Interesting concept.

I think the "Project Log" should be in the "Project Details" instead, though... as it seems a bit vague until getting to the log regarding kids casting off toys, which makes a lot of sense.

Similar to @Duane Degn's analysys, many of these concepts could be implemented sans-machine... e.g. each week a teacher would hold an auction of the children's unwanted toys... 

But, the vending-machine idea is groovy. Before the kids' toys idea, I had a visual of something like a vendor-bot in a maker-space... Don't need that Arduino anymore? Exchange it for someone else's discarded Wifi gizmo...

Tickets... Thermal-printed, with a barcode or QR code, seems a heck of a lot easier/cheaper than RFID ;)

Good point E, I need to re-work the description some to incorporate that and some of the stuff from the conversation with Duane as well.

Was on tech sabbatical over the weekend so didn't make much progress in the physical world, but did do some thought experiments that I plan to turn into something real in the next few weeks.

Yeah I go back and forth almost daily on the card-vs-printed (or other) thing for the tickets,   At the end of the day most of the internal aspects are the same so it will probably come down to whatever I can get my hands on first.  That said the design should be able to accomodate either so you could select whatever method best suited the local application.

Jason,

I'm not sure if you're aware of the inexpensive MRFC522 MIFARE 13.56MHz RFID readers or not. These devices are surprisingly inexpensive. They might be a good option to keep track of credits. I suppose it would be possible to keep the credits on the card  (I think this is done with some parking meters)  or you could use the cards to identify the users and keep track of the credits in some sort of database.

I know there's lots of code available for these readers. I ported some Arduino code to use with the Propeller (my personal favorite microcontroller). The Propeller code can be found here:

http://forums.parallax.com/showthread.php/154851-Propeller-Object-for-NPX-s-MFRC522-MIFARE-13.56MHz-RFID-Readers

I'll likely move the code to my GitHub in the future.

I'd like to learn more about the MIFARE cards & readers.  I did  little reading on them awhile back but I wasn't able to get a clear idea of which readers were compatible with which cards/fobs/tags/etc.  I found some that did look very inexpenive, but I wasn't sure if they would work or not, so I'm excited to talk to someone who's worked with them :)

I think something like an RFID or magstripe card has a lot of advantages in simplifying the device so if it can be used for a reasonable cost I'd prefer to go that way.  Do you have a recommended card/reader combination that you prefer?

One of my favorite "cheap parts from China" stores is ElectroDragon. They charge a little shipping but the shipping is generally faster than "free shipping" sites.

Here's a reader I think would work.

http://www.electrodragon.com/product/mifare-rc522-rfid-card-readerdetector-ic-card/

I purchased one myself but I don't think I've used it yet (I've similar ones off ebay). I also purchased these cards:

http://www.electrodragon.com/product/5pcs-mifare-one-1-ic-s50-13-56mhz-contactless-card-rfid/

These cards are broken up into sections of "sectors" and "blocks." 

IIRC, there are 16 sectors each with four 16-bit blocks. I think only 3 of the 4 blocks can be used for data and the other block is used for the sector's passwords.

Of the 1K of EEPROM memory in a card about 25% (IIRC) is used for card overhead.

Each card has a unique (I think) 32-bit serial number.

There's plenty of room on the card to store information like how many credits have been earned, the ID number of the last item shared (so the machine can check to see if it should add credits) and information about their past activity with the machine.

The information on the card can be secured with passwords (I think there are two different passwords for each sector).

I think these cards can be "hacked" relatively easy these days so you won't want to use them for high value data but I'd think they'd work fine for a neighborhood sharing library.

Thanks for the links Duane, I'll definately be checking that out!

Cool idea, I like the credit system.  I don't see why you wouldn't make these internet connected eventually, but they certainly don't have to be.  People like spur of the moment window shopping, I could see it being similar to Redbox, I still can't believe how popular that is considering Netflix exists.  But people like making a split second decision to get a DVD when they are walking out of the grocery store or convenience store.

What's your plan for dealing with junk nobody wants piling up?  Would be interesting if you got -1 credit if your item was never taken and had to be disposed of.

I hadn't considered negative credits Alex but that's an interesting variable to play with.

I'm not opposed to the machines being connected, I'm just sort of deferring thinking about it too much until we get the basics working.  There are some very cool low-cost ways to do the machine-to-machine/IoT stuff now but I wanted to make sure I didn't unintentionally design the device with a dependence on connectivity.  One thing I've wanted to do with our Little Library is have a way for it to notify people when a new book is available (either directly or via a feed of some sort) but it was hard to come up with a reliable way to do it that didn't require constraints on how the library worked or raised some potential privacy issues :)

I think you're doing it right, often the projects that get thrown up here are so over constrained with features.  Get the basic idea working and build on it after that.  Good luck with it!

    I don't understand why you'd want or need the added complexity of having "credits." If someone wants something in the machine, what's to stop them from placing a piece of trash in a compartment? If the machine is going to dispense things people value, it would need some way of assigning value to the items placed in it to prevent abuse of the system.

    Why not just open shelves?

    I recalling seeing a news story about a place near a dump where people placed items they no longer wanted but they thought  the items didn't belong in the dump. Anyone could take these items. Unfortunately this "reuse" area of the dump was closed down. The people running the dump didn't want to take care of what they considered trash.

    I really think something like this needs some human intervention. IMO, secondhand stores already fill this need. They charge a small price for items which pays the employees wages and other costs associated with storing items awaiting new owners.

    If one wanted to receive something in exchange for the item being discarded then one could turn to a pawn shop. The pawn shop gives the person "credits" which in turn can be used to purchase other items in the pawn shop. In this case, money is used as credits.

    If one wanted to bypass pawn shops and secondhand stores (such as  Goodwill or Deseret Industries), then perhaps the item could be listed in Craig's List with the tag "free stuff"? 

    I really didn't want to come across so negative, but it sure seems like you want to make some sort of automated pawn shop (without the loans).

Hi Duane, thanks for the feedback.

There are some subtlties to the credit system that I probably did a poor job explaining, but I'll give it another try :)

The earliest ideas we had were along the lines of the open shelve system you describe.  The original motivation for the credit system was to help reduce the amount of human intervention, but it might not be obvious how these two are related.

Credit is only dispensed after an item has been "purchased" by another user, so in order to receive something from the system you need to leave something behind which is valuable to another user.  This means you need to come back and check to see if your ticket has earned any credit.  This "coming back" is important, because it establishes a habit of visiting the device which encourages further use both for deposits and purchases, which keeps goods flowing and increases the value of the device to the community.

There's another impact that the credit system can help with by providing a "rate limiting" function on the device.  We have a Little Free Library in our yard where neighbors can pick up and leave behind books in a free and unmetered fashion and this works well, so we decided to try an experiment along the lines of Recyclotron by placing some small toys and other non-perishable foodstuffs in the library and indicating that they are free and that others are free to leave things they do not need.  This worked to some degree, but there was a strange effect where occasionally (perhaps once a week) someone would clean-out all these items and then throw them down the street.  I don't think this was anything other than mischief, but by having something like the credit system in place the amount of waste this kind of activity has is limited to only items that have "depreciated" to being available for free.

That said of all the aspects of the device the credit system is the one that we think needs the most research and testing.  We've spent a lot of time thinking about how best to implement it and mentally modeling how it would be used, but I believe the best way to know how it should work is to field test the simplest implementation and then make adjustments based on data and feedback from use by real people.  So the plan is to implement a device capable of the credit system described initially, but exactly how that works will be sorted out as we go along, and the way that credits are dispensed, collected and decay are all adjustable based on what works (and can even be adjusted on a per-installation basis to work best for the environment the device is installed in).

Recyclotron is certainly inspired by Craigslist, resale shops, etc. (I've described it as Craigslist in a vending machine :).  The key differentiator is that there are substantial start-up and operational costs involved in providing a shop, which puts limits on where these shops can exist and how accessible they are.  Recyclotron can fill these gaps by providing an alternative with a lower up-front cost and almost no ongoing operational cost.  It's also something that can be integrated into existing environments which I think increases the likelihood of casual use (as opposed to a place you only visit deliberately).

The other downside to existing resale systems is that by their nature the rely on using real currency (necessary to pay employees, rent, etc.).  The advantage to an abstract credit system is that you can tune it to suit the users in ways that you can't with real currency that need to be accounted for (for example, the credit system could pay out 1 credit for each item even if the item's value decays to 0, if that helps keep the system useful).

That said I don't see Recyclotron competing directly with existing resale systems but applying the principle to environments where resale shops are not practical, at small scales and in a distributed fashion, and available to people who don't have access to currency.  At the same time it provides physical distance between buyer and seller as well as a fixed physical location which provides more safety than the one-on-one exchanges needed for purely virtual systems like Craigslist or Freecycle.

Thanks again for the feedback and I'll work on incorporating some of this into the project description if you think it's helpful,

- Jason

  Thank you for your thoughtful reply. Your description of your experience with a neighborhood lending library was very interesting.

  I think you've managed to change my mind about this project.

  I'll start following your project. I hope things go well.