The number of nodes on the
system at any time consists mainly on the
following things (in descending order):
The math here is really quite
simple. Let's take the numbers from right now:
August 8, 2001
925,709 nodes
500,176 writeups
37,626 users
Approximately 500 other items (
superdocs,
themes,
nodelets,
usergroups,
aliases,
edevdocs, theme settings and containers,
dbtables,
nodetypes, etc). This is a rough count from the big group at
List Nodes of Type (edev only there cowboy).
- 925,709 Writeups are still nodes as counted by the stats nodelet, so lets sub off the number of writeups from the number of nodes. 925,709 - 500,176 = 425,523
- 425,523 Users are a node, so let's subtract them off. They have a homenode, but it is not a separate thing; it is really their user id being displayed through its default display method.
425,523 - 37,626 = 387,907
- 387,907 Let's also take off our generous estimate off of the other nodes. 387,907 - 500 = 387,407
- 387,407 Now what does this number mean. This is a rough number of e2nodes (the stuff nodeshells and node parents are made of). Let's try to find a statistic: the number of writeup per e2node. 500,176 / 387,407 ~= 1.29 writeups per node.
Is 1.29 writeups per node a sane number? With the amount of
nodeshells being and one writeup nodes (obscure webby references* included) balanced out with huge
gathering writeups,
daylogs, and other
e2node titles that tend to gather a huge writeup
inclusion, I think that we have a suitable "sane" number.
It is important to remember that
nodeshellers have nothing to do with that
statistic. Writeups will never catch up to
nodes (because nodes contains the number of writeups). The
terminology the stats nodelet uses, and what the vernacular here on
E2 throws around in the concept of a "node" are mildly different, and thus it can be
misleading. Whenever you node something for the first time, you actually create two
nodes, one for the
parent shell, and the other for the
writeup. They have the same
title, but different
nodetypes.
Notes:
*Webster definitions still count. They are merely writeups of type definition.