My objective is to improve Scotty to include many of the desireable traits
of NNM and improve on some of its weaknesses - especially level 2 layout.
I agree with the last commenter's evaluation of Scotty vs NNM. Scotty is
inferior in several ways and superior only in being OSS. However, I think
it is a good start. Just the snmpwalk function is worth the price of
admission.
Some observations to date:
1. A simple modification will change Scotty so it uses an equivalent to
oid_to_sym and oid_to_type to map discovered nodes into appropriate icons.
This I have done.
2. The "ipmap" program (tkined) uses a single canvas and handles "groups"
within a single display environment. This makes large networks more
difficult to represent and navigate. It probably should be modified to open
sub-windows to expand icons.
3. The current structure is not partitioned into the framework of
cooperating processes that NNM has. In particular, the monitor is built
into the map, causing the problems referred to by the last commenter. It
needs to be split up so the monitor updates the status database and the
mapper displays from the status database.
4. I haven't made rigorous tests, but my early observations see the Scotty
monitor function being much more efficient than NNM. They apparently have
put some effort into "multi-threading" the monitor function so it works
several nodes in parallel.
5. I have also translated the ip_layout module into (approximate) English.
I believe I have discovered some errors in this module and there are a
couple of sections that I may not yet fully understand. However, it is in
many ways already better than the HP layout algorithm.
That's about as far as I have gotten with Scotty so far.
-- !! This message is brought to you via the `tkined & scotty' mailing list. !! Please do not reply to this message to unsubscribe. To subscribe or !! unsubscribe, send a mail message to <tkined-request@ibr.cs.tu-bs.de>. !! See http://wwwsnmp.cs.utwente.nl/~schoenw/scotty/ for more information.