[Coco] Y2K fixes ever developed or posted?

[Theodore Evans (Alex)]

On Dec 15, 2003, at 3:21 AM, Robert Gault wrote:

>> That six byte field for the date is good until *2155* without any 
>> changes, or 2027 if you insist on looking at it as a signed value.

> It is amazing how the same arguments show up over and over with no one 
> learning from past experience. The reason we had a Y2K problem to 
> begin with is exactly because people said the equivalent of, "I'll
> probably be in the ground for 70 years or so..."

I didn't say anything about using any technique to extend the life of 
the current arrangement.  Merely stated something about the the clock 
structure for OS-9.  The fact is that there is no time structure which 
takes up a finite amount of space that will not eventually overflow.  
The only question is that of how much is enough.

> Extending an inadequate technique for a few more years should not be 
> dignified by calling it a fix. Postponing the inevitable instead of 
> removing it as a possibility should only be used during the time 
> period a true fix is being implemented.
>
> With the trivial addition of a century byte to the OS-9 defs 
> structure, the OS-9 software clock would be good to the year 9999. If 
> that's not good enough two bytes should more than cover the collapse 
> of our civilization.

There are a number of places where the date is used where there is no 
space for an extra byte.  For example the file descriptor sector 
(probably *the* most important place to put a time stamp).  In any case 
OS-9 doesn't use BCD for date storage, one byte can hold 256 values, 
not 100 and two can hold 65,536 values, not 10,000.