Bureaucracy:	Computer time usage   --   L. Earnest, December 1973

To find out how much computer time has been  used by any given group,
say
.R BUREAUCRACY		(actually, just "R BUREAU" will do).
The program will ask for dates and for PPN lists.  If you give a null
response  (CR)  to both  requests,   the  program  will automatically
generate a  summary  of  computer  utilization by  project  from  the
beginning of the current year through yesterday.

Data covered include number of PPNs, number of logins, job hours, CPU
minutes, core  K minutes, and "Doubloons"  (an approximate measure of
overall cost on  an arbitrary scale).   The entries  are in order  of
decreasing doubloons, except that "All others" is always last.  After
typing all  this junk out, it asks for a file name, in case you would
like to save a copy.

			DATES

Alternatively, you can  "roll your own".   The raw data is  stored by
PPN for each month over the past year and for each year back to 1971.
When it asks  for dates,  you can say something like "Jul:Oct", which
may or may not work.  If  the current date is December, this gives
you data  for the  current year.   If the current  date is  May, last
year's data is used.  If you say that in July, August,  or September,
it will simply complain.  It doesn't hurt to ask.

If you say just "May" then you get data for the last instance of that
month only.   If you say  ":May" you get data  from 1 January through
May.  If you say 72 or  1972, 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.

			PPNs

When the program asks for PPNs, you can say something like:

Bureaucrats:JMC,LES,JAF; Bad guys:GUE,[NET,SYS]; Classes: [206],[225]

which will gather data on three groups, separated by ";". Labels (one
or more  words ending in ":") may appear  anywhere in the group list.
If you omit labels,  the groups will be  named "Group 1", "Group  2",
etc.

More generally, the following elements may appear in PPN lists:

group label:pn,[pj],[pj,pn],[*,pn],[pj,*],@file,@@,/SUMMARY

Nothing need be capitalized.  [pj] is of course equivalent  to [pj,*]
and pn means [*,pn].  If a given PPN qualifies for membership in more
than one group, then [pj,pn]  takes precedence over [pj,*], which  in
turn outranks [*,pn].  If you  mention exactly the same thing in more
than  one group the  program will grumble, but  will ignore conflicts
and continue.

Now about the "@file" gizmo: This lets  you compose a horrendous list
as a  text file and gobble it into the middle  of the line.  The file
name may contain a PPN.  Any TV/E directory or SOS line  numbers will
be ignored.   While the on-line  response is restricted to  one line,
the  text file may  contain many  lines.  It  will be read  as if the
carriage returns don't exist. You can go indirect any number of times
and  any  number  of  levels,  though  string  space  may  freak  out
eventually.

The "@@"  form says "go read  the telephone list".   This will assign
everyone to  groups on the basis of project affiliations given in our
telephone directory.   Individuals who are  in more than one  project
are  arbitrarily pushed into  the first  one on  their list.   People
without a listed project are dumped into "Miscellaneous".  Of course,
this still leaves  quite a few  "outlaws" who are  not on any  of our
authorized  user lists.   You can precede  the "@@" with  one or more
group  lists of  your  own,  which  will  take  precedence  over  the
directory groupings.

If you stick  a "/SUMMARY" or  even just "/s" at  the end of  the PPN
string,   then  the output  will  list only  the total  doubloons and
percentages of total for each group.  If you leave this off, you  get
listings of actual time used for each element on the list.  Thus, to
get a detailed listing of all major projects, respond to "PPNs" with
"@@".

			Looping

After the program  finishes and (upon  request) writes out a  copy of
the the results,  it loops back for another run.  On subsequent runs,
the null reply to questions means  "just like last time".  Thus,  you
can easily  ask for data from  a different time period  with the same
cast of characters, or for the same time with another group.