Evolution of electronic circuit construction technology
I think it worth pointing out that my personal experience of how electronic circuits are constructed shows a steady movement away from putting chips in sockets, and towards soldering them directly to the PCB. As far as I know, much of the electronic service industry works on the basis of swapping out complete boards, rather than repairing at component level.
In the 1980s, my designs were all through-hole, with no surface mount. CPU chips and ROM would be in sockets. Other ICs might also be in sockets. In those days, for embedded computing, you put the machine code in EEPROMs using an external programmer, and then plugged them in. We used Z80 systems quite a lot.
Later, there were CPU chips with internal programme ROM. The ROM tended to come in two varieties: EEPROM that could be reprogrammed after UV erasing, and one time programmed (OTP). The EEPROM devices were quite expensive, and only used in development. The OTP devices were cheaper, and used in production. Obviously, if you wanted to upgrade firmware in an OTP device, you would have to replace the chip, which tended to mean that it was inserted in a socket.
The movement towards surface mount technology meant most chips were not in sockets, but soldered straight to the board. The problem of firmware upgrades was fixed by having chips that are in circuit programmable. This means you can solder the CPU to the PCB, then erase and programme it multiple times via a header on the PCB. The PIC micros I use these days use a five pin header, and a fairly simple programmer, driven from USB. Some chips can have firmware directly loaded via USB. This is about where my designs are today. All IC sockets are gone.
Though soldering chips to the PCB makes component level repairs more difficult than when you have chips in sockets, the kit to desolder chips can still be affordable. It is basically a hot air blower. A certain amount of that goes on in our production line. However, you have to consider whether it is worth taking the time to repair a board that might only be worth a few quid. Boards that fail production test are usually put on one side, then a skilled technician will do some kind of triage to select which boards look worth repairing, and glean a bit more out of the scrap. The production line is generally run on a right-first-time basis, to avoid adding to the scrap pile in the first place.
Where things start to get difficult for component level repairs is with modern surface mount technology such as ball grid arrays (BGAs). As far as I know, a hand-held hot air blower will not really cope with desoldering such devices. You need special kit, which is expensive. I worked on a board with a BGA processor, and densely packed tiny components. It was not my design. The wrong type of RAM had been fitted, but replacing that ruined the BGA joints nearby, and the board was bricked. This was done via a contract manufacturer I have some respect for. They have the specialised kit, and even they could not do the job. Ever since then, I have avoided anything like a BGA in new designs. My colleague who does production engineering and quality control is in agreement on this. We could deal with BGAs if we had the X-ray inspection kit and so on, but as we don't have that now, no BGAs please.
It could be that for some electronic kit, what people want is the right to swap out circuit boards, and not try to do component level repair. Manufacturers will tend to do servicing by swapping out whole boards, so consumers want to be able to buy spares and do that themselves. In that case, all this discussion of desoldering chips is not so relevant. However, this level of repair does mean that you are almost certainly tied to the original manufacturer when it comes to getting spares, because these boards are not generic components that you can buy anywhere. At present, I don't think manufacturers are legally obliged to sell spares to consumers, and maybe that is what Right to Repair is about.