Re: reinventing ASCII?

ram@[EMAIL PROTECTED]
 (Stefan Ram) writes:

> "Charlie Gordon" <news@[EMAIL PROTECTED]
> writes:
>>>I'd make '0' - '9' and 'A' - 'Z' contiguous ;)
>>Do you mean '9' + 1 == 'A' ?

Probably.  And while it would make hexadecimal conversion a bit
simpler, I don't find it really im****tant.

>   Some aspects of the original design where:
>
>       - Digits and symbols they are usually paired with 
>         on a keyboard should differ only by one 
>         bit in order to simplify keyboard design
>         (i.e., »1!«, »3#«, »4$«, and so on).

This is certainly not relevant anymore.  Besides, the placement of the
symbols on the ****fted numeric keys is not the same in all countries.
For example, the standard Danish keyboard layout has '(' and ')' over
8 and 9 insted of over 9 and 0.  Some older typewriters don't even
have separate 0 and 1 keys, requiring you to use O and l instead.

>       - All (uppercase) letters should reside within 
>         a single 5-bit block.

Less relevant now than then, but not completely irrelevant.

>       - The lowest four bits of a digit should represent
>         its value.

This is on par with having A-Z directly after 0-9: It makes text to
number conversion slightly easier, but not by much.  I write c-'0'
rather than c&0xf to find the value of a digit, as I find the
intention much clearer.  And it doesn't take any longer to do on
modern hardware.

>       - Due to technical reasons, »0000000« and »1111111«
>         had to be control codes (NUL and DEL, respectively). 
>         Therefore, the outer regions of the code were used 
>         for control codes, while printable characters were 
>         placed in the middle.

This is almost irrelevant now, except I shouldn't wonder if some
hardware would break if it was changed.

	Torben