Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: Big B, little b




Geoff,

Pat Thaler kicked off this thread with a question about what we call the
64/66 code (i.e. 64b/66b or 64B/66B). 802.3 2000 may be the dog (I don't
like the implication :-) but the issue in P802.3ae is much broader than
just what 802.3 2000 says about 8B/10B. Additionally, Bob Grow said, to
this reflector on more than one occasion, that he is going to submit a
No vote and a TR comment if I use 8B/10B in clause 48. Bob is also
suggesting that the dog needs a bath.

Best Regards,
Rich
     
--

Geoff Thompson wrote:
> 
> Colleagues,
> 
> I do not understand why this is all being discussed, much less at this length.
> The abbreviation for octet into 10 bit block coding was established in P802.3z.
> That decision is entrenched throughout the existing IEEE Std. 802.3 as
> "8B/10B".
> The current project, P802.3ae (tail), is an Amendment to IEEE Standard
> 802.3 2000 Edition (dog).
> 
> Dog wags tail.
> Tail does not wag dog.
> 
> I have heard nothing to make me believe otherwise.
> We get points with the world for how little we change, not how much.
> 
> Geoff
> 
> At 07:14 AM 11/28/00 -0500, Mark Ritter wrote:
> 
> >Some history might help to resolve this thread.
> >The '8B/10B' acronym demonstrably does not trace back to IBM. We simply
> >adopted  the code classification scheme mX/nY which had been widely used in
> >technical literature long before the Fibre Channel code was developed. X
> >and Y stand for the number of transmission levels (Binary for two, Ternary
> >for three, etc.), m and n indicated the number of symbols at the respective
> >levels. So 8B/10B implies that 8 binary symbols are translated into 10
> >binary symbols. Codes of the class 4B/3T translate 4 binary symbols into
> >three ternary symbols. This notation is used in major textbooks and IEEE
> >publications as any search for the terms quickly reveals. In a context
> >where nothing but binary systems are discussed, one might want to simply
> >drop the big B and shorten the acronym to 8/10. Both 8/10 and 8B/10B are
> >deficient insofar as they are not referencing a specific 8B/10B code of
> >which there are many. How about FC8/10 as abbreviation for Fibre Channel
> >8B/10B code?
> >
> >Albert Widmer, IBM T.J. Watson Research Center. EMAIL: widmer @us.ibm.com
> >
> >Mark B. Ritter
                                 
------------------------------------------------------- 
Richard Taborek Sr.                 Phone: 408-845-6102       
Chief Technology Officer             Cell: 408-832-3957
nSerial Corporation                   Fax: 408-845-6114
2500-5 Augustine Dr.        mailto:rtaborek@xxxxxxxxxxx
Santa Clara, CA 95054            http://www.nSerial.com