Re: RBB: Really Big Bitvectors

From: Ian Macintosh (ian@iman.internet.co.nz)
Date: 12/08/95


On Sat, 9 Dec 1995, Paul Cole wrote:

> Now for the method that does work but is highly proprietary.

> But close inspection would show that you can use a mere 2 bits for these 4
> values becuase RUNNING is only possible when you are WALKING and DIVING is
> only possible when you are SWIMMING. Also, I note that WALKING is the same
> thing as NOT FLYING. So you could <though I don't, too much work> use 2
> bits to represent this. 
> 
> XY where
>     X -- 1 = SWIMMING, 0 = WALKING
>     Y -- 1 = DIVING,   0 = RUNNING
> or 00 = NOT SWIMMING and NOT RUNNNING = WALKING
>    01 = NOT SWIMMING and RUNNING
>    10 = SWIMMING     and NOT DIVING
>    11 = SWIMMING     and DIVING
> 
> Thus I've just implemented all 4 STATES in 2 BITS.
> 
> This is very proprietary and imho, not terribly useful.  It is only 
> presented here as a method for representing more than X FLAGS in an Xbit 
> bit field.

That is not proprietary.  I was using logic like that back in 1981.  To
lay a claim of 'proprietary' on that is ludicrous in the extreme.  I was
not the only one doing that type of logic either, as I recall seeing it in
an old accounting system written to run on the TRS-80 model I. 

Your comments on bitfields, etc, is however, highly correct.  Perhaps you 
could add that when viewing a situation where any combination is 
possible, it is in reality a permutation.

	Ian.



This archive was generated by hypermail 2b30 : 12/07/00 PST