I have read draft-bormann-6lowpan-ghc-04.
The draft is well written, simple and useful, and I support the draft
to be an RFC.
I have one comment on code byte representation in the draft.
Please see below.
The code byte representation such as "0kkkkkkk" is confusing.
Which does it mean 1) or 2) below?
1) 00000000 (for k=0b0) or 01111111 (for k=0b1).
2) the least significant 7 bits of the code byte value has a
decimal value of k, where the right-most bit is the least significant bit.
If it means 2), then representation of code byte should be changed
to be more accurate. Here is the suggested change for the code bytes
definition in section 2:
"
The most significant bit: The leftmost bit in the code byte.
c_i: The i-th significant bit of code byte
(c_0 is the most significant bit)
sum_{i=x}^{y}: sum with increasing variable i from x up to y by 1.
* (wildcard bit): The bit can be 0 or 1.
The code bytes are defined as follows:
Code Byte Action Argument
----------------------------------------------------------------
0******* Append k = sum_{i=1}^{7} c_i*2^(7-i) k bytes
bytes of data to bytecode argument (k<96). data
1000**** Append sum_{i=4}^{7}{c_i*2^(7-i)}
bytes of zeros.
10010000 STOP code
101***** Set up extended arguments for a backreference:
sa += sum_{i=4}^{7} c_i*2^(10-i);
na += c_3*2^3
11****** Backreference:
n = na + sum_{i=2}^{4} c_i*2^(4-i) +2;
s = sum_{i=5}^{7} c_i*2^(7-i) + sa + n;
append n bytes from previously output bytes,
starting s bytes to the left of the current
output pointer; set sa=0, na=0.
-----------------------------------------------------------------
"
Best Regards,
Yoshihiro Ohba
_______________________________________________
6lowpan mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6lowpan
