perm filename SOB.LES[UP,DOC]2 blob sn#648671 filedate 1982-03-19 generic text, type C, neo UTF8
COMMENT āŠ—   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	Jan 1978	SOB (Son of Bureaucracy)	LES
C00009 ENDMK
CāŠ—;
Jan 1978	SOB (Son of Bureaucracy)	LES

SOB first asks for the date range, then for the output form.
If you answer each request by typing <carriage return>, it will
automatically show a summary of computer utilization by project from the
beginning of the current year through yesterday.

Accounting records for this program begin in May 1977, but reasonably
accurate data exists only from November '77 on.  When SOB asks for dates,
if you say "NOV", to get data for the most recent November.  If the
current month is November, this gets data from November 1 through
yesterday.

Similarly, "Jul:Oct" covers a four month period this year (if past
October) or last year (if earlier than July).  If you ask for those dates
during July, August, or September, however, you will get an error message
because the information isn't available; it doesn't hurt to ask.

If you say ":May" you get data from 1 January through May.  If you say 77
or 1977, you get data for that entire year.  If You mention the current
year, or if you just hit <carriage return>, you get year-to-date data.

			OUTPUT OPTIONS
To get a summary of computer usage by project, just type <carriage
return>.  This lists projects in order of decreasing "doubloons" (a
linear function of computer cost components on a somewhat arbitrary
scale).

In calculating this cost, "B time" charges (weekday mornings and
evenings, weekend afternoons) are reduced by a factor of 2 and "C time"
is reduced a factor of 3.  The doubloon cost is also shown as a percentage
of overall computer utilization (excluding system services such as
spooling, accounting, and news service).

Next, the number of logins is given and the percentage of A time, B time,
and C time, based on the doubloon calculation with no discounting.
The last three items are the average core size, CPU utilization (CPU time/
job time), and load average while computing.

There are three possible commands: GROUP, PN, and MOST.  All commands and
switches may be abbreviated so long as they are unambiguous (one letter
will do).  Cases are ignored.  The default command, described above, is
equivalent to GROUP without an argument.  If an argument is given, then it
is matched against the leading characters of the various group names to
select one (e.g. the command "g hand" will retrieve data on the Hand-eye
project).

The PN command with a programmer ID as an argument will get data on that
individual alone.  Without an agument, it exhaustively lists the accounting
data in order by PN, omitting null entries.

The MOST command without an argument lists the 20 heaviest users.  An
integer argument causes that many of the heaviest users to be listed.

			SWITCHES
The /DETAIL switch has effect only on the GROUP command and causes data
for individuals in each group to be listed in order of decreasing cost.
The /FILE switch will subsequently ask for a name and direct output there.
/SPOOL and /XSPOOL do what you would expect.

The /DISK switch causes disk space allocation and doubloons/50 tracks to
be printed.  The /SUMMARY switch prints only name, doubloons, disk space
allocation, and doubloons/disk allocation normalized by the system
average.  The /SUMMARY switch implies the /DISK switch.

The /SORTDISK switch sorts the data by the doubloon/space ratio rather
than by doubloons.  The /SORTDISK switch implies the /DISK switch.  The
/PPN switch causes sorting to by by PPN.  The /DOUBSORT switch sorts
solely by doubloons.  All sort switches are sticky, that is they affect
all subsequent requests in the same run until another sort switch is
given.  The default is sorting by doubloons.

The /RAW switch is mostly for debugging and causes raw accounting data to
be printed together with a (rough) explanatory heading.

The /REMOTE switch selects only people who have remote login time and shows
how much.