[sigcomm] considerations for reviewing extended papers

[Greg Minshall]

Mark,

2 things.

1.  to my knowledge, we are *not* considering "non-referred" papers (which 
includes tech reports and, as of i last knew, internet drafts).  the question 
is precisely "workshop papers", i.e., papers submitted to a venue for earlier, 
draftier, less complete work, with reviewing, however.

i think everybody's formulation (C-SUMi(Wi) > delta) && (C >> CONST_LPU) makes 
sense.

(where "C" is conference paper contribution; "Wi" is contribution of workshop 
i paper; ">>" is "much greater than"; "CONST_LPU" is "least publishable unit".)


2.

> But, my strongest opinion is that whatever we do we need to be
> clear and explicit---which is tough because all the words are
> overloaded.

network protocols need to be clear and explicit.  here, we should think more 
in the terms of the British admiralty (say) giving instructions to an 18 
century captain they wouldn't see for 3 years (in our case, to multitudes of 
PCs over some small number of years).  there needs to be enough flexibility 
for different PCs to adapt to the needs of their moment.

cheers, Greg