Re: Root level prodir combined results

[[e-mail]]

Well

I don't want to fire out a war

let me just make a few statements
  . we appreciate the unfunded effort, and the results of probing
  . it is not because this is unfunded that it is not worth
      to make an extra effort (very modest) to include
      in each message a couple of lines such as
         this period results were affected by a link problem from
	   rpobing node XXX which account for a rough x% of
	   unavailability

This is certainly obvious when results from different
probing sites show a significant inconsistancy
and must be extremely easy to detect whereas
it is far less obvious once it is aggegated.

As you are not funded, and if you have no time
to "interpret" the results, then I suggest that you
distribute the raw data from each prrobing site.
This will be much more helpfull to the DSA managers
helping them to detect problems (including network problems)
and even less unfunded efforts.

let's take a nexample

if french master availability is 90% from all probing sites
   then there must be a french problem ( up to Katy to
   check wether Renater, the Jussieu LAN, the dione
   computer or the pizarro software is the culprit)

if the same master availability is 97% from all sites but one
then the search for the origin of the problem will
be better focused in between the probing site and
the french master 
   if this probing site stats are all lower than the ones
    from other pprobing sites then OK it is a probing site problem
  if it is specific to the french master, then something should
    be investigated (specific network link, specific interop
    problem...)

I thus suggest that
   either the unfunded effort is complete enough and the analysis
      is done
   either the raw data is distributed so that, anyone is able
     to perform his own analysis

regards,

-- PAP
ZZ