conversion from binary to decimal

disp('conversion from binary to decimal');
z=input('give a floating point number ','s'); 

n=strfind(z,'.'); %index of the . 
z1=z(1:n-1); %z1 vector
disp(z1);
z2=z(n+1:length(z)); %z2 second part after .
x=bin2dec(z1); %bin 2 dec transformation of the first part of the number (before .)

%transformation of the decimal part of the number
b=0.5;
y=0;
for i=1:length(z2) 
 y=y+(z2(i)-48)*b;  %what is -48, it is set for precsion?
 b=0.5*b;
end

fprintf('Il numero in base 10 e'':\n');

-48 is set for a better precision in the conversion from binary to decimal or there is another reason?

Thank you!!

Answer 1

[David Hill]

No, the -48 is converting the ascii character '0' or '1' to a number.

[John D'Errico]

And that happens because '0' happens to be ascii 48.

+('0')
ans =
    48

[Walter Roberson]

In all character sets that I have been able to find since the 1930's, IBM's EBCDIC is the only one that did not encode the characters '0' through '9' as consecutive positions in the table with '0' being the low value (EBCDIC coded '1' as the low value and '0' followed '9' if my memory is correct)

Therefore for all computer languages that you are ever likely to encounter, you can rely on the fact that the character code for a particular single digit "Arabic numeral" is that many positions after the position for the character code for '0'.