Sorry your browser is not supported!

Begin: Choosing your type of landscape
The first step is to decide how you will design the landscape.  The reason being this is because there are different ways to implement the mouse interface that you'll use to select spots in 3D on the landscape.  You could use a matrix, but to have accurate 3D mouse positioning you will be limited to a completely flat ground.  If that works for you, then use it.  It'll keep things more simple later on.  However, if you want a rough terrain with some hills and valleys, you'll need to use an object.  You could use the new terrain commands recently added to DBP or use the memblock matrix code that myself and a few others(Van-B,Kevil) have written.  By using an object, 3D mouse coordinates are easily attained by using the pick object and get pick vector commands.  For those that wish to stick with the standard flat matrix, here is the code used to translate a 2D mouse position into a 3D coordinate.  The author of this code snippet is unknown to me.


These variables must be declared at the beginning of your program.
screenwidth2# = screen width() / 2.0
screenheight2# = screen height() / 2.0
scalefactor# = screen height() / 1.2

The offsetx#, offsety#, offsetz# variables is position at which the camera is pointing to.


_Click_Matrix_Position:
   mouseposx#=mousex() - screenwidth2#
   mouseposy#=screenheight2# - mousey()

   dist#=sqrt((camera position x()-offsetx#)^2+(camera position y()-offsety#)^2+(camera position z()-offsetz#)^2)

   if mouseposy#<>0
      vectorang#=atanfull(scalefactor#,mouseposy#)
   else
      vectorang#=90.0
   endif

   if vectorang#+camera angle x()>90.965
      ratio#=mouseposy#/(sin((vectorang#+camera angle x() )-90.0))
      cursorposy#=ratio#*sin(180.0-vectorang#)
   else
      cursorposy#=1000000.0
   endif

   hyplength#=scalefactor#/sin(vectorang#)
   cursorposx#=(((ratio#*sin(90.0-camera angle x() ))/hyplength#)+1.0)*mouseposx#

   if mouseposx#<>0
      angtotal#=wrapvalue(camera angle y()-atanfull(mouseposy#,mouseposx#))
   else
      if mouseposy#<0
         angtotal#=wrapvalue(camera angle y()+90.0)
      else
         angtotal#=wrapvalue(camera angle y()-90.0)
      endif
   endif

   cameraang#=wrapvalue(360.0-camera angle y() )

   cosang#=cos(cameraang#)
   sinang#=sin(cameraang#)

   movex# = (dist#/scalefactor#)*((cosang#*cursorposx#)+((-1*sinang#)*cursorposy#))
   movez# = (dist#/scalefactor#)*((sinang#*cursorposx#)+(cosang#*cursorposy#))

   rem 3D position stored into these variables
   movex#=movex#+offsetx#
   movez#=movez#+offsetz#
   movey#=0

RETURN

So to use the code properly, you must know where the camera is pointing. Easily done, just pick a spot on your matrix to point the camera at so you have the desired angle of view that you want.  Then whenever you call this routine, the 3D position of your mouse cursor will be stored into the movex# and movez# variables.

Now for those who wish to use an object for their terrain.

rem object number of your terrain
object = 1
null = pick object(mousex(), mousey(), object, object)
targetX# = camera position x() + get pick vector x()
targetY# = camera position y() + get pick vector y()
targetZ# = camera position z() + get pick vector z()

That's it.  Much easier isn't it.  So why would anyone want to use a flat matrix with less flexibility when it takes more code to use?  The reason I said a flat matrix is easier won't be fully explained until later, but just think about position buildings on your terrain.  If you know it's all flat, then how you place the object is quite simple.  But think about placing a house or building on a random matrix.  A bit more complex task, if you want it to look right.  From now on in code, I will only use this function name, get3DMouse(), and these variables, targetX#, targetY#, targetZ#, when referring to either of the above 3D mouse selection routines.


Step 2: Move your character
Great, we can now get a 3D position from our mouse, but how do we make our character walk to the targeted spot?  This is easier than some people might think.  Here, see for yourself.

rem object number of character
character = 15
rem speed at which character will move
speed# = 3.0
mouseFlag = 0
moveFlag = 0

DO
   _mouse_click = mouseclick()

   rem if user right-clicks
   if _mouse_click = 2 and mouseFlag = 0
      mouseFlag = 1
      get3DMouse()
      moveFlag = 1
      rem get the angle between the character and targeted spot
      angle# = atanfull(targetX# - object position x(character), targetZ# - object position z(character))
      rem rotate character to point towards target
      yrotate object character, angle#
   endif

   if _mouse_click <> 2 then mouseFlag = 0

   if moveFlag = 1
      rem get character's new position
      x# = newxvalue(object position x(character), angle#, speed#)
      y# = getTerrainHeight()
      z# = newzvalue(object position z(character), angle#, speed#)
      position object character, x#, y#, z#
      rem if character is within range of target, then stop moving
      if getSquaredDistance#(x#, z#, targetX#, targetZ#) < (speed# * speed#)
         moveFlag = 0
      endif
   endif

   sync
LOOP

The getSquaredDistance() function returns the distance between two points.  But instead of using sqrt() to get the actual distance, it returns the distance value still squared.  We can get away with doing this because we're only using the distance for comparison.  Just remember to square the other comparison value.  If the character's speed is 3, then it'll stop moving if it's within 3 units of the target.  As you can see, we only need to call get3DMouse() when the user clicks the mouse.

So to use the code properly, you must know where the camera is pointing. Easily done, just pick a spot on your matrix to point the camera at so you have the desired angle of view that you want.  Then whenever you call this routine, the 3D position of your mouse cursor will be stored into the movex# and movez# variables.

Now for those who wish to use an object for their terrain.

rem object number of your terrain
object = 1
null = pick object(mousex(), mousey(), object, object)
targetX# = camera position x() + get pick vector x()
targetY# = camera position y() + get pick vector y()
targetZ# = camera position z() + get pick vector z()

That's it.  Much easier isn't it.  So why would anyone want to use a flat matrix with less flexibility when it takes more code to use?  The reason I said a flat matrix is easier won't be fully explained until later, but just think about position buildings on your terrain.  If you know it's all flat, then how you place the object is quite simple.  But think about placing a house or building on a random matrix.  A bit more complex task, if you want it to look right.  From now on in code, I will only use this function name, get3DMouse(), and these variables, targetX#, targetY#, targetZ#, when referring to either of the above 3D mouse selection routines.

newover#=over#-(screen width()/2)
newup#=-up#+(screen height()/2)
bestover#=newover#/2
bestup#=newup#/2

function placeObjectWithMouse(objType as byte)

if object exist( 2 ) = false then exitfunction

setObjectAtMouse()

if mouseclick() = true
  if onButton = false

    rem Find an object number
    repeat : inc exist, 1 : until object exist( exist ) = 0

    rem Make so the placing object become more invisible
    set alpha mapping on 2, 20

    rem Array stuff
    array insert at top objects()
    ac = array count( objects() )
    objects( ac ).objectNum = exist

    objects( ac ).objectAngleX = object angle x( 2 )
    objects( ac ).objectAngleY = object angle y( 2 )
    objects( ac ).objectAngleZ = object angle z( 2 )

    rem The objects
    select objType
      case 1

        make object box objects( ac ).objectNum, objects( 0 ).sizeX, objects( 0 ).sizeY, objects( 0 ).sizeZ
        position object objects( ac ).objectNum, object position x( 2 ), object position y( 2 ), object position z( 2 )
        rotate object objects( ac ).objectNum, objects( ac ).objectAngleX, objects( ac ).objectAngleY , objects( ac ).objectAngleZ

      endcase
    endselect

    sync : repeat : until mouseclick() <> 1

  endif

else

  set alpha mapping on 2, 90

endif

endfunction

function placeObjectWithMouse()

if object exist( 2 ) = false then exitfunction

setObjectAtMouse()

if mouseclick() = true

  rem Find an object number
  repeat : inc exist, 1 : until object exist( exist ) = 0

  rem Make so the placing object become more invisible
  set alpha mapping on 2, 20

  rem Array stuff
  array insert at top objects()
  ac = array count( objects() )
  objects( ac ).objectNum = exist

  objects( ac ).objectAngleX = object angle x( 2 )
  objects( ac ).objectAngleY = object angle y( 2 )
  objects( ac ).objectAngleZ = object angle z( 2 )

  rem Place the object
  make object box objects( ac ).objectNum, 10, 10, 10
  position object objects( ac ).objectNum, object position x( 2 ), object position y( 2 ), object position z( 2 )
  rotate object objects( ac ).objectNum, objects( ac ).objectAngleX, objects( ac ).objectAngleY , objects( ac ).objectAngleZ

  sync : repeat : until mouseclick() <> 1

else

  set alpha mapping on 2, 90

endif

endfunction

function sizeObject()

if object exist( 2 ) = false or keystate( 82 ) = true or keystate( 83 ) = true then exitfunction

objects( 1 ).objectAngleX = object angle x( 2 )
objects( 1 ).objectAngleY = object angle y( 2 )
objects( 1 ).objectAngleZ = object angle z( 2 )

rem Size the object at the z-axis
if keystate( 72 ) = true
  delete object 2
  inc objects( 1 ).sizeZ, 1
  make object box 2, objects( 1 ).sizeX, objects( 1 ).sizeY, objects( 1 ).sizeZ
  rotate object 2, objects( 1 ).objectAngleX, objects( 1 ).objectAngleY , objects( 1 ).objectAngleZ
  color object 2, rgb(255, 0, 0)
  setObjectAtMouse()
  set alpha mapping on 2, 90
endif

if keystate( 80 ) = true
  delete object 2
  dec objects( 1 ).sizeZ, 1
  if objects( 1 ).sizeZ < 1 then objects( 1 ).sizeZ = 1
  make object box 2, objects( 1 ).sizeX, objects( 1 ).sizeY, objects( 1 ).sizeZ
  rotate object 2, objects( 1 ).objectAngleX, objects( 1 ).objectAngleY , objects( 1 ).objectAngleZ
  color object 2, rgb(255, 0, 0)
  setObjectAtMouse()
  set alpha mapping on 2, 90
endif

rem Size the object at the x-axis
if keystate( 75 ) = true
  delete object 2
  inc objects( 1 ).sizeX, 1
  make object box 2, objects( 1 ).sizeX, objects( 1 ).sizeY, objects( 1 ).sizeZ
  rotate object 2, objects( 1 ).objectAngleX, objects( 1 ).objectAngleY , objects( 1 ).objectAngleZ
  color object 2, rgb(255, 0, 0)
  setObjectAtMouse()
  set alpha mapping on 2, 90
endif

if keystate( 77 ) = true
  delete object 2
  dec objects( 1 ).sizeX, 1
  if objects( 1 ).sizeX < 1 then objects( 1 ).sizeX = 1
  make object box 2, objects( 1 ).sizeX, objects( 1 ).sizeY, objects( 1 ).sizeZ
  rotate object 2, objects( 1 ).objectAngleX, objects( 1 ).objectAngleY , objects( 1 ).objectAngleZ
  color object 2, rgb(255, 0, 0)
  setObjectAtMouse()
  set alpha mapping on 2, 90
endif

rem Size the object at the y-axis
if keystate( 78 ) = true
  delete object 2
  inc objects( 1 ).sizeY, 1
  make object box 2, objects( 1 ).sizeX, objects( 1 ).sizeY, objects( 1 ).sizeZ
  rotate object 2, objects( 1 ).objectAngleX, objects( 1 ).objectAngleY , objects( 1 ).objectAngleZ
  color object 2, rgb(255, 0, 0)
  setObjectAtMouse()
  set alpha mapping on 2, 90
endif

if keystate( 74 ) = true
  delete object 2
  dec objects( 1 ).sizeY, 1
  if objects( 1 ).sizeY < 1 then objects( 1 ).sizeY = 1
  make object box 2, objects( 1 ).sizeX, objects( 1 ).sizeY, objects( 1 ).sizeZ
  rotate object 2, objects( 1 ).objectAngleX, objects( 1 ).objectAngleY , objects( 1 ).objectAngleZ
  color object 2, rgb(255, 0, 0)
  setObjectAtMouse()
  set alpha mapping on 2, 90
endif

endfunction