Re: Proposal for new last call issue: Some unprocessed headers should stay

Jacek Kopecky wrote:

> I think the table without the 'Assumed' and 'mU' columns was 
> better because rows with Assumed=No are all the same and rows 
> with mU=Yes are again all the same. So the original table seems 
> to be more concise and expressing the same information, 
> therefore better.

I tend to agree for "mU"; that's why I did not add it in my first 
revision of Gudge's table.

I respectfully disagree for "Assumed", only if for an education 
exercise. I think there is general tendency to collapse "Assumed" 
and "Understood" into one single concept, which we should resist.

The aim was also to point out that only 1 row out of 3 are 
different for "relay" and "next", hence the great similitude. 
IMO, this was less apparent in Gudge's original table, since 
there were only 2 rows and 1 was different.

Also, see below.

> Anyway, by not having Assumed=No row for the Relay role are you
> suggesting the interpretation that relay must be assumed by 
> everybody?

Indeed it was my assumption that this role had to be assumed by 
all nodes, just like "next". Did I read too much into Noah's (et 
al) proposal?

Interestingly, haven't you just given a justification for the 
extra "Assumed" column?  ;-)

Jean-Jacques.

Received on Monday, 21 October 2002 04:32:31 UTC