There are various methods used depending upon the part type used and/or the types of program memory used.

1) Some parts have EPROM code memory. These are typically programmed via a programmer device which uses a device specific algorithm which is often a "parallel load" process.

2) Some hardware configures a EPROM based part with a loader in the EPROM and then there is a SRAM memory organized outside the part that can have program code downloaded to it via the loader. This loader often works over the UART port.

3) Some older hardware configurations used processors that had no onboard program memory and so EPROM components are mounted in sockets on the external program bus. These program EPROM parts would be programmed via a EPROM programmer device and the parts then inserted into the sockets on the target hardware.

4) With the advent of standalone FLASH chips there are hardware configurations similar to #3 where FLASH chips are programmed separately and then inserted into the sockets of the target hardware. As a point of interest the earliest FLASH chips were pin compatible with EPROM chips and so could be used in the same sockets.

5) Further progress has led to the fact that FLASH could be re-programmed in circuit so a FLASH programmer was put into the 8051 OR into a first protected block of the FLASH and then the circuit design permitted download of code via the UART port and code could be written to the external FLASH chips which could now be soldered to the boards.

6) Many varieties of the 8051/8052 chips can be manufactured with the program confgured right into the metal mask of the semiconductor chip. For this configuration no "download" is required because the chips are "ready made" and soldered right into the board.

7) FLASH memory becomes available right on board the 8051 type parts and many designs arranged the microcontrollers into 40 pin DIP sockets or 44 pin PLCC sockets. In these the parts are programmed in an external programmer and then plugged into the sockets of the target hardware.

8) FLASH has then evolved to the possibility that a special port interface of a serial nature can access and program & re-program the FLASH via a special procedure while the chip is soldered into the target hardware. Today this is a rather quite common implementation. This is often reffered to as ISP (In System Programming) and is commonly implemented via a JTAG port.

9) Some manufacturers put special code on-board their parts in a read-only-memory that is only accessable if the part is booted out of reset via a special procedure. These implementations can and often do support FLASH programming in this mode over the UART port with the special ROM code doing the actual programming function. This is another form of ISP.

10) In the latest incantations of parts the manufacturers have made the circuitry necessary to program the on-board FLASH available to the normal user program simply by allowing the program to enable it by setting some special register bits. These then support a re-programmability that is called IAP (In Application Programming). The method of feeding new program code to the part can be as creative as the designer desires and you will find designs where the download occurs over I2C, SPI, UART, CAN, Dual Port Memory, just to name a few.

So..........as you can see there are MANY ways.

Today, if you are clever, you will choose a part that has on-board FLASH that is of size sufficient to meet your needs, 2K all the way to 128K byte parts are available, that can at least be programmed via ISP methods. And if you desire subsequent field upgradeability you will pick a part that has IAP capability. Many of the newer parts also have on-board RAM up to 8K or even 16K bytes of RAM too so this means that many, if not even most, applications can be designed to be true "single chip" solutions !!

Good Luck
Michael Karas