# HG changeset patch
# User Paul Boddie <paul@boddie.org.uk>
# Date 1684329975 -7200
# Node ID 9a19c0cb0c45936397ce58ca4152a8371eb815e6
# Parent  ab28336894bcdf1f8f1c9173dd8f61d72a7213d5
Moved the new section to be adjacent to a related section, adding some notes.

diff -r ab28336894bc -r 9a19c0cb0c45 ULA.txt
--- a/ULA.txt	Tue May 16 15:54:20 2023 +0200
+++ b/ULA.txt	Wed May 17 15:26:15 2023 +0200
@@ -1483,6 +1483,39 @@
 character values (repeatedly) and those to read the bitmap data to reproduce
 the characters on screen.
 
+Enhancement: 40-Column Text Modes by Interleaving Screen and Bitmap Accesses
+----------------------------------------------------------------------------
+
+A simplified form of the above interleaved character/bitmap reading scheme.
+This was also suggested in a discussion here:
+
+https://stardot.org.uk/forums/viewtopic.php?p=393243#p393243
+
+The ULA could be run in high-bandwidth mode to fetch character codes from
+screen memory in one cycle and then to use the character code to look up a
+pixel row of a character bitmap, reading that bitmap slice in the following
+cycle. The bitmap would be converted to pixel values that would then be
+emitted over the subsequent two cycles concurrently with the preparation of
+the next character's pixels.
+
+  2MHz cycle: 0 1 2 3 4 5 ...
+  Reads:      C B C B C B ...
+  Pixels:         a   b   ...
+
+The memory access to bitmap data would be computed as follows, assuming the
+normal eight pixel height and single-byte encoding of character bitmaps:
+
+  bitmap address = bitmap table base + (character code * 8) + index
+
+Each successive pixel row on the screen would expose the appropriate row in
+the character bitmap, with this index looping from 0 to 7 repeatedly as shown
+previously. Spacing between character lines could be introduced as already
+done in MODE 6.
+
+Character bitmap data would be stored in RAM, since this is the only possible
+source of data for the ULA as delivered. The use of ROM would require changes
+to the broader system architecture.
+
 Enhancement: MODE 7 Emulation using Character Attributes
 --------------------------------------------------------
 
@@ -1554,31 +1587,6 @@
 MODE 7 support could be made available, potentially as a hardware option
 separate from the ULA.
 
-Enhancement: 40-Column Text Modes by Interleaving Screen and Bitmap Accesses
-----------------------------------------------------------------------------
-
-Suggested here: https://stardot.org.uk/forums/viewtopic.php?p=393243#p393243
-
-The ULA could be run in high-bandwidth mode to fetch character codes from
-screen memory in one cycle and then to use the character code to look up a
-pixel row of a character bitmap, reading that bitmap slice in the following
-cycle. The bitmap would be converted to pixel values that would then be
-emitted over the subsequent two cycles concurrently with the preparation of
-the next character's pixels.
-
-  2MHz cycle: 0 1 2 3 4 5 ...
-  Reads:      C B C B C B ...
-  Pixels:         a   b   ...
-
-The memory access to bitmap data would be computed as follows, assuming the
-normal eight pixel height and single-byte encoding of character bitmaps:
-
-  bitmap address = bitmap table base + (character code * 8) + bitmap row
-
-Each successive pixel row on the screen would expose the appropriate row in
-the character bitmap, with this "bitmap row" looping from 0 to 7 repeatedly.
-Spacing between character lines could be introduced as already done in MODE 6.
-
 Enhancement: Compressed Character Data
 --------------------------------------