You can buy that cool tablet today, and its useful life is probably three years on the outside. Something new and cool will be available in 2014 (no pre-announcements here, just predictions) and you'll want to buy it. Perhaps you'll use a vendor's trade-in program to do something with the old one -- after you've conveniently moved the data to your new machine. We hope.
There's a huge opening for someone to get rich, developing a usable taxonomy of parts and materials so that products can be safely and profitably devolved. The way you do it is clear: find a method to describe parts in such a way that they can be taken apart and recycled or safely disposed of. The avalanche of tech products is unlikely to stop, and we expect even less time with them before the new thing arrives to tempt us.
You bought. Someone now has your old machine, with its data removed. What's done with it is then, is something ranging from devolution to landfill fodder. Inside the derelict are a number of precious metals, and depending on the battery technology, a lump of lithium, nickel, and/or other metals. Many smaller bits inside will become reduced to smaller and smaller bits until they're either disposed of in a pile (in the ocean, landfill, etc.) or smelted and separated into base elements. It's an inefficient and labor-intensive process. Plastics can be reused, as well as the stickers and box that an item arrived in.
Lots of derelict products are shipped to SE Asia, where the labor cost of this inefficient process helps compensate by being comparatively low. It also leads to huge piles of ex-computer gear parts that pollute the groundwater in hideous ways. People are poisoned in the scavenging process, not to mention the evil piles of computer dung that are nuclear waste without the isotopes. What's needed is a way to mark directly, every part in a machine. Some parts will be more lucratively recycled. Importantly, those parts that are environmentally damaging, or those that require special devolution processes can be aggregated so that they don't cause interim pollution, and recyclers can benefit from scale of devolution of hazardous materials. Today, we use primitive marks to denote very basic (typically plastic) product composition. We have hazardous materials markers and identification and other markings to identify objects that can be either recycled or are hazardous/dangerous-to-handle.
My suggestion: use advanced barcodes to identify everything by a recycling mark that can be rapidly identified for devolution. The marking doesn't have to be on an easily visible area, but it needs to be revealed somehow. The marks can be tiny, almost microscopic, yet recognized by modern bar code scanners. They could identify either specific categories of product materials, or by actual part number.
In the first case, generic markers can identify tens of millions of generic product identifications, making devolution and separation into elements for recycling vastly simpler than it is today. Specific identification then differentiates subsystems and elements that need specific handling requirements, or perhaps have vendor/manufacturer-specific (even mandated) devolution processes (including rewards).
Another reward potential is that most consumer and industrial products could benefit from the same marking scheme that would permit rapid and accurate product devolution. Junkyards across the world are full of unidentifiable bits and pieces of products gone by, ranging from building cranes to old Volkswagens to refrigerators and no one knows what this stuff is. There are various tests for precious metals (often using primitive magnets) and certain plastics, but many materials aren't easily identified. So they rot, rust, and ooze back into the environment. Materials identification methodologies won't be tough to deploy, and a government mandate seems unnecessary because the motivation to make money from recycled materials exists now.
If we don't do this, then the chances of high-efficiency recycling becomes reduced vastly, and piles of useless and hazardous ex-computer junk become taller. Just as every bill of materials includes parts and sources, we could devolve products when their lifecycle is over systematically. What's needed is an agreement to employ this methodology to the production process: deproduction. The devil of the details will come. Barcodes exist. Now we need a product identification taxonomy, a method to affix material markings, and a database access method that tells the devolvers how to make money.