... roughly said, most of the semiconductor manufacturers use the subsequent scheme (example is your 1.):

1. prefix, denoting manufacturer¹ - P is for Philips
2. first two-three digits or characters, denoting function group - 89 is for Flash microcontroller
3. in digital circuits, a middle letter denoting technology and/or voltage - C is for CMOS (here: as opposed to NMOS, which was an older, more power consuming technology, albeit in those times somewhat faster)
4. several digits and/or characters denoting model - 51RD2
5. suffix of several characters/digits, denoting temperature range² and/or package - BN
5a. sometimes a suffix denoting a variation or improvement of the original circuit - (nothing here)
6. after a dash, maximum operational frequency - nothing here
7. the full order number also includes additional suffixes for circuits packaging - e.g. tape and reel, boxed, etc.

Digital circuits tend to follow this scheme somewhat closer than analog circuits (where much more diversity in all respects is seen than in digital); however, there is no fixed scheme and manufacturers are free to establish their own numbering. Notable exceptions:
¹ - Intel in their microcontrollers (although it does not make them anymore) marks package/temperature etc. by the prefix. This is a bit suprising, as Intel was very influential in the early world of digital microelectronics, and much or the marking of digital circuits can trace its roots to the Intel marking [1]
² - the famous TI SN74xx basic logic ICs had in fact also 54xx and 84xx variants, where the first digit denoted the temperature range

Quite many of the markings have also evolutionary background. Your first example is a very nice example of this (we are on an 8051-specific site so I might perhaps risk a somewhat longer historical excursion...?)

P - as said above, it stands for Philips

8 - Intel's marking for processors and their companion circuits. This is "out of the original scheme" numbering, adopted after the success of 8008/8080 processors, which owe their name to a "play" on the 4bit-4004 -> 8bit-8008 line

9 - a continuation of Intel's marking of memory technology (0 was rom-less (sometimes also mask-rom), 1 was RAM (as in the 2114 RAM), 3 was definitively mask-ROM, 7 was EPROM, 8 was EEPROM and then 9 is FLASH

C - denoting CMOS (as opposed to NMOS or bipolar)

51 - the 8051 microcontroller was originally seen as a modernised member of the 8048 family (which already had 8049 and 8050 members, both doubling the amount of code memory of its predecessor)

R - Intel's marking of the 8051 variants with added internal-XRAM (called "ERAM" by Intel)

D - continuation of Intel's marking of amount of code memory in the '51Rx line - RA had 8kB, RB - 16kB, RC - 32kB, RD - 64kB

2 - Philips' way of denoting the 6-clk (double clock) option

B - temperature range (marked sometimes as "Commercial", roughly 0-70°C)

N - package - DIP


JW

----------
[1] http://download.intel.com/technolog.../art_1.pdf