On Thu, 18 Mar 1999 [EMAIL PROTECTED] wrote:
> Allow me to throw my hat into this one. It looks like many people are
> looking for a way to roll up the reporting to what would be considered
> a summary for a particular subdomain without knowing in advance how
> many dots the name contains. Some time ago we used WWWSTATS to
> produce our log file stats. WWWSTATS uses an approach of allowing you
> to drop the left most portion of the name, which in many cases
> resolves to a specific user. The beauty of this approach is that you
> don't have to know the length in advance. The program did it
> automatically. For those people who still want some additional
> customization, you'd still need to make the subdomain command
> available.
>
I'm thinking as I write here:
It's an interesting idea. I wonder how well it works though? The problem would
be that different organisations' internal naming structure might be quite
different. For example, just looking at the top of our current logfile, we
have etide03.microsoft.com which would become microsoft.com, followed on the
very next line by s06.austin.ibm.com which would become austin.ibm.com, which
seems to be one layer deeper in the hierarchy -- so you would be comparing
apples with oranges again. This is why I took the hierarchical approach. You
might also get ursa.cus.cam.ac.uk -> cus.cam.ac.uk, which is more like
austin.ibm.com, but seems even deeper -- microsoft is a company, austin.ibm is
one division of a company, and cus.cam feels like just one computer, it just
happens to run on two different servers.
These shouldn't be taken as more than scattered thoughts. But I want to be
careful to find solutions which are appropriate to more than just 80% of
companies in the .com domain.
--
Stephen Turner [EMAIL PROTECTED] http://www.statslab.cam.ac.uk/~sret1/
Statistical Laboratory, 16 Mill Lane, Cambridge CB2 1SB, England
"Ad infinitum, if not ad nauseam." (Interviewee, BBC Radio 4)
--------------------------------------------------------------------
This is the analog-help mailing list. To unsubscribe from this
mailing list, send mail to [EMAIL PROTECTED]
with "unsubscribe analog-help" in the main BODY OF THE MESSAGE.
--------------------------------------------------------------------
