Bob's Blog

The original author's email (2 years old) is on the page you linked to. I'll bet he'd like to hear about this.

How do you play the game?

How do you play the game?

The 2600 version (mine) isn't playable as a game yet - no collisions and no scoring - download the A8 ROM to check out the gameplay.

 

Briefly:

You want to reach the top of the building; you move up one floor by completely crossing the building.

Hit an elevator and you lose a life and start over at the beginning of the floor.

When you reach the top you start over at the bottom with the elevators moving at faster speeds.

Scoring: a timer counts down - when you complete a floor whatever is left on the timer is added to your score.

 

Elevator Repairman was a type-in game (machine language, though!) from ANALOG magazine and, as the author mentions, was based on Spy's Demise, which was a commercially released A8 game that is basically the same though prettier looking.

So it's kinda Squish'em-ish, just without'ems to squish?

So it's kinda Squish'em-ish, just without'ems to squish?

Kinda...doesn't scroll, though. And the movement is all left-right rather than up-down.

Spy's Demise was a game that looked cool, but never really seemed to be all that much fun. Seems like a somewhat interesting kernel challenge. I can see a few nice approaches:

 lax (elevator1),y
 lda shiftleft,x
 ora (elevator2),y
 sta PF1
 lax (elevator4),y
 lda shiftleft,x
 ora (elevator3),y
 sta PF2
 lax (elevator5),y
 lda shiftleft,x
 ora (elevator6),y
 sta PF2
 lax (elevator8),y
 lda shiftleft,x
 ora (elevator7),y
 sta PF1

This requires having a shape table in ROM with ample blank space above and below the elevators. Time required is 17 cycles per playfield word, or 68 cycles for all four. Should be adaptable for a 2LK (the PF stores couldn't be timed as illustrated). Using two shape tables would save 16 cycles.

That's an interesting idea, John; thanks for sharing. The method I'm using is just as fast - a little faster for PF2 - but uses two shape tables:

   lda (ElevPtr1),Y
  lsr
  ora (ElevPtr2),Y
  and #%00110011
  sta PF1		;+17

...

  lda (ElevPtr3),Y
  ora (ElevPtr4),Y
  and #%10011000
  sta PF2
  and #%10000000
  ora (ElevPtr5),Y
  and #%10011000
  sta PF2		;+27

...

  lda (ElevPtr7),Y
  lsr
  ora (ElevPtr6),Y
  and #%00110011
  sta PF1		;+17

ElevPtrs 1, 4, 5, and 7 point at a shape table holding this value: %01111000 (with a bunch of zeros before and after)

And ElevPtrs 2, 3, and 6 point to this shape: %10000011

 

As you can see from the binary/screenshot I posted, I draw PF1 and PF2 on alternating scanlines.

As you can see from the binary/screenshot I posted, I draw PF1 and PF2 on alternating scanlines.

 

Slightly easier sometimes is to draw half of the objects using each playfield register on each scan line. This is the approach used for the score digits in Space Invaders. Not sure whether that approach would be particularly suitable for this game, but something to consider.