RP> HR>RP>Some one said that there is a memory penalty for name space,
RP> HR>RP>there is a bigger penalty on disk space, which with our 4 Gig
RP> HR>Bigger?  Hmmm..
RP> HR>The memory calculation is:
RP> HR>.023 * volume size (in MB) / blocksize ... for each DOS volume
RP> HR>.032 * volume size / blocksize ... for each volume w/added name space
RP>I don't see 40% there, you do not use both figures together with name 
space,
Blind, are you?     Of course you don't use both figures on the
same volume, but if .032/.023 isn't 1.3931, then one of us needs to go
back to the fifth grade.
RP>calculatiopn assumes worst case, ie., every file name is long and at 50% 
f
RP>maximeum length, which does not happen in practice.
RP> HR>I consider an increased memory requirement of approx. 40%
RP> HR>QUITE significant, and your quarter of a percent of drive space
RP> HR>inconsequential!
An additional entry is made in DET (Directory Entry Table) for each
added namespace, regardless of whether the filenames are actually long
or not!  The length of the filenames matter not one iota.
With ONE added namespace, the DET is twice the size as it is with just
the native DOS name support. Add two namespaces and its TRIPLE the size.
RP>Yip. The increas in memory (assuming worst case as above) would be 33% 
(appr
Where in the world do you get 33% anyway?  Make this stuff up as you go
along do ya?
RP>of the directory cache memory, which is nowhere near the total memory.
Which is why adding a namespace only requires an additional 40% and not
TWICE the total system memory.
If the math is to deep, God help you!
If you disagree with the SPEC, talk to the people
that wrote it....Novell
HTR
---
 . OLX 2.2 . Teaching a pig to sing wastes your time & annoys the pig
---------------