28 Aug
2002
28 Aug
'02
1:22 p.m.
On Wed, 28 Aug 2002, Ethan Dicks wrote:
With my early experience having come from the
Commodore world, I admit
that the hi-res mapping on the Apple ][ struck me as odd, too; but when
I programming professionally on it, our graphics guru whipped up a fast
and simple row->address lookup routine and we didn't have to worry about
it from an application standpoint.
Yes, it adds a few cycles to the plotting of points, and it is quite lame,
but we must look at the history of why it is like this, as Ethan describes
below.
The "logic" is that the video refresh
circuit doubles as a DRAM refresh
counter. By decoupling the sequential nature of the CPU's view of the
hi-res page and the video output circuit's view of the same memory, the
timing lines up and results in a simpler refresh circuit (i.e., you
don't have to refresh every byte in a DRAM chip, just every row, every
so often; the mapping is driven by the bit geometry of the 4116 DRAM and
how many columns there are per row). I forget how many chips Woz saved,
but I think it was between 1 and 4.
Woz was all about efficiency in design: a true hacker. The hack was very
clever for the day (1977) and saved on chip costs, which were still high
back then. Also, Woz believed that anything that could be done in
software should, and stuck with a minimalist hardware design attitude.
Yes, it's totally lame to those who came up on computers with normal,
sequential graphics screens, but as Ethan mentioned, a simple look-up
table (consisting of $180 bytes) and some convenient 6502 opcodes (load
accumulator indirect indexed) fixes the problem quite nicely and easily.
Sellam Ismail Vintage Computer Festival
------------------------------------------------------------------------------
International Man of Intrigue and Danger http://www.vintage.org
* Old computing resources for business and academia at www.VintageTech.com *