NetWare(R)
Accounting
Analysis
System

Author:
S Sharpe
4 Drovers Way
Peebles EH45 9BN
Scotland

100614.3441@compuserve.com

**************************************************
The Authors disclaim all warranties
as to this software, whether express
or implied, including without limitation
any implied warranties of
merchantability, fitness for a particular
purpose, functionality, data
integrity or protection.
**************************************************
Click on a topic of interest highlighted below:

Packing List
Installation
Quick start
Introduction
Netware(R) accounting data
Multi-server analysis
Entering data analysis options
Data analysis
Analysis methodology
Boot.log
Problems/Limitations
Command line options
**************************************************

Packing list
------------
This software consists of the following files:

 README.1st - A text file including a quick start guide (also included in 
 the manual)
 PCAUDIT.exe - The analysis program
 PCALIAS.nas - An example PC alias file
 USRALIAS.nas - An example User alias file
 MANUAL.hlp - A manual explaining how the program works in Windows Helpfile 
 format
 MANUAL.txt - A text printout of the Windows Helpfile
 NAAS.hlp - The helpfile used by the application to provide context-sensitive help at run-time
 FORM.txt - A form for obtaining further copies or higher capacity software
This version includes a number of minor enhancements since ver 1.02, 
including better handling of XMS memory and more error checking for 
incorrectly configured PCs.  The major enhancement has been the inclusion 
of linked list processing to speed analysis of large networks, e.g. where 
the number of PCs or Users exceeds 200.  This slightly slows the speed of 
analysis for small networks, but vastly speeds the analysis process for 
large or very large networks.
This will be the last DOS version of the NetWare Accounting Analysis Program.  A Windows version will be issued in the near future, and this is likely to use a static database compatible with Microsoft Access for more efficient data analysis over a period.

Installation
------------
To install this software, simply copy the contents of the distribution disk 
to a convenient directory on your local hard disk or network drive.  The 
software will operate whether you are logged in to the Netware(R) network 
or not, so operation can conveniently be carried out from a local hard disk.  Alternatively, type 
A:INSTALL
Note:  This program will not operate within a Windows(R) DOS-box unless you 
have enabled XMS memory emulation using a suitable PIF file.  Refer to your 
Windows User Guide.
When operating the program from DOS, the program requires access to XMS 
memory.  This requires you to have installed a suitable memory manager, 
such as HIMEM.sys.  The actual memory required depends on the licence limit 
of the software purchased.  A check is carried out at run-time and the 
program will not execute if there is insufficient memory.
Note that a 486 or faster CPU is recommended, because of the CPU-intensive 
nature of this application.  You may run the program on a 286 or 386, but 
a warning will be issued.  The program speed is also very much affected by 
your drive performance, and it is recommended that you process files on 
local hard disks rather than via network files.

Working directory
Please note that the program requires access to a number of support files 
for proper operation.  To access these files the program MUST be executed 
from its installation directory.  If you attempt to execute the program from 
a different directory an appropriate error will be issued.

Alias files
At this stage you can optionally create some alias files to help decipher 
your analysis.  An alias file is a lookup file that is used by the program 
to give more informative information about objects (PCs or users) that are 
listed in report files. 
PC alias file:  this is called PCALIAS.NAS and consists of plain ascii text 
(e.g. as created using the DOS EDIT or Windows Notepad programs).  One alias 
per line can be specified using the following format:
station address=alias text
where station address is the ethernet address of a PC (e.g. 0001B3459BEF) 
and alias text is any identifying text for this PC (e.g. the name of the 
owner, department ID etc.).  If the NAS file cannot be found in the same 
directory as the PCAUDIT.exe program, alias columns are left blank in the 
final report.  If a NAS file exists, each result is compared with the alias 
file and the alias text displayed where a match was found.
User alias file:  this is called USRALIAS.nas and has a similar format to 
the PC alias file.  However, bindery object ID is used instead of station 
address, e.g.
0B13FE=My PC

Quick Start
-----------
To try the program out straight away, do the following:

1. Move your accounting file to a handy location on your hard drive (that 
also holds the installed software), e.g.
     COPY F:\SYSTEM\NET$ACCT.DAT C:\ACCT\MYSVR.DAT
     DELETE F:\SYSTEM\NET$ACCT.DAT

2. Simply type:

	PCAUDIT

You will be presented with a number of prompts to define your analysis.  
Enter the following at these prompts:

You see...>> You enter...
Specify source file:.>> MYSVR.DAT
Specify ADDITIONAL source file (return to continue):.>> Hit enter or return
Output format:  U=by user  P=by PC [by PC]: >> Hit enter or return
Output format:  P=plain text  C=csv [CSV]:  >> enter P
Give output file for results: PC.TXT.>> Hit enter or return
Give start date (mm/dd/yy) [02/01/95]:.>> Hit enter or return
Give end date (mm/dd/yy) [02/14/95]:.>> Hit enter or return
Prime time START (hh:mm) [08:00]:.>> Enter the start of a normal working day
Prime time END (hh:mm) [16:30]:.>> Enter the end of a normal working day

If you are operating with a NetWare client shell loaded (whether or not you 
are logged in) you will also see:
Give (optional) user to send completion message to:.>> Hit enter or return

All parameters entered correctly? (Y/N).>> Y

The program will then start to analyse the system account file for your 
current Netware(R) server.  Progress is displayed and output will be stored 
in PC.TXT in the current directory.  When the program starts to find records 
beyond the end date (or the end of the file is reached) the output will be 
formatted and the program will stop.
If you get a message asking if you want to adjust analysis periods, simply 
hit return or enter to select this option.  The total number of days is 
automatically adjusted to match the total found in the analysed data.
Use the DOS editor EDIT or the Windows editor NOTEPAD to view the resulting 
text file.  Alternatively repeat the analysis hitting return or enter at the 
output format, the result will be the default comma separated value file 
that can be directly imported into Microsoft(R) Excel or Lotus(R) 1-2-3.
A more detailed description of program operation is given in the following 
sections.  At most points in the program's operation you may hit F1 for help.

Introduction
------------
The Netware(R) Accounting Analysis System was written because the authors 
could find no comparable product readily available to analyse the Accounting 
records of Netware servers.  Accounting records of Login and Logout events 
are recorded by all Netware servers with accounting enabled (or 'installed' 
as the manuals term it).  It is not necessary to enable billing for this 
feature to be used.
This analysis program does not simply list the accounting records in an 
easier to read form:  it attempts to provide a useful analysis, answering 
management questions such as 'how many PCs are used more than 50% of the 
time during normal working hours'.
If you have too few PCs and suspect some are under-utilised, this is the 
ideal program for your needs.  If you want to buy more PCs and want to argue 
the case that your existing ones are over-utilised, again this program is 
for you!

Netware versions
The program has been tested on accounting files from a variety of Netware 
servers, including v3.11, 3.12 and 4.01.  Tests have not yet been carried 
out on Netware 4.1 or 2.x.  If your particular accounting files prove 
problematic, feel free to zip them to a reasonable size and mail them to 
us for further tests, including a description of your Netware version and 
billing/accounting options installed.

Netware accounting data
-----------------------
Netware can potentially record a large quantity of accounting data, ranging 
from login and logout records to charging data of various sorts.  Accounting 
data can also include records of intruder detection, server boots and 
console time changes.  Refer to your Netware manuals for more information.
All Netware accounting data is recorded in the file sys:system\net$acct.dat.  This file generally grows until someone notices it, at which point the data is either deleted or archived and forgotten about.  A new accounting file then starts growing...
If you cannot find this accounting file,  you need to install accounting.  
Login in with supervisor-equivalent privileges.  Run syscon.  Select the 
accounting option and list the accounting servers.  If your server is not 
present, press Insert and add it.  Accounting is now enabled for your server.  You need do no more to use this program.

Analysing accounting data
Limitations

Analysing accounting data
-------------------------
Having enabled accounting on your Netware server, the next step would be to 
analyse the data.  However, Novell provide only a basic list facility 
through the PAUDIT program.  This program lists the contents of the current 
net$acct.dat file in ASCII format, three lines per record.  The result can 
be very verboseand not much use.  What's more, it's hard to analyse any 
accounting data in a file of different file name or location.
There are shareware accounting programs aboutmost notable of which is 
PAUDIT2.  This improves on paudit by allowing any accounting data file to 
be listed, and by providing the data in a more useful format as one line 
per record.  It also provides simple character-graphic display of login/
logout patterns.  However, that is the extent of its usefulness.
If you want to make intelligent use of the accounting data you must now 
import the data in to a database and write clever SQL statements, or write 
a program to analyse the data (whether via paudit or paudit2) yourselfor 
use this program.
The emphasis in this first version of the Netware Accounting Analysis 
Program is to provide useful network utilisation figures for PCs and Users.  
Specifically, graphs such as the one shown below:

This sort of graph can be produced directly from the results of the analysis 
produced by this program.  Having produced this graph one often then asks 
so which are the PCs that are only used 1% of the time or less?.  The raw 
data to find these is also recorded by the program.  This is presented in 
the form of the PCs station address (or ethernet address).  Use the Netware 
USERLIST /A command to see who has which station address.
Of course, PCs that are used very little are unlikely to appear as logged-in 
machines when you try to find them.  One approach to knowing your station 
addresses might be to record the station address and login names to a file 
for a period at login by modification to the system login script, 
f:\public\net$acct.dat.
Alternatively (and this is the approach we use), you could implement an 
automatic inventory manager on your servers, which scans PC hardware and 
software each time a user logs in then updates a central database.  It is 
then a simple matter to query this database by station address to find out 
the normal user etc.  The software we use is part of Intel's LanDESK Manager 
suite.
A further advantage of this sort of system is that you can then look at the 
software installed on a particular PC to see if it is likely that use is 
made of the PC when not logged in to the network.  Our particular policy is 
to install no software (except network client software) on each PC.  This 
means that when the program shows very low utilisation, it actually means 
the PC is being used very little.

Limitations
-----------
One of the most irritating limitations you will find with this program is 
that, whereas PCs are identified by station address (which is quite useful), 
users are identified by bindery object ID (which is a pain).  It is hoped 
to address this in the next version, if enough interest in doing so is shown 
by registered users.  If this is important to you, let us know when you 
register.
Bindery object ID is the form in which Netware records the stats.  To 
convert to login ID would require access to the binderyand since the 
program was originally designed to allow analysis on stand-alone PCs it 
was not possible to ensure this.
If you want to convert to login ID, either make a list of your login ID's 
Vs object IDs by looking in other information in the user information 
section of SYSCON, or take a look in the f:\mail directory.  The f:\mail 
directory contains a directory for every user on the server, named by object 
ID.  To find out the user ID related to a particular object ID, simply list 
the trustees of the appropriate mail directory (by using TLIST).  This is a 
tedious method of preparing the user alias file - a program to do this will 
be included in the next version of the program if enough interest is shown 
by registered users.
Data earlier than 1992 will not be analysed.  Attempts to analyse such data 
will result in an 'corrupt or out-of-range data' error.
Finally, the overall program dimensions result in limitations of 2500 PCs in 
any one analysis sweep (irrespective of number of files being analysed) and 
5000 users (ditto).  The actual limit for your program depends on the 
licence fee paid.

Multi-server analysis
---------------------
In a typical medium to large company users will log in and attach to 
multiple Netware servers.  To cope with this situation the program can 
be configured to intelligently analyse and merge multiple data files.  
Two data merge options are provided:
 
 Inclusive data merge
 Exclusive data merge

Inclusive data merge takes each data file in turn and merges all results 
together:  if PC A (or user B) has data records in more than one data file 
then the statistics relating to that PC are added together.  Conversely, 
Exclusive data merge intelligently merges data.
Exclusive data merge takes each data file in turn and ignores data for those 
PCs (or users) where it appears in subsequent files.  Should new PCs (or 
users) appear in subsequent files, these will be included.

Why provide Exclusive data merge?
Imagine the situation where a user logs in to more than one server, e.g. 
where applications are on one server and data on another.  Typically adding 
up the utilisation of each server by a particular PC or User will result in 
login time being calculated as twice the real time that the PC was used or 
the User was occupied.  Since this program attempts to provide an accurate 
picture of PC or User utilisation - not server utilisation - this situation 
must be accurately managed.  Hence the option for exclusive data merge.
Thus where data comes from more than one data file the logic flow for 
Exclusive data merge goes something like: 
1. Read file A
2. Store data for PCs or Users in file A
3. Read file B
4. Check if each PC or User already exists in file A
		- if Yes: ignore the data
		- if No: store the data
5. Repeat in turn for the other data files (if any).    

Entering data analysis options
------------------------------
Various options are requested when you first define your analysis.  Some of 
these are truly optional, others require a particular choice to be specified.  The meaning of these options and the available choices are summarised below, in the order in which they appear when you first run the program.

Specify source file
Output format
Give output file for results
Give start/end date
Working days in period
Start/end of normal working day
User to notify
Confirm parameters

Specify source file
-------------------
This is the first data file you want included in your analysis.  You can 
enter up to eight data files for inclusion in the analysis.  Order is only 
significant when using the exclude data merge option (see previous section).  You may specify drive, path and file name, e.g. f:\local\system\net$acct.dat.
Note  Do not directly use the f:\system\net$acct.dat file.  Using this file 
directly will result in accounting errors because the analysis program will 
lock the file and prevent Netware from adding records whilst the analysis 
is being carried out.  This will result in console error messages and lost 
data.  To avoid this problem first copy the accounting file to another 
location.
Having specified the first data file, you will be prompted to enter 
subsequent data files with the message 'Specify ADDITIONAL source file 
(return to continue)'.  When you have no more files to specify (e.g. if 
there is only one file to analyse) simply hit return or enter to continue 
to the next option.
If you enter a file that does not exist, you will be told!  
As you enter valid data file names you will see that your selection is 
displayed towards the top of the screen.  All option selections will be 
listed here for your information.  Each file name is numbered and the date 
of the first data record found is displayed for your convenience.
Also for your convenience you can list data files in the current working 
directory by hitting F2.  You can then highlight and select your desired 
data file.

Output format
-------------
These options specify the combination of output format options that you 
want for this particular analysis.  You can specify no options (in which 
case all the default ones are chosen) or any combination.  The options are 
as follows:

Output format:  U=by user  P=by PC
U = record output summarised by User bindery object ID
P = record output summarised by PC station (ethernet) address - the default

Output format:  P=plain text  C=csv
P = the output file will be in plain text format
C = the output file will be in a format suitable for loading in to a 
spreadsheet or database, i.e. comma separated values - the default
Thus simply hitting enter or return for these options will select all 
defaults:  by PC, CSV format text.

Give output file for results
----------------------------
This option allows you to specify where the analysis results will be stored.  Again you may specify a drive, path and file name.  A default file name is chosen for you.  This is based on the output format options you chose in the previous step as follows:
Option(s)       Default file name
1: P  2:C       PC.CSV
1:U  2:C                USER.CSV
1:P  2:P                PC.TXT
1:U  2:P                USER.TXT

You may enter a different file name (optionally including drive and path) 
by simply typing over the top of the default file name.  Alternatively 
simply hit enter or return to select the default name displayed.
Note  If a file name identical to the one entered already exists, it's 
contents will be lost.  You will receive no warning of this.
A check is made to make sure the output file specified is valid and can 
indeed be written to before this step is complete.  A warning is used if 
either of these problems is identified.

Give start/end date
-------------------
Give start date
This option allows you to specify the date from which records will be 
included for analysis.  The date is inclusive and is specified in the 
format month/day/year (i.e. following the North American style).  The 
day, month and year must be specified in fixed format as two digit numbers.
Thus the date 1st of March 1995 would be entered as 03/01/95.
Some basic checks are included at this stage, e.g. you cannot enter a start 
date later than the day you run the program!  By default the program 
suggests the date of the first data record found in the last data file 
specified (!).
Give end date
This option allows you to specify the date to which records will be included 
for analysis.  The date is inclusive and is specified in the format month/
day/year (i.e. following the North American style).  The day, month and year 
must be specified in fixed format as two digit numbers.
Thus the date the 5th of March 1995 would be entered as 03/05/95.
Some basic checks are included at this stage, e.g. you cannot enter an end 
date that is earlier than the start date.  A default date of today's date 
is provided for your convenience.  If you do not wish to select this, simply 
type in your own date.  If you do wish to select this default date, simply 
hit enter or return.
When you have entered both start and end date the selection will be 
displayed in the top windows together with a calculation of the total 
number of (inclusive) days between the start and end date.  This total 
period is useful when answering the next option.

Working days in period specified
--------------------------------
When calculating the utilisation of network resources such as PCs, or the 
extent staff use the network (when selecting analysis by user), one is 
mostly interested in how much use is made during normal working hours.  
This is particularly true when looking at percentage utilisation.  
Consideration of percentage use for all time (total time) would provide a 
significant low bias.
Thus to calculate percentage utilisation during working hours it is 
necessary not only to know what the working hours are (see next two options) 
but also how many days in the period were spent working these hours.  Hence 
the inclusion of this option.
Thus if you have specified 7 days as the start to end period, this might 
equate to 5 working days.  In some cases working days are not equal, for 
example you may start 08:00 and finish 16:30 Monday to Thursday, but finish 
at 13:00 on Fridays.  In this case you may enter 4.5 working days with a 
normal working day specified as 08:00 to 16:30.
It is not possible to enter more working days than the total days in the 
period!  By default the program suggests a possible number of working days 
based on a calculation of how many weekdays are in the period specified.
Note that if you abort the analysis early, or the program finds data 
analysis finishes with data earlier than the end data you specify, you will 
be prompted to accept an adjusted total/prime period calculation.

START & END of normal working day
---------------------------------
START of normal working day
This option is used to specify the time of day that your normal working day 
starts - the start of 'prime' time for analysis purposes.  The time must be 
entered in '24 clock' format, e.g. 08:00 for 8am or 14:00 for 2pm.  If you 
hit enter or return, the default start time of 08:00 will be selected.
END of normal working day
This option is used to specify the time of day that your normal working day 
ends - the end of 'prime' time for analysis purposes.  The time must be 
entered in '24 clock' format, e.g. 08:00 for 8am or 14:00 for 2pm.  If you 
hit enter or return, the default start time of 16:30 will be selected.

Give (optional) user to send completion message to
--------------------------------------------------
The analysis can take many minutes to complete, depending on the speed of 
your PC, network or local drive and the time period selected.  To ease the 
analysis process you can run the program whilst logged in to a Netware 
server and get the program to tell you via Netware 'SEND' when analysis is 
complete.  To do this you must have SEND in your search path (normal 
location: f:\public) and specify the user ID of the person receiving 
notification at this option.
If you hit return or enter, no completion message will be sent.
You will not see this message if the program cannot detect the presence of 
a NetWare client shell (netx).

Confirm all parameters are correct
----------------------------------
When you have successfully entered all the above options you will be 
prompted to confirm all is OK before the program will commence analysis.  
If you enter 'N' at this option you will be given the option of leaving the 
program at this point.  If you hit 'Y' will be allowed to re-start the 
option selection process from the top.  
If you enter 'Y' or simply hit return or enter the analysis will start.

Data analysis
-------------
When all options have been successfully entered the program commences 
analysis.  You will see a screen looking something like this:

+------------------------------------------------------------------------------+
                     Netware Accounting Analysis System                       
+------------------------------------------------------------------------------+
    Alias files+No PC Alias file, No User Alias file.
(1)   PROCESSED+RSD6.DAT                                         02/16/95 15:31
 Output formats+By PC station addr, CSV text.
    Output file+PC.CSV
	 Period+02/16/95 to 03/07/95 = 20 total day(s)   14 working day(s)
     Prime time+08:00 to 16:30 (Monday to Friday)
Completion tell+WORSNOP_SC
	       
	       
	       
	       
	       
	       
 Started search+10:19:36  Records analysed: 32 (total)  32 (matched)
  Match reached+03/07/95 08:01:06      PCs: 4                          SW v1.20
--------------------------------------------------------------------------------

S:Stop analysis  F1:Help

		    Licence limit:  2500 PCs,  5000 Users.
Various elements of progress information are presented for your information.  These are updated every 10s and include:

Started search
This is the time that the analysis actually commenced.

Records analysed
This is the total (tot) and matching (match) number of records found so far.  The matching records are those that match the specified criteria for start and end date.  These figures are updated every 3 seconds.  If the analysis completed between times this figure will continue to be updated even though analysis has actually completed.

Match reached
When records matching the start and end time period specified have been 
found this line will show the date/time of the latest records found.  This 
data is updated every 3 seconds.  As with the previous data, if the analysis 
completed between times this figure will continue to be updated even though 
analysis has actually completed.

PCs
This is total number PCs found in the data file(s) so far.  If output format 
'by User' is specified then this will change to 'Users' and display the 
total number of Users found in the data file(s) so far. 

Stopping analysis
You can pause then either stop the analysis or continue at any point.  To 
pause the analysis simply hit 'F1' and wait for up to 3 seconds.  The 
program will pause and display a help screen.  Hit any key to continue 
analysis.
To end analysis early at any point simply hit 'S'.  The program will still 
provide the output file specified, but this will only include data matching 
the start/end time criteria that have been read from the source data files 
up to this point - i.e. it will almost certainly be incomplete data.
This option is useful if you are trying out analyses and wish to see a 
partial result as a guide.

Adjustment of period data
When the data analysis is complete, or when you hit Stop to interrupt the 
analysis, the program will take the date from the last data analysed and 
decide if the total/prime periods require adjustment.  If they do, you will 
be presented with an option to adjust the periods.  If you answer Yes (or 
hit return, since this is the default option) the program will replace the 
end date and total/prime periods with those calculated.
This option is useful if you are not sure when the data in your accounting 
file(s) finish, or when you want to exit the program early but still use 
the data analysed so far.
As with the initial prime period calculation, the number of days in the 
prime period is calculated as being the number of weekdays in the period 
so far analysed.

Analysis methodology
--------------------
Period calculations
The key to the Netware Accounting System is the method by which period 
calculations are made.

Prime time period calculations
An additional feature of this version of the System is its ability to 
provide analysis based on normal working hours (or prime time).

Full data
Both full and summary analysis is presented.

Centile calculations
The summary analysis is presented in centile format.

Period Calculations
-------------------
The program takes individual login and logout records from the raw Netware 
accounting file(s) and attempts to match them.  Where a match is found a 
period is calculated.  These individual periods are then summed together 
with by PC station address or User bindery ID.

The following assumptions are made:
A valid period consists of:
 A login followed by a logout record either for the same PC or User;
 A login followed by a further login record for either PC or User.
If a logout record is followed by a second logout record, this data is 
ignored.
If a logout record is followed by a login record, and no previous login 
record was found, the start of the current period is reset to the login 
record time.
Note that where a user logs in before the end of the analysis period and 
logs out after the end of the period, it is possible to record a usage of 
0 minutes.  This is because the login is recorded but no matching logout 
was found.
Accounting data is recorded to the second.  This program rounds the data 
to the nearest minute for analysis purposes.

Prime time period calculations
------------------------------
The prime time - or period of normal working hours - is calculated by 
comparing the start and end date/time of the current login session with 
the day of the week and the start/end time of the prime time specified at 
the beginning of the program.
Login time spanning Saturday or Sunday are ignored (it's assumed normal work 
only takes place Monday to Friday!).  The program then calculates the actual 
portion of the login session that encompasses the normal working hours 
period.  For example, consider the following session:
Login at:       Friday 10:00
Logout at:      Monday 09:00
If normal working hours are 08:00 to 16:30, the prime time period for this 
session would be:
From Friday: 10:00 to 16:30     = 6.5 hours
Saturday and Sunday are ignored = 0 hours
From Monday: 08:00 to 09:00     = 1 hour
I.e. Total prime time period for this session = 7.5 hours.

Full data
---------
The full data presented consists of:
 PC station address or User bindery object ID
 Minutes of usage either in total or during working (prime) time
 Percentage usage either in total or during working (prime) time.

Centile calculation
-------------------
The centiles are calculated as follows:
0-10% centile:  Number of PCs (or Users) with total or prime time utilisation percentage in the range:
		0% <= utilisation < 10%
0-20% centile:  Number of PCs (or Users) with total or prime time utilisation percentage in the range:
		10% <= utilisation < 20%    etc.
Thus you will see that the spread of users to higher centiles is greater for 
prime time utilisation than total time.  This is because typically users use 
a network more during normal working hours (or at least they do where I 
work!).
Note that PCs or Users with 0 minutes of total time will be excluded from 
this calculation, so as not to adversely bias the 0-10% centile.  These 
will typically be users who have only a login or logout record during the 
accounting period, with no corresponding logout or login record having been 
found.  Thus the existence of the PC/user has been recognised but no 
accounts actually collected yet.

Session period centiles
A centile calculation is also carried out to provide some breakdown of how 
long sessions typically last for your network.  This provides the number of 
sessions (i.e. login/logout periods) by how long those sessions lasted.  
Thus your can see if most sessions are less than 30 minutes in length (for 
example) or whether you have real power users who login for hours at a time 
(we have both, giving a 'bath tub' sort of graph).

Boot log
--------
When you have run the analysis you will find a file called 'boot.log' in the 
directory that you ran the program from.  This is an by-product of the main 
analysis that you may find of interest, and for this reason it has been left 
in the program.  The log will record all accounting records of the type 
'server boot' or 'server down':  i.e. a record of server outage for the 
analysis period specified.

Problems
--------
If you have any problems this may be because:
1. You don't have supervisor privileges and hence cannot read the file 
f:\system\net$acct.dat.  Ask someone who does to copy this file to an area 
you can read.
2. You are using the Netware(R) billing option - the software has not been 
tested with this option active.
3. You have unrecognised accounting information in your accounting file.  
Some third-party programs can write to the accounting file.  If you suspect 
this is the case please forward a copy of the sample accounting file 
together with details of your server environment (Netware(R) version, 
billing and accounting options etc.).
4. The program takes too long to analyse.  Try a faster PC.  In the longer 
term the next version will include optimised sort routines that speed up 
analyses by up to 300% - but only on request from registered users.

Command line options
--------------------
The only valid command line option currently available is /P.  This allows 
trace data for a specified PC to be logged to the file TRACE.LOG.
To use this facility start up the program specifying the PC station address 
of the PC to be traced, for example:
PCAUDIT /P=00001B34671F
In this case all sessions (login/logout periods) for the PC with this 
station address will be recorded in the file TRACE.LOG.  If you have 
correctly entered the command line option you will see the opening input 
screen as normal, but with the addition of a line highlighting that a trace 
is in progress and specifying the address being traced.
Note that 'by PC' must be specified as the output format option for this 
trace to work.  'By user' will give invalid results.

The following data will be recorded:
1. Session start date and time, date in the format mm/dd/yy
2. Session end date and time, date in the format mm/dd/yy
3. Calculated total number of minutes in the period
4. Calculated total number of minutes during normal working hours in the 
period.
**************************************************
