Hardware Maintenance in the Digital World
Hardware maintenance is a physical process. When a connection is broken, severed, or jostled loose – someone must reconnect the wire. When a circuit board is damaged by heat or moisture, someone must replace the board to return the unit to service. None of this has changed since the ENIAC (1st computer) began the computer era. The concept of logic gates created with plug boards are still in use – only the production method has changed to switch from very large wires to printing tiny wires into a silicon wafer.
Today’s repairs, aka “Break-Fix” and “Hardware Maintenance” are different only in matters of scale and sophistication. More diagnostic routines are provided with the machine to facilitate self-testing of connections. More parts are designed to be pulled and replaced easily so as to reduce the labor costs demanded by highly trained technicians. Problems with chip designs are still resolved by the hardware designers and reflected in new iterations of the chip.
What has changed, mostly over the past decade, is how Original Equipment Manufacturers (“OEMs”) have changed their marketing policies to drive customers exclusively to their high-margin repair services. The equipment has not changed – it is faster and more powerful but not fundamentally different. A broken wire in 2013 is still a broken wire.
The common theme among all OEMs seeking to control their aftermarket for machines, parts, and break-fix options is restrictions on access to the bits of code that make the machine operate by claiming this code as “Intellectual Property” (IP). Buyers are fearful of being out of compliance on any software issues and acquiesce readily to this illogical marketing construct.
All the routines necessary to operate a machine can be delivered as hardware “baked” in a chip. As machines have become vastly more complex, the opportunity for errors is similarly increased. Rather than have to replace a chip with a flaw (such as Intel has occasionally needed to do), modern machines are manufactured with modest programmable update capability to make corrections to flaws faster and less costly to deliver. Flaws can still be corrected with a physical “Engineering Change” (“EC”) but the physical replacement is the most costly option.
Therefore, most chip (hardware) code errors are fixed using media delivered patches and fixes. Today’s media is the internet. Prior ECs were delivered using tape media, disks, etc. depending on the most common external media ports available on the machine. EC’s were always free because they were (and are) making corrections to the specifications of the machine. Some OEMs may have charged a minor media and shipping fee.
ECs are clearly hardware when they are included in a chip. If the lifecycle of a product is long enough, all chip (machine) code errors are incorporated into the next iteration of the physical chip design. The older version with all patches and fixes applied is identical to the new one. It is illogical to agree that the original chip is hardware and the fixed chip is hardware, but the intermediate repairs (fixes) are IP and do not belong to the buyer as part of their ownership.
In the used/refurbished equipment market – the terminology “At Current EC Levels” is a common requirement in a transaction. If the machine is not up to EC levels – all the applicable patches and fixes have to be applied by the equipment owner before the buyer will complete the transaction. Making sure a machine is up to all current EC levels (aka application of all known patches and fixes) is similarly necessary for the repair of the equipment. Many parts may be in storage that are missing some patch or fix which must be applied during the repair so the user has a fully functional machine. This is why blocking access to ECs is so harmful to both the secondary market and the self/independent repair market.
In the automobile world we call such patches and fixes “Recalls” and not ECs. A recall is not a desirable thing for owners or manufacturers. At best, it is annoying to have to bring in a vehicle for a recall – whether to a flawed chip design (in an engine control module), or a faulty window washer part. At worst, one might be the first user of the infamous Firestone tires involved in a fatal accident.
Manufacturers of IT products have avoided using the terminology of Recalls but the obligation of the manufacturer to deliver products that work to their specifications remains the same.
Other types of “IP” are also being lumped together with essential machine code patches and fixes that may or may not have any impact on the repair of the machine. This is a marketing choice on the part of the OEM and not a technical one. All OEMs have the option to separate “Recalls” from upgrades and enhancements or other functions.
To be sure, there are other problems limiting equipment owners in their choice of independent repair and resale, but at the most basic level if the relationship between machine code and IP is not resolved such that owners have unlimited access to all patches and fixes needed to facilitate a complete repair, it will not matter what other competitive problems remain.
Today's OEMs are inconsistent. While Apple, Dell, Microsoft and HP still freely make updates available, others like Oracle/Sun Microsystems and IBM do not. The Digital Right to Repair Coalition is introducing the Digital Right to Repair Act 2013 to ensure consumers have rights to repair what they own.