped them.
When I first saw the women's photos, I immediately noticed that there was just some intangible way in which they were similar and I thought they were equally attractive. Perhaps the shape of their eyes and distance between them. To me, the above photos appear so similar that they could be the same person. Note the shape of the nose (especially the bridge of the nose and brow area), cheekbones, chin, forehead, eyebrows, ears, and face in general in each photo. In addition to similar facial features, they are both thin and athletic. If full length photos of Brianna were available, plus the exact measurements of each woman, one might develop a physical profile or perhaps a simulation of their gait.
The question is, what male facial morphology and other biometrics would present a statistically significant correlation with physical attraction to these women? Here's the plan: create separate photo collages of the victims so that all angles are represented, and if resources permit create a 3-D model or video simulation of the victims walking. Go to a college Psychology department and conduct a test protocol on male students. Seek out the men who self-report strong physical attraction to both women and also respond physically (e.g., microexpressions captured on video, pulse rate, posture, etc.). Follow up with the best matches and collect their biometric data. For later research, conduct a behavioral interview. Compare the facial features of the best matches and create several profiles. E.g., "men with super-orbital ridge larger than 10mm, chin that protrudes beyond tip of the nose, jawline that extends 50mm or more below ears, eyes that cover 4-5% of their faces, and athletic build respond most strongly to both photo collages." Take the profiles to investigators and sketch artists. If the serial killer is still operating, canvass the community and warn local women with biometrics and facial morphology similar to the victims that they are most at-risk and show them what the perpetrator might look like without disguises. If no test were done and only victim information were available, likely victims could still be warned.
It would also be interesting to examine the victims' and perpetrators photos in known cases of serial murder and kidnapping in which the victims have features most similar to the present ones. Are Ted Bundy's victims more feminine looking than those of Jeffrey Dahmer? How does the facial morphology of the BTK killer and his victims compare?
Can someone use sex offender registration photos and databases to conduct preliminary research in this area? I know it sounds morbid and I don't know whether this is close to any research or even possible, but how could I even ask for access to something like that without sounding like a nut?
2009-03-12 10:11 PDT - Visualization of Theoretical Fields of View Among Points in Multi-Dimensional Array Representing Terrain Features Using gd Graphics Library
Coming soon. Portions © 2005 University of California.
2008-12-30 13:47 PST - Electromagnetic Field Visualization Using Remotely Operated Probes
There are a number of webpages and applications devoted to analyzing stationary EMF's. See here for example. But what does the blast of an electromagnetic pulse bomb look like and how quickly does it increase/decrease in intensity? Ever been curious where the dead spots in a moving cell phone jammer's coverage are? Ever wonder where there could be a safe path through a minefield that uses EMF to detonate mines? Like in Twister (1996), send a cloud of probes in to record data on the field by flying through it, falling through it, sitting on the ground while the source flies overhead, etc. Use the field and position data from the probes to reconstruct field strength at each point and generate a visualization of the field. How about a Bluetooth P2P network among the probes for extra accuracy?
2008-12-30 11:09 PST - System for Automatic Detection and Tagging of Vehicles of Interest in Visually Augmented Highway Patrol Cruiser
[Update 2009-05-22 16:41 PDT: The Northrop Grumman Joint Strike Fighter, a state of the art jet coming out in the next couple of years, has this system on a much better and more complicated level. See, this is how I know I'm on the right track with these ideas and this blog :)
]
Problem: when the authorities put out alerts on particular cars, it seems to be a matter of chance whether the car (and the suspect inside) is found, if at all, even though all criminals who use a car have to drive it away from the crime scene. If they don't dump the car, then at some point someone will drive it again.
Solution: turn every police car and city vehicle in the state into a mobile searching device that automatically detects, tracks, and tags vehicles of interest that are within range. For enforcement personnel, give the driver a head-up display or reflective windshield that provides an overlay indicating vehicles of interest with a reticle, as well as information on range, speed, etc. Link it with map software to route the nearest officer to the nearest priority target automatically. The military has software like this to detect enemy tanks on the battlefield, so I know a civilian application is possible. For maximum effect, track speed of all vehicles within range and display attributes such as whether the registration is valid. I will do a mockup of the HUD some time soon.
Implementation: radar gives general direction and shape information for the surrounding environment to the onboard computer. Use computer vision to recognize the general shapes of cars as opposed to uninteresting objects, so that optical scanning can be directed to specific areas of the field of view to be more efficient. (Incidentally, I read in a DARPA presentation that the military wants software to recognize people by the distinctiveness of their gait...wow) Read the license plates off cars optically and match them against a cached list of wanted vehicles. Use a laser for range and speed information, and draw appropriate icons on the head up display to notify the driver. Send the results to a central database and notify peers in the area.
2008-12-30 07:54 PST - Bloodborne Medical Nanorobots
This is back from 2005 when, one day, it just popped into my head: nanorobot "spores" that act as cancer cell assassins but leave normal cells alone. They would use surface proteins to identify the cancerous cells. Although it turns out someone had already thought of this in 1999, which I didn't learn until reading the January 2009 issue of National Geographic, I think it's wicked cool.
Update 2009-05-26: A much better solution is to modify a naturally occurring virus to attack cancer cells- it's good to know there is progress in this area.
Update 2009-01-12: someone has done this in a better fashion. I didn't realize this was even possible, but it's so cool!
2009-03-29 9:10 PDT - Purple Rain Prince Tribute Guitar
I'm a Prince fan. My favorite song is "Purple Rain." Back in 2006, I wanted to learn guitar, and I wanted a guitar that would let me play "Purple Rain." So I designed and built this custom Prince tribute guitar from components, some of which were special orders. Purple Rain features the same EMG active pickups used by Prince in his Cloud guitar as well as his newer powder blue Stratocaster. With maple construction, the guitar sounds just like the Cloud, but with the Strat design and gold hardware it has more modern looks. The paint is high quality BMW Interlagos Blue automotive paint, which coats the whole guitar. It's a beautiful purplish blue metallic finished in automotive clearcoat. The neck has symbol fret markers in metallic silver, and the headstock has the famous Purple Rain logo. Overall, this is my best creative project because it actually came to fruition and I received positive feedback from real musicians who demoed it. I especially like the gold-plated Original Floyd Rose tremolo and matching Schaller tuning machines. Perhaps the most frustrating thing for me is that I can't play guitar, and never will because I don't have time! It's too bad because I could play parts of the Purple Rain guitar solo (see here for my favorite version of the solo) when I was taking lessons, and it sounded great.
My only regret with this guitar is that I should have hired someone to do symbol fret inlays instead of painting them on. They are a little sloppy and because the guitar is clearcoated now, I can't change them. On the plus side, the clearcoat is really durable and there is no visible wear from the strings rubbing against it. It's also a slick, fast neck, especially after I waxed and buffed it like a car. The paint has just the right amount of purple in it- I didn't want a straight purple guitar because, well, that would be a little much. Purplish blue is enough.
2010-05-30
a short story I wrote