Yes, that's certainly an option, but was never very popular for cost reasons.
When memory got a little cheaper, it became less of an issue whether you had 11
bytes per character instead of (1 + a character-genrator). A graphic image of a
character would need about that much, at least 10 bytes, I think, if you want
full decenders. Back in the Apple and CP/M days, inexpensive microprocessors
were too slow to do that sort of thing at a reasonable rate. Later on it was
not uncommon to have different font sizes/styles/colors in the same graphic, but
that was a different era. It was never a secret that one could do this, but the
cost factor was what made character generators popular for presenting text.
Character generators (ROMs) weren't cheap either, BTW.
Dick
----- Original Message -----
From: "Pete Turnbull" <pete(a)dunnington.u-net.com>
To: <classiccmp(a)classiccmp.org>
Sent: Wednesday, July 18, 2001 12:54 AM
Subject: Re: Light Pens ...
On Jul 17, 20:22, Richard Erlacher wrote:
If the refresh memory is to support text and
graphics, the pipeline must
be
two-forked.
Not necessarily. In a BBC Micro, for example, the 6845 is essentially
generating a stream of pixel addresses, since what's stored in the screen
memory is the bitmap of the character, not the character code. That's how
you mix text and graphics on the same screen. When text is written to the
screen, the OS looks up the bitmap(s) of the character(s) and writes the
individual pixels to screen RAM.
--
Pete Peter Turnbull
Network Manager
University of York