A new member of the colorful family of objects that populate urban China’s streets is spreading gradually across the country’s largest cities – Beijing, Shanghai, Shenzhen. Notable is that such an advance has taken place around libraries, and that there is sufficient perceived demand/interest for enabling people to check out books 24 hours a day (or that the company perceiving said need is owned by the family/associate of someone influential*). In that vein, I can’t help but puzzle at the nature of the dark matter that has allowed these not-so-cheap machines to take root across both urban (Beijing, Shenzhen, Guangzhou) and rural (remote and resource-starved Hui Autonomous Region in Ningxia Province) contexts, each with a very different set of needs and each bearing vastly different reputations for serving their citizens.

Such a machine intrigues for its theoretically duplicate means of surveying the inventory, seemingly to make the device more able to be easily engaged. A user could casually search through (at least part of) the inventory on display with a passing glance through the large transparent pane on the front, and access to the entirety of the inventory is possible through two additional means; a keyboard that appears when prompted on the touchscreen located next to the front pane, and a physical keyboard composed of cold, hard-to-depress keys that resemble steel Chiclets and come with a delay between pressing a key and seeing it appear on the screen sufficiently lengthy so as to offer time for a string of several (short) obscenities. While the front display/keyboard is at torso height for an approximately 6-foot user, the physical keyboard on the side is seemingly positioned such that it is just out of comfortable reach for users who rely upon a wheelchair for mobility.

Users verify their credentials with a pass of a library card over an RFID reader (which the company also makes, and which come with their own unique set of dark matter constraints, as outlined by Adam Greenfield in an excellent post of his). If you happen to not show up with such a card and see a title you’d like, a you may be granted a library card in exchange for feeding your credentials to the machine. Once done with your book, bring it back to the machine from which you rented it and deposit it into the slot (a physical library branch also works, if you happen to know where to find one). Every few days, an employee of the library restocks the machine with new books and loads the plastic bins of returned books into their van to return them to the nearest library branch for re-cataloging. A great deal of moving parts to keep lubricated, from a service design perspective, but the library employee with whom I spoke while he was restocking the machine considered the additional time and labor involved worthwhile for the library.

Two questions:

1) How does this system learn? For this particular machine in Shenzhen international airport’s, many of its 400 volumes are skewed towards city and country guides, “survival” language lessons, and travelogues (with the odd volume on programming in Python and the study of speech therapy). If a book is nearly always rented out from a particular machine, would more of those books be stocked in that machine? What are the differences in reading habits of an passengers in an airport in Shenzhen and residents of several blocks of identical apartment buildings in a quiet Beijing suburb, and how do these machines (and their managers) respond in kind?

2) What is the lifespan/utility of the “exhaust” of said system? A reader’s destination could be fairly easily inferred from examining the reading material they’ve chosen for the journey from the above library machine in Shenzhen. Far more (and far different) information could be used when examining the reading habits of citizens in a potentially unstable area such as Ningxia (or Xinjiang).

* this company