Qin Wu wrote:
[Qin]: I can understand 8-octets accommodate 1-octet record, but How does 8-octets data block accommodate 12-octet record, I can not understand.
The draft does not say that data block sizes are multiples of 8 octets, and indeed they need not be multiples of 8 octets. Data blocks are merely concatenations of octets, with no padding or alignment. If a field requires only 1 octet, exactly 1 octet is appended to a data block. If it requires 8 octets, exactly 8 octets are appended to the data block. If it requires 12 octets, exactly 12 octets are appended to the data block. So it doesn't matter how many octets a field requires: whatever number that is, that's the number of octets that appear in the resulting data block.
[Qin]: how does 44 octets header can be inserted into such "header for 32-bit transitions"? Does header for 32-bit transitions stand for 4 octets header?
The header for 32-bit transitions is a 44-octet header. There is no "insertion" of one header into another; there's just the one 44-octet header.
[Qin]: You are using plura, i.e., "32-bit **transitions**".
Sorry, I've lost context. I guess you are suggesting that we make a minor wording change to clarify a point about the 32-bit transitions, but I don't know what the point is. Could you suggest a specific wording change that would make things clearer?
