`RGR(1)` and `RGR(2)` gave the window in pixels and nothing gave columns, rows or the cell size -- so anything placing a character *and* a sprite at the same spot had to hardcode a number measured by hand against whatever font the host loaded. The Breakout in examples/ does exactly that, `CW# = 16`, and it is the one thing in that listing that breaks on a different font or window. **`RWINDOW` is BASIC 7.0's own answer and had never been implemented here.** `RWINDOW(0)` is the current text window's rows and `RWINDOW(1)` its columns. `RWINDOW(2)` reports a C128's 40 or 80 column screen mode, and this interpreter has neither -- refused by name, because answering 0 would be a plausible lie, which is worse than a refusal that says why. The cell size in pixels is `RGR(3)` and `RGR(4)`, beside the surface's own dimensions rather than on `RWINDOW`. Two reasons: a cell size is a fact about the surface, and `RWINDOW` reports the *window*, so dividing `RGR(1)` by a column count stops being right the moment a program calls `WINDOW`. Both read a new optional `grid` entry point on `akbasic_TextSink` -- columns, rows, cell width, cell height -- implemented by the akgl sink and forwarded by the tee, in the shape `moveto` and `window` already had. NULL everywhere else, so both verbs refuse by name against a sink with no grid. `akbasic_sink_init_ stdio()` clears it for the same reason it now clears the other two. Measured on the standalone build: `RGR(3)` answers 16 and `RWINDOW` answers 50 columns by 37 rows -- the three numbers the Breakout listing had written out as constants -- and `RWINDOW` follows a `WINDOW` call while `RGR(3)` does not. tests/console_verbs.c drives the answers through a stand-in sink with a grid, since the harness sink is stdio and has none; tests/graphics_verbs.c covers the new `RGR` fields, their refusal, and the moved range bound. The `c excerpt=` block in docs/10-embedding.md moves with the header, which is `docs_examples` doing its job. TODO.md section 6 item 31's second half, struck. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
23 KiB
17. Tutorial: Breakout
This chapter builds a complete game — three lives, six rows of bricks, three level
layouts, a high score and an attract mode that plays itself — out of nothing but the
verbs in the earlier chapters. The finished listing is
examples/breakout/characters/breakout.bas,
and it is worth having open beside this.
The ball and the paddle are sprites; the bricks, the HUD and the messages are characters in the text grid. That split is the first decision and everything else follows from it, so it is Step 1.
Run the finished game with the SDL build:
$ ./build-akgl/basic examples/breakout/characters/breakout.bas
| Key | Does |
|---|---|
| left / right | move the paddle |
| space | start a game, then launch the ball |
| P | pause |
| Q or escape | quit |
Chapter 18 builds the same game again out of sprite artwork, and takes a completely different shape. Read this one first: it is the smaller of the two and it is where the rules that constrain both are explained.
Step 1: Decide what is a sprite and what is a character
Two layers of this interpreter are redrawn from state it keeps for you, every frame, without your program doing anything:
- the text grid, repainted row by row by the akgl sink, and
- the sprites, walked slot by slot after it.
The drawing verbs are not one of them. DRAW, BOX, CIRCLE and PAINT go straight
into the renderer's back buffer, and the text layer paints over the whole window on its
way past — so in the standalone SDL build nothing you draw with them is ever visible.
That is recorded as a defect (TODO.md §9 item 3) rather than a design, but it is the
interpreter you have.
So this game uses no drawing verb at all. Anything that has to move smoothly is a sprite; anything that can live on a 16-pixel grid is text. Ball and paddle move; bricks, score and messages do not.
If you want artwork on the screen, that rule flips and the whole architecture changes with it. That is Chapter 18.
Step 2: Measure the screen once, at the top
CW# = RGR(3)
CH# = RGR(4)
SCW# = RGR(1)
SCH# = RGR(2)
COLS# = RWINDOW(1)
ROWS# = RWINDOW(0)
Six numbers, none of them assumed. RGR(1) and RGR(2) are the window in pixels,
RGR(3) and RGR(4) are one character cell, and RWINDOW gives the text grid in
columns and rows — see Chapter 6. Everything below is derived from
these, including the ball's speed relative to the wall, so the game fits whatever window
the host opened and whatever font it loaded.
The listing in examples/ still has CW# = 16 written out, measured by hand against
the bundled C64_Pro_Mono at 16 points, because when it was written a program had no way
to ask — and it was the one thing in it that would break on a different font. That is
what RGR(3) and RWINDOW are for; TODO.md §6 item 31 is the history.
Lay the rest of the geometry out in the same block — the wall in cells, the play area in pixels:
BRW# = 4
BCOLS# = 10
BROWS# = 6
BTOP# = 4
BLEFT# = (COLS# - (BRW# * BCOLS#)) / 2
TOPY# = 2 * CH#
MAXX# = SCW# - 8
PW# = 144
PY# = 528
A brick is four cells wide, so ten of them are forty cells, and BLEFT# centres that in
whatever COLS# turned out to be. Integer division truncates here, which is exactly what
a cell index wants.
Step 3: Declare the names a subroutine has to answer through
A GOSUB gets its own scope. Assignment walks up the chain and finds an outer name, but a
name first seen inside a subroutine is created in that subroutine and dies at
RETURN — so a routine cannot answer its caller through a name it invented itself.
That is why the game declares this block before anything calls anything:
DIM BR#(60)
DIM LAY$(6)
DIM BSG$(6)
SCORE# = 0
LIVES# = 3
BX# = 0
BY# = 0
HIT# = 0
BI# = 0
HIT# and BI# are how HITTEST (Step 8) answers the routine that called it. Declared
out here they are one variable the whole program shares; left to HITTEST, they would be
two that vanish at RETURN and the caller would read zeros for ever, with no diagnostic
of any kind. Chapter 4 has the scope rules.
Creating a scalar inside a scope is free, so nothing else has to be declared up front:
X# = 0
FOR T# = 1 TO 20000
GOSUB SUBA
NEXT T#
PRINT "OK " + X#
END
LABEL SUBA
LOC# = 1
X# = X# + LOC#
RETURN
OK 20000
Twenty thousand calls, each creating a local. This did not use to work. A scalar drew
storage from the same 4096-slot pool an array does, scope exit gave back the variable but
not its storage, and a program creating one name per tick died in about half a minute —
naming whichever line was unlucky rather than the line that was wrong. It is what killed
the first version of this game after twenty-five seconds. TODO.md §6 item 30 has the
history; tests/value_pool.c is what keeps it fixed.
An array is different, and still costs slots for good. DIM inside a subroutine takes
pool storage that never comes back, because a pointer into a record can outlive the scope
that declared it — see Chapter 16:
X# = 0
FOR T# = 1 TO 6000
GOSUB SUBA
NEXT T#
PRINT "OK " + X#
END
LABEL SUBA
DIM LOC#(4)
LOC#(0) = 1
X# = X# + LOC#(0)
RETURN
? 8 : RUNTIME ERROR Array of 4 elements does not fit in the 0 remaining value slots
So the rule that survives is narrow: DIM at the top, not in a loop. Which is where
you would have put it anyway.
Step 4: Build the wall, and write a row whole
Levels are DATA, one string per row, 1 for a brick and 0 for a hole:
LABEL LAY1
DATA "1111111111"
DATA "1111111111"
DATA "1111111111"
DATA "1111111111"
DATA "1111111111"
DATA "1111111111"
LABEL LAY2
DATA "0011111100"
DATA "0111111110"
DATA "1111111111"
DATA "1111111111"
DATA "0111111110"
DATA "0011111100"
RESTORE takes a label — it wants a line number and a label evaluates to one — so a
layout is just a named place in the DATA stream, and adding a fourth level is a block
and one IF:
LABEL LOADLAY
N# = MOD((LEVEL# - 1), 3)
IF N# = 0 THEN RESTORE LAY1
IF N# = 1 THEN RESTORE LAY2
IF N# = 2 THEN RESTORE LAY3
FOR R# = 0 TO 5
READ LAY$(R#)
NEXT R#
RETURN
Now draw it — and here is the second rule that shapes this listing:
CHAR terminates the row where it stops. A row of the grid is a C string, so writing
at column N erases everything from N+1 on, and writing past the terminator of a short
row draws nothing at all. There is no way to poke one character into the middle of a row
and leave the rest alone.
So a row is rebuilt whole and written from column 0, in one CHAR:
DIM LAY$(6)
DIM BSG$(6)
BSG$(0) = "[##]"
BSG$(1) = "[##]"
BSG$(2) = "[==]"
BSG$(3) = "[==]"
BSG$(4) = "[--]"
BSG$(5) = "[--]"
LAY$(0) = "1111111111"
LAY$(1) = "1111111111"
LAY$(2) = "1101111011"
LAY$(3) = "1111111111"
LAY$(4) = "1011110111"
LAY$(5) = "0110110111"
S$ = ""
T$ = ""
R# = 0
C# = 0
FOR R# = 0 TO 5
GOSUB BUILDROW
PRINT S$
NEXT R#
END
LABEL BUILDROW
S$ = " "
FOR C# = 0 TO 9
T$ = MID(LAY$(R#), C#, 1)
IF T$ = "1" THEN S$ = S$ + BSG$(R#)
IF T$ = "0" THEN S$ = S$ + " "
NEXT C#
RETURN
[##][##][##][##][##][##][##][##][##][##]
[##][##][##][##][##][##][##][##][##][##]
[==][==] [==][==][==][==] [==][==]
[==][==][==][==][==][==][==][==][==][==]
[--] [--][--][--][--] [--][--][--]
[--][--] [--][--] [--][--][--]
That block runs anywhere, with or without a graphics device, because PRINT and CHAR
are building the same string. In the game the last line is CHAR 1, 0, R2#, S$ instead,
and the leading spaces are " " * BLEFT#.
The rows are told apart by shape, not by colour. CHAR parses a colour argument and
ignores it; the sink draws in one colour. [##], [==] and [--] are what makes a
60-point row look different from a 10-point one.
Killing a brick clears its cell and redraws that one row:
LABEL KILLBR
BR#(BI#) = 0
LEFTN# = LEFTN# - 1
PTS# = (BROWS# - BR2#) * 10
SCORE# = SCORE# + PTS#
GOSUB DRAWBR
GOSUB DRAWHUD
IF LEFTN# < 1 THEN STATE# = 2
RETURN
LEFTN# is counted up when the wall is built and down when a brick goes, so "is the
level clear" is a comparison rather than a scan of sixty cells.
Step 5: Make the ball and the paddle
SPRSAV takes an integer array of 63 bytes: three per row, twenty-one rows, most
significant bit leftmost. A clear bit is transparent — see
Chapter 8.
Two patterns are all this game needs. The ball is an 8 by 8 disc in the top-left corner
of the pattern, so the sprite's position and its pixel rectangle are the same thing and
no offset arithmetic is needed anywhere. The paddle is a solid 24 by 6 bar with SPRITE's
x-expand bit set, which makes it 48 wide; three of them laid end to end are one 144-pixel
paddle.
DIM SB#(63)
DIM SP#(63)
I# = 0
D# = 0
FOR I# = 0 TO 62
SB#(I#) = 0
SP#(I#) = 0
NEXT I#
FOR I# = 0 TO 7
READ D#
SB#(I# * 3) = D#
NEXT I#
FOR I# = 0 TO 17
SP#(I#) = 255
NEXT I#
SPRSAV SB#, 1
SPRSAV SP#, 2
SPRSAV SP#, 3
SPRSAV SP#, 4
SPRITE 1, 1, 2
SPRITE 2, 1, 15, 0, 1, 0
SPRITE 3, 1, 15, 0, 1, 0
SPRITE 4, 1, 15, 0, 1, 0
MOVSPR 1, 108, 40
MOVSPR 2, 48, 80
MOVSPR 3, 96, 80
MOVSPR 4, 144, 80
DATA 60, 126, 255, 255, 255, 255, 126, 60
The three segments meet with no seam, which is the whole point of the arrangement. The
game writes the same 63 bytes out as DATA a row at a time, with the bit pattern drawn
in a comment beside it; the loops above are the same two patterns spelled shorter so the
figure fits on a page.
Moving them is four MOVSPRs, and each coordinate is worked out into a variable first:
LABEL SHOWSPR
MOVSPR 1, BX#, BY#
MOVSPR 2, PX#, PY#
X2# = PX# + 48
MOVSPR 3, X2#, PY#
X3# = PX# + 96
MOVSPR 4, X3#, PY#
RETURN
Do not write MOVSPR 3, PX# + 48, PY#. MOVSPR reads a leading sign as move by this
much rather than move to here — see Chapter 8 — and
an expression that starts with one is the relative form. Computing into X2# first is
unambiguous.
Sprites are positioned by their top-left corner in device pixels, so ball, paddle and walls are all plain pixel arithmetic. Only the bricks are in cells, and exactly one routine converts between the two (Step 8).
Step 6: Write the main loop as a GOTO loop
LABEL TICK
GOSUB READKEY
IF STATE# <> 0 THEN GOTO DISPATCH
IF PAUSED# = 1 THEN GOTO TICKEND
GOSUB MOVEPAD
IF HELD# = 1 THEN GOSUB HOLDBAL
IF HELD# = 0 THEN GOSUB MOVEBAL
GOSUB SHOWSPR
LABEL TICKEND
SLEEP 0.02
GOTO TICK
LABEL DISPATCH
IF STATE# = 1 THEN GOTO LOSTLIF
IF STATE# = 2 THEN GOTO LEVELUP
IF STATE# = 3 THEN GOTO BYE
IF STATE# = 5 THEN GOTO NEWGAME
STATE# = 0
GOTO TICK
DO ... LOOP would be the natural way to write this and it is the wrong one. Every
loop and every GOSUB pushes one of 32 scopes, and a state change that jumped out of a
DO — losing a life, clearing a level — would leak its scope every single time. Thirty-two
lives later the program stops. A GOTO loop pushes nothing at all, so the deepest chain
in play is the six GOSUBs of a brick hit.
Nothing here GOSUBs into the state changes either. They are branch targets that run and
then GOTO TICK, so the stack is empty again by the time the next frame starts.
SLEEP 0.02 is the frame pacing, and it does not block: it records a deadline and
the step loop declines to advance the program until the clock reaches it, so the host
goes on drawing frames and pumping the keyboard the whole time. Fifty ticks a second is
what that asks for. Chapter 18 needs a much sharper instrument than this one and builds
it out of TI#.
Step 7: Read the keyboard through a countdown
GET is the only non-blocking read there is. There is no key-up event and no held-key
state — what you get instead is the host's own key repeat arriving as ordinary key-down
events, about one a tick while a key is held down.
So drain the queue every tick, and let a keystroke refill a countdown:
LABEL READKEY
FOR KI# = 1 TO 8
GET K#
IF K# <> 0 THEN GOSUB HANDKEY
NEXT KI#
RETURN
LABEL HANDKEY
IF K# = 1073741904 THEN PDIR# = 0 - 1
IF K# = 1073741904 THEN PDEC# = 12
IF K# = 1073741903 THEN PDIR# = 1
IF K# = 1073741903 THEN PDEC# = 12
IF K# = 32 THEN GOSUB KEYSPC
IF K# = 112 THEN GOSUB KEYPAU
IF K# = 113 THEN STATE# = 3
IF K# = 27 THEN STATE# = 3
RETURN
LABEL MOVEPAD
IF PDEC# < 1 THEN GOTO PADCLMP
PDEC# = PDEC# - 1
PX# = PX# + (PDIR# * PSPD#)
LABEL PADCLMP
IF PX# < 0 THEN PX# = 0
M# = SCW# - PW#
IF PX# > M# THEN PX# = M#
RETURN
The paddle coasts to a stop over twelve ticks when the repeats stop, so held reads as
held and tapped reads as a nudge — with no held-key state anywhere. 1073741903 and
1073741904 are the right and left arrows; the codes are the host's, and Chapter 12's
GET entry says where they come from.
Note 0 - 1 rather than -1. Unary minus on a literal is fine, but the habit of writing
the subtraction out is worth forming here: this parser reads a - b + c as a - (b + c),
so every mixed + and - in the game is parenthesised as a rule rather than where it
happens to matter. It also matches only one unparenthesised AND or OR per
expression. Both are in Chapter 13.
Step 8: Move the ball one axis at a time
LABEL MOVEBAL
OX# = BX#
BX# = BX# + BVX#
IF BX# < 0 THEN GOSUB WALLL
IF BX# > MAXX# THEN GOSUB WALLR
GOSUB XBRICK
OY# = BY#
BY# = BY# + BVY#
IF BY# < TOPY# THEN GOSUB WALLT
GOSUB YBRICK
GOSUB PADHIT
IF BY# > LOSEY# THEN STATE# = 1
RETURN
X first and then Y, each move tested on its own, so a ball arriving at the corner of
a brick reverses one axis rather than both. OX# and OY# remember where it came from,
which is what a collision backs it out to.
The brick test converts a pixel to a cell, and it is the only place in the program that does:
LABEL HITTEST
HIT# = 0
RC# = TY# / CH#
IF RC# < BTOP# THEN RETURN
RB# = RC# - BTOP#
IF RB# > (BROWS# - 1) THEN RETURN
CC2# = TX# / CW#
IF CC2# < BLEFT# THEN RETURN
CB# = (CC2# - BLEFT#) / BRW#
IF CB# > (BCOLS# - 1) THEN RETURN
BI# = (RB# * BCOLS#) + CB#
BR2# = RB#
IF BR#(BI#) = 1 THEN HIT# = 1
RETURN
Callers hand it a point in TX#/TY# and read HIT#, BI# and BR2# back out — which
works only because Step 3 created all three outside the routine.
Test the ball's leading edge, not its centre:
LABEL XBRICK
TX# = BX#
IF BVX# > 0 THEN TX# = BX# + 7
TY# = BY# + 4
GOSUB HITTEST
IF HIT# = 0 THEN RETURN
BX# = OX#
BVX# = 0 - BVX#
GOSUB KILLBR
RETURN
A ball tested by its centre sinks four pixels into a brick before it turns, and it looks like the brick was hit late. The cost of the leading edge is that a brick clipped at the very corner can be missed by up to seven pixels' worth of ball; that is the better trade of the two, and it is worth knowing which one you took.
Step 9: Make the paddle bounce mean something
LABEL PADHIT
IF BVY# < 1 THEN RETURN
BB# = BY# + 8
IF BB# < PY# THEN RETURN
IF BY# > (PY# + 10) THEN RETURN
RX# = PX# + PW#
IF (BX# + 8) < PX# THEN RETURN
IF BX# > RX# THEN RETURN
BY# = PY# - 8
BVY# = 0 - BSPD#
Z# = ((BX# + 4) - PX#) / (PW# / 5)
IF Z# < 0 THEN Z# = 0
IF Z# > 4 THEN Z# = 4
PVX# = BVX#
BVX# = (Z# - 2) * 3
IF BVX# <> 0 THEN GOTO PADAIM
BVX# = 2
IF PVX# < 0 THEN BVX# = 0 - 2
LABEL PADAIM
GOSUB BEEPP
RETURN
Five zones across the face and the zone decides the angle: Z# is 0 to 4, so the new
horizontal speed is -6, -3, 0, +3, +6. Where you catch it decides where it goes, which is
the entire game.
The dead-centre case needs its own handling. A ball with no sideways speed rises and falls down one column for ever, which is how attract mode used to hang itself — so the middle zone keeps the direction the ball came in on and gives it a shallow angle instead of zero.
There is a watchdog for the same failure in a subtler form. Ten seconds without touching a brick means the ball has found an orbit that misses the wall, so it gets a new angle — at the next paddle bounce, never in mid-air. Changing it in flight looks exactly like the ball hitting something that is not there:
LABEL UNSTICK
NUDGE# = 0
STALL# = 0
RMAX# = 4
GOSUB RANDOM
BVX# = (RND# * 3) - 6
IF BVX# = 0 THEN BVX# = 3
RETURN
RANDOM is a linear congruential generator, because this dialect has no RND — and
no INT, SQR, ASC or TIMER either. TI# is jiffies off the host's clock and is
uptime rather than zero-based, which makes it a fine seed:
LABEL RANDOM
SEED# = MOD(((SEED# * 1103515245) + 12345), 2147483648)
RND# = MOD((SEED# / 65536), RMAX#)
RETURN
The multiply stays inside a 64-bit integer for any seed under 2^31, and the answer comes
from the middle bits because the low bits of a power-of-two modulus barely move.
Integer division truncates for free, which is the INT you do not have.
Step 10: The HUD and the messages
Same rule as the wall: build the whole line, write it once.
V# = 0
P$ = ""
H$ = ""
SCORE# = 1250
LIVES# = 3
LEVEL# = 2
HIGH# = 9900
V# = SCORE#
GOSUB PAD6
H$ = " SCORE " + P$
H$ = (H$ + " LIVES ") + LIVES#
H$ = (H$ + " LEVEL ") + LEVEL#
V# = HIGH#
GOSUB PAD6
H$ = (H$ + " HIGH ") + P$
PRINT H$
END
LABEL PAD6
P$ = "" + V#
DO WHILE LEN(P$) < 6
P$ = "0" + P$
LOOP
RETURN
SCORE 001250 LIVES 3 LEVEL 2 HIGH 009900
"" + V# is how a number becomes a string: + concatenates a string with a number, and
that is the conversion — see Chapter 5. Every + after
the first is parenthesised, for the reason in Step 7.
A centred message is the same idea, and needs no padding on the right because CHAR
terminating the row is what erases the rest of it:
LABEL SHOWAT
L# = LEN(MSG$)
C2# = (COLS# - L#) / 2
S$ = MSG$
IF C2# > 0 THEN S$ = (" " * C2#) + MSG$
CHAR 1, 0, MROW#, S$
RETURN
LABEL CLRAT
CHAR 1, 0, MROW#, " "
RETURN
Clearing a message is a single space written at column 0 — one character, and the terminator behind it takes the rest of the row with it.
Step 11: Ask for sound once, then believe the answer
SOUND refuses on a machine with no audio device, and an untrapped refusal ends the
game. Ask once at startup, record the answer, and never ask again:
SND# = 1
TRAP NOAUDIO
SOUND 1, 2000, 1
TRAP
IF SND# = 1 THEN PRINT "SOUND IS AVAILABLE"
IF SND# = 0 THEN PRINT "NO AUDIO DEVICE, PLAYING SILENTLY"
END
LABEL NOAUDIO
SND# = 0
RESUME NEXT
NO AUDIO DEVICE, PLAYING SILENTLY
That output is the stdio build, which has no devices at all. On the SDL build the same
program prints the other line. TRAP with no argument disarms the handler again — leave
it armed and the next refusal anywhere in the game is silently swallowed. See
Chapter 4.
After that every sound is guarded by the flag it recorded:
LABEL BEEPB
IF SND# = 1 THEN SOUND 1, 12000, 2
RETURN
SOUND's frequency argument is a SID register value, not hertz — the pitch is
register * 1022730 / 16777216, and Chapter 7 has the arithmetic.
Work the notes out once and write them down in a comment beside the numbers.
Step 12: Give it an attract mode, and test with it
The title screen plays itself, and this is not decoration — it is how the game gets
tested without a hand on the keyboard. DEMO# is the only difference between the demo
and a real game:
LABEL DEMOPAD
TX# = (BX# + 4) - (PW# / 2)
TX# = TX# + DOFF#
D# = TX# - PX#
IF D# > PSPD# THEN D# = PSPD#
M# = 0 - PSPD#
IF D# < M# THEN D# = M#
PX# = PX# + D#
RETURN
LABEL DEMOAIM
RMAX# = 81
GOSUB RANDOM
DOFF# = RND# - 40
RETURN
The demo paddle tracks the ball but aims off-centre by a random amount, refreshed at every bounce, so it plays like something with a hand on it and occasionally misses. A demo that tracks perfectly never loses a ball and therefore never tests losing one.
Leave it running for two minutes and watch the score climb: that exercises the ball, the bricks, the level change and the speed-up — and it is the only thing that exercises them together, for long enough. Every defect this game found took thousands of ticks to show itself, and a green test suite proved nothing about any of them.
The machine does not get on the scoreboard. Attract mode keeps score so the wall behaves, but only a player's score is ever the high one:
LABEL HISCORE
IF SCORE# > HIGH# THEN HIGH# = SCORE#
RETURN
There is nowhere to save it. There is no disk in this interpreter — see Chapter 9 — so the high score lasts as long as the process does.
The rules this listing never breaks
Every one of these was learned by breaking it. They are the checklist to write your own game against:
| Rule | Because | Where |
|---|---|---|
DIM at the top, never in a loop |
An array created inside a scope costs pool slots for good; 4096 ends the run. A scalar is free | Step 3 |
| Build a text row whole and write it from column 0 | CHAR terminates the row where it stops |
Step 4 |
| Tell rows apart by shape, not colour | CHAR ignores its colour argument |
Step 4 |
Compute a MOVSPR coordinate into a variable first |
A leading sign means move by, not move to | Step 5 |
Loop the game with GOTO, not DO |
32 scopes, and jumping out of a loop leaks one | Step 6 |
Parenthesise every mixed + and - |
a - b + c parses as a - (b + c) |
Step 7 |
One AND or OR per expression unless parenthesised |
The parser matches one | Step 7 |
| Keep a line under 32 tokens | The 33rd kills the interpreter with a stack trace, not a BASIC error you can TRAP |
— |
Span a loop across lines, and never FOR I = 1 TO 1 |
A whole loop on one line does not loop; equal bounds run zero times | Ch 13 |
| Probe for a device once and remember | An untrapped refusal ends the program | Step 11 |
Where to go next
- Chapter 18 builds Breakout again with sprite artwork, powerups and a coloured HUD. Every architectural decision comes out different, and the chapter is about why.
- Chapter 13 is the full list of what this dialect does differently from BASIC 7.0.
- Chapter 14 is the interpreter itself — the step loop, the pools and how to debug a program that stops for no visible reason.
TODO.md§6 and §9 carry the defects this game found, each with a reduction that fits on a screen and what a fix would touch.
