This file contains notes from Charles Powell <cp98@zelda.wss.yale.edu>
(the author of this code).

[Blame the order of the paragraphs or any incomprehensibility on me (Bruce).
I just plugged paragraphs from separate e-mail messages from Chuck into
this one file.]

I have forwarded along to Bruce in a separate email a proposed
extension to SYSSUMM.  The extension allows you to store the data
collected by SYSSUMM in a database (right now just MySQL but ostensibly
any database supported by Perl's DBI module).  The extension would be
completely optional and would only require that you have the database
and DBI.PM loaded on the server.  No changes are required on the client
(remote) side whatsoever.

The two scripts I have submitted are fully functional but do not yet
implement all the features I envision.  They are pretty safe and
unlikely to cause any damage in a production environment and certainly
do not require you to change the way you currently do business.  They
are extremely ugly code and need a lot of abstraction work and thoughts
towards efficiency.

As some "teasers" of what I want to do consider the following
scenarios.  It's Weds PM when you get the word that there is a new
exploit being run against pam-0.64.  You want to use SYSSUMM to find out
which of your machines have this particular software installed so you
can install the new version.  Or you get an irate phone call from a user
who says their machine is carping about an IP conflict with a machine
with MAC/EHA address 00:0f:98:78:0e how do you find that needle in the
proverbial haystack?

There are only 3 files involved, a table structure file that tells you
how to set up your adatbases in MySQL.  A "slurper" that currently takes
feed from a find command to get all the syssumm info and load it into the
database.  Finally a chuckles.pl which is the equivalent of sysquery.pl it
would go wherever was appropriate.

The only other things needed (and only on the server) would be MySQL
(although that is pretty portable and a lot of databases are supported)
and the DBI.pm module.

Right now I plan on kicking out about a new release every week or so to
quickly add in functionalities.  I am submitting the code hoping that at
least a few other people (hopefully those with more talent in Perl than
I) will want to add their talents to maintaining.  If the interest is
there I think we can push it along quite quickly.  If no one is
interested or it is a bad idea for others then we'll just let it
publicly die.

My understanding is that Bruce is going to add it to the CVS repository
and MANIFEST as his schedule permits.  Look forward to hearing from you.

