[Coco] Why do we need a CoCo 4? (Long irrelevant rant)

[Steve]

Without an U register none of games would run at all.  But first a 
little coding history ...

Back in the 70's, I create a form of Object Oriented Code (OCC) in 
assemble language.  Every thing from the player's ship to the shot 
fire at him were control by uniformed data-blocks of information. To 
create a new object in the game all I need was to setup a new 
data-block filled with the data for that element of the game.

The first byte was the control flags with bits that would let the 
system know if the data-block was in use (alive) or should be display 
on the screen and so on. That was followed by bytes would hold other 
information like the display image type, movement type and position. 
Every bit of information that the object needed was stored in that 
block of data.

What made this system work so well on the Z-80 (and 6809) was their 
16-bit index registers with an 8-bit offset.  I would load the Index 
pointer with the address of the first Data-Block and call that would 
work on the data.  Once the call was done, I move to the data-block 
till all the data-blocks were worked on.

Each call knew just where to find the information they needed because 
the data-blocks had the same layout.  The data at located at 4 bytes 
in was the movement type for the object so an MOVETYPE,IX would get 
that movement code to run on this object.

The bottom line was this data-block system created very tight code 
since all the calls used the same data format.  Also, there was very 
little chance of having a runtime error do to pulling the data from 
the wrong location.

Getting back to why I used the U register for my pointer and not the 
X or the Y registers.  First, the X & Y registers can be used for 
more than index registers, like 16-bit counters since they set the 
zero flag when the LEA -1,X hits zero.  Besides, I had no need for a 
user stack since all my data for an object was stored in the data-block.

Steve (Zaxxon) Bjork

At 07:22 PM 12/30/2008, you wrote:
>Joel
>
>That is what I said. if the HC12 had the U pointer register then it 
>would be the CPU for the next Coco. Infact I told one of the 
>managers of the HC11 line that one the HC11's biggest defects 
>was  it lacked that U pointer register and he agreed with me.
>
>The HC12 is about 80% source code compatible. Object code is not 
>compatible. 6809 code would have to be reassembled. You do have to 
>deal with the PSH*/PUL*  instructions.
>
>james