The Dexcom G4 is a continuous glucose monitoring system (CGMS) worn by type 1 diabetics. It has a USB port and Windows only software for downloading and analysing data.
A team of developers / users have created an open source realtime remote monitoring, alarm and glucose prediction system for it. It uses a USB OTG cable to connect the Dexcom to a commodity Android phone, which runs a custom Android app to upload the data to MongoLab. From there, a Node.js powered dashboard gives viewers - typically parents of diabetic children - a way to monitor glucose levels remotely. In addition, a Pebble watch app lets the wearer check their glucose by just looking at their wrist, instead of having to use the bulky Dexcom monitor.
Impressive! And something I could see myself doing at some point, too...
Any idea if anything like this is possible for a Medtronic CGM? Their latest model has very good pump integration but I want real-time electronic data too.
Awesome. I don't have one of these yet, but I'll certainly check this out when I do. The most important thing is to know that the hardware support is there -- to paraphrase xkcd, the rest is software and I'm sure it will happen eventually.
Is that illegal? Will those parents be charged with altering safety devices in case of accident? I'm perfectly aware how good it is, but insurances and labs don't really care about that.
I had a teacher make seating charts in Bingo Card Creator by exploiting the fact that, for the free trial, the random perturbation of the words is deterministic. She put 1 through 25 in the word list, observed where they showed up on the bingo card, then replaced them with students' names to get the desired seating chart. (How do I know this? Because she bought BCC and that replaced the deterministic random seed with an actual random seed, causing the seating chart to break and her to send me a support email the next time she tried printing it.)
Given that one of my Internet buddies runs a table planning software company I probably should have bought her a license on general principles.
as a tangent... one of my former teachers would call register by assigning us numbers at the start of the year, then everyone would say them in sequence. a lot faster than the classical approach. :)
When I did phone support for a barcode scanning company I had one customer trying to use an undecoded scan engine to guide robots across a floor.
A scan engine works by sweeping a laser across the barcode and using the light intensity at different points to determine where the bars and spaces are relative to each other. A modern scan engine will usually output what data was in the barcode, but an undecoded scan engine outputs a continuous stream of data indicating the light intensity received from the laser.
If I recall correctly, they had a bunch of timing marks on the floor that the robot would follow, and they were scanning these using the engine. You could keep a robot on-track in this way because, if the laser is angled relative to the timing marks a large part of the sweep isn't going to contain any marks and so the robot needs to correct. Then it's just a matter of servo'ing down the line.
I would never have thought to do something like that so it was probably the best call I'd gotten all year.
Shortly after launching Cronitor.io, (a dead-simple SaaS solution for monitoring cron jobs), we noticed a large number of tracking events coming from a specific user -- 3000x more than we normally see after a user starts monitoring a job.
The normal use case is to curl your tracking URL at the end of your crontab line, so seeing 25k events coming in over 48 hours clued us in that Cronitor was being used for something different than we expected.
Turns out, the user needed to do a big one-time batch processing of thousands of records and wanted to keep track of how many were done, and wanted to be alerted when the process was complete. By curling his tracking URL at the end of his event loop, he could leave behind his workstation and trust that he would be notified when the batch was done.
Please try and contain yourself. I know you want to take every opportunity to promote your project, but comments like this one come across as blatantly spammy.
As William Gibson put it: "The street finds it's own uses for things."
Back before cell phones, Motorola sold a lot of walkie-talkie type radios. Most of these were 'ruggedized' (encased in big thick yellow/orange rubber padded boot). One day, Motorola engineers came up with a 'better' padded boot they thought would help the radios survive a fall.
But they got complaints from the Air Force. So they investigated..
It turns out that the Air Force mechanics need to 'chalk' their airplanes to track state (bombs loaded, fuel loaded, etc). But the mechanics were always loosing their chalk. So they would just 'scrape' the plane with the (padded) back of their radios, and it would leave a (non-permanent) mark like chalk. The 'improved' padding didn't work the same way, so they asked Motorola to keep selling the old padding (and they did).
(Heard this a long time ago. Really fuzzy on all the details, but I'm sure it's true.)
I sold surveillance cameras when I was in University and had a customer build something similar to a GoPro using a bullet camera, a portable mini-dvr and battery packs. I don't recall the quality being very good, but he was an interesting guy!
How viable is it to offer a surveillance camera installation service while in college? I'm sure there are people who own an expensive house and don't know how to setup an advanced DVR such as ZoneMinder.
That was my theory - people don't feel secure, surveillance equipment can be complicated and so people would gladly pay for someone to come in and set everything up for them. It worked out fairly well for me, but I feel that it only worked well for me because I got lucky and had some good strategic options:
a) Someone I knew opened up a store that sold surveillance equipment. I aligned myself with his store, trained his staff and booked my appointments out of it.
b) After a few months, I started to focus on small business instead of residential installs.
c) I have a combination of strong technical and writing skills. Whenever I finished an install, I would give the client a customized report where I explained what I did, documented what types of cameras we installed, documented the lines of sight that we covered, went step by step through everything she would need to do, and gave my contact information. Those reports became biz dev.
Having a location was especially valuable to me because the surveillance industry is all about trust. Without a storefront, I would have been some guy in his late 20s who worked out of his apartment and knew quite a lot about breaking into places.
I moved into the small business market because there were unexpected quirks with the residential market. To put it gently, there are really three kinds of people who need surveillance cameras. The majority (I hope) own a home and are scared of being taken advantage of. They may own some collectible cars, or they may have a valuable baseball card collection and they want some additional security. Some people have in-home child care and want to be able to check up on the child care provider. Those people are excellent. Another group of possible clients are drug dealers/miscellaneous criminals who have property they want to protect. And, a third group lives in a different reality than I do.
One particular customer wanted to hire me to install hidden cameras in his ex-wife's house. He thought that she was cheating on him and wanted to catch her in the act. And yes, this was his ex-wife. Or, another fellow wanted cameras because he was convinced that the RCMP broke into his house at night and injected him with mind control drugs to keep him addicted to drugs. And several men came in wanting hidden cameras installed in their bedrooms, pointed at their beds. There must have been a rash of pillow thefts going on or something....?? ;)
Long story short, sometimes the residential side just felt really dodgy and I was always very concerned that people were lying to me. And, since the kinds of people who need security are often paranoid (they have a reason), it was very hard to read residential customers. Consequently, I left a massive amount of money on the table. But, it didn't seem worth the risk to me because the storefront and the fact that I had an inventory of security cameras to work with gave me access to the small business market.
Small business is all the fun of residential with some added bonuses.
a) Small business setups are significantly easier to plan than in-house setups. For example, how would you set up hidden cameras in a two storey, 1950 square foot house when a client is concerned that her babysitter is abusing her children? Do you put hidden cameras in every room? If so, do you use all wireless and risk getting into interference/interception problems? Does the customer have the budget for a 16 channel DVR? On and on...
b) Small business owners are significantly easier to deal with. For the most part, they're either pissed off with shoplifters or they're scared one of their employees is stealing from them. This is easier to work through than someone who is concerned about gangsters breaking into his grow op.
c) Small business owners tend to know other small business owners, so if you can get six setups on your own, those six people will help you line up your next projects.
The biggest downside to small business is that if you don't have a storefront to work out of, they can be a very hard group of people to market to.
Long story short, I had some success, but I feel like my particular circumstances had more to do with the success than any particular strength of the market. This is a strong market, but it's a strong market with some problems that can make it somewhere between dangerous and stressful to serve.
Thanks for the in-depth answer, I appreciate it. While this is exactly what I am looking for, I want to know what is the minimum amount of equipment needed to get any jobs. For example the most important are the cameras, but to mount them you would need a ladder. To be able to bring a ladder to sites you need a long van. Then you need experience as an electrician if the house in question does not have the Ethernet cables installed. All these expenses add up to 100k plus 100k for a storefront at a minimum, correct?
The ladder isn't really that big an issue, you can get a decent collapsible or telescoping ladder that will travel pretty easily in a hatchback, small wagon or minivan. Something that will work in most offices with 8-10' suspended ceilings will probably set you back ~$100, for higher ceilings you're going to pay more and probably need a minivan, as well as some muscles to wrangle it around (e.g. Werner MT-26 or Little Giant 12026 which collapsed are ~6ft and $300-500). Having 2 ladders will often make a huge difference in your productivity.
You don't really need that much electrical experience for running low-voltage Ethernet, but you will need some and you need common sense, some basic construction skills, more common sense, some planning skills, and a bit more common sense. Keep in mind that you're going to be putting holes in walls, putting up new faceplates, etc. and you don't want to put them in crooked or leave a mess. Keep in mind that it's one thing for your friends to wonder "What was he thinking?" and a completely different thing for your customers / referral sources to wonder the same.
You'll need a vehicle, but it probably doesn't need to be new. Ladders, call it $1k. Cable fishing equipment, drills, saws, etc. maybe another $1k. Consumable supplies (boxes of cable, jacks, patch cables, wall plates, low-voltage mounting brackets, patch panels, etc.) will easily take another few thousand (particularly using plenum cable). You'll want to keep some quantity of equipment (cameras, etc.) on hand to do immediate fixes/replacements for your customers unless you have a good local supplier, but you may be able to avoid keeping a large stock as long as your orders have enough lead time (and you have the ability to purchase as needed). Insurance, licensing, etc. will chew some, but even so I'd say you're still looking at well under $100k to start, and certainly not $200k unless you're in a very expensive market.
With questions like those, though, I'd recommend spending at least some time working for someone already in the business.
While any specific use isn't very interesting, an honorable mention should go to the army's P-38 can opener. It was the minimum amount of sheet metal wrapped around just barely enough of a cutting surface to open a can of C rations.
As they were cheap and easily available due to being standard in a pack of rations, they ended up being used for... everything. The link below mentions a few of the simple can opener ended up being used, if you want a few examples and has a picture of the tool.
I write software used by organizations that help individuals with developmental disabilities. Sometimes the organizations have to babysit the individuals (day programs), or have them help with work, while their guardians are out working during the day.
I recently got a call from a staff at an org that was having an individual help with her workload by using our software - she was honestly complaining that an update we had pushed out had made things so much more efficient, so that the individual they were babysitting/keeping busy/working finished their work so much faster that they didn't know what else to do with her. She wanted us to revert our update back to a state of being less efficient. (Of course we responded that we can't support that type of organizational workflow...)
We had no idea our system was used as a tool for keeping people busy for the sake of being busy and not causing trouble!
A company I occasionally do support for has a profitable legacy system that runs on 486 or lower PCs. Supplier of product out of busiess. One PC died and they couldn't get any more. I found a manufacturer of current PC104 boards with a 486 CPU. moved the software across to those. It nearly worked but would occasionally crash. I asked a genius friend to have a look at it. He got out his trusty DOS debug floppy disks, footled around for a few days working out what was wrong (slightly incompatible implementation of serial port in the SoC on the PC104 boards, plus crap code in the original system), hand-assembled a workaround and wrote it in to blank space on the original executable, and presto! the system has been running reliably on the new hardware for a couple of years now.
I also thought the site was going away courtesy IKEA legal, but apparently based on user feedback, IKEA is reconsidering. The original legal action was voted up here on HN last month but the follow up not so much.
I thought Ikea specifically didn't want to stop the site from posting useful and interesting ways to use their furniture, they were just required to protect the use of their trademark in the name of said site.
A friend's family owns a plastic manufacturing facility. They buy a lot of Coca Cola, for cleaning their manufacturing equipment (of grease, oil etc) after each production round. Much cheaper than the industrial cleaning agents.
A team of developers / users have created an open source realtime remote monitoring, alarm and glucose prediction system for it. It uses a USB OTG cable to connect the Dexcom to a commodity Android phone, which runs a custom Android app to upload the data to MongoLab. From there, a Node.js powered dashboard gives viewers - typically parents of diabetic children - a way to monitor glucose levels remotely. In addition, a Pebble watch app lets the wearer check their glucose by just looking at their wrist, instead of having to use the bulky Dexcom monitor.
http://nightscout.github.io/posts/01-quickstart/