Re: Disk II Table Translation

["Bryan Parkoff" <nospam@nospam.com>]

"Michael J. Mahon" <mjmahon@aol.com> wrote in message 
yPudnVdjOalvU1bYnZ2dnUVZ_tGsnZ2d@comcast.com">news:yPudnVdjOalvU1bYnZ2dnUVZ_tGsnZ2d@comcast.com...
> Bryan Parkoff wrote:
>>     I notice that table translation for Disk II has 81 valid out of 256 
>> ASCII codes, but 64 values are used only.  Here is 17 unused values 
>> below.
>>
>> 92, 93, 94, 95, 99, 9C, A4, A5, A9, AA, C9, CA, CC, D2, D4, D5, & E4
>>
>>     Why didn't Disk II use 17 values for data buffer?  I am aware which 
>> values are used for address field and data field.
>
> As Alex pointed out, powers of two are what matter.
>
> But these codes (except for D5 and AA, which appear only in headers) are
> not used because they do not meet the requirement for *only* one pair of
> consecutive 0 bits and *at least* one pair of consecutive 1 bits in bits
> 0-6.

>In addition a very short piece of code in Disk][ ROM builds the
>table of valid bytes.

    It makes sense.

> Relaxing the "at least one pair of 1 bits" rule would allow 95, A5, CA,
> and D4 as additional "delimiters".  In practice, the Disk ][ Controller
> can reliably read other nibbles as well, but not enough to allow 128
> different nibbles, since that would require *all* combinations of bits
> 0-6.

    Do you allow A9, D2 and D5 like you said 95, A5, and CA?  Does it mean 
that 8 bit value can't have two pair of zero bits?  It can only have one 
zero bit and pair of two zero bit combination in any 8 bit value.

Bryan Parkoff