Is this a bad time for this question?
Why the terms PGA != PLA != PLD != ???
PGA - sometimes it is "Programmable Gate Array" some times "Pin Grid
Array"
(former is a chip architecture the latter is package type)
Does this just represent different
manufacturers trying to control
market lingo?
On the contrary, it is a reaction to that.
The first time I was exposed to programmable logic it was with "PAL"s.
These are fuse programmed and-or-invert gates tied to an array of
interconnects. The term PAL was trademarked by the parent company (whose
name escapes me at the moment MMD?, but AMD bought them) and so the generic
term was "PLA" (programmable logic array). When the networks within the
chip became more complex they weren't really "arrays" anymore and so the
generic term for a logic device that could be programmed was "PLD"
(programmable logic device). This included memories with programmable chip
select logic on chip, and traditional PLAs etc. Erasable devices were
called EPLDs. Then "Gate Arrays" became dense enough to be practical and
the chips were referred to as "ASICs" (Application Specific ICs) or Gate
arrays, or programmed gate arrays (PGAs) but the last was confusing with
the package type. Then gate arrays that could be programmed in the field
(versus mask programmed) made their debut and they were called Field
programmable gate arrays (FPGAs). Devices that were still logic cells that
could be combined but had higher densities are called Complex PLDS or CPLDs.
--Chuck