Resource Evaluation

Network Working Group                                         J. Pickens
Request for Comments: 519            Computer Systems Laboratory -- UCSB
NIC: 16818                                                     June 1973


                          Resource Evaluation

Abstract

   In the spirit of RFC # 369, Evaluation of ARPANET resources, a new
   test group was organized at UCSB to take a detailed look at specific
   network resources and develop initial site dependent and function
   dependent MINIMAN's (Concise User Manuals).  As the group was again
   composed of novices, initial effort revolved about basic procedural
   indoctrination.  In the period between January and March 1973 a
   number of resources were investigated with varying degrees of
   success, as to availability, proper usage, sample problem solutions,
   and access to help and documentation.  Included in this paper are a
   summary of the projects undertaken, initial suggestions at MINIMAN
   composition, and suggestions for future test groups.  As these groups
   are attempting to perform a useful function for the ARPANET
   community, comments and suggestions are requested.  Copies of the
   reports described herein are available on request from the Computer
   Systems Laboratory at UCSB.

Resources Investigated by the Group

   I.  APL

      APL was investigated primarily at MULTICS.  UCSD was also
      scheduled for evaluation but not carried out.  APL at MULTICS was
      used to solve a few trivial problems.  Most effort revolved about
      the difficult task of obtaining any available documentation.  The
      octal codes for APL characters were obtained and mapped into the
      OLS keyboard.  A side goal of the project, the comparison of APL
      with OLS, was begun but progressed very little.

   II.  Basic

      Basic was investigated at a number of TENEX sites.  Differences
      between sites were pointed out and necessary file manipulation
      commands were documented.  An integration problem was written at
      one site, sent via FTP to another site, and then run again to show
      comparative execution times and compatability.  Non-PDP/10 sites
      were investigated but no report was submitted.

   III.  TSO

      IBM's Time Sharing option was exercised at UCLA-CCN.  Interesting
      results were obtained regarding cost and execution time.
      Available commands were documented and a PL/1 program was written
      and executed.

   IV.  MIT-MATHLAB

      This, the most successful of the projects, involved documentation
      of help, file manipulation, and MACSYMA access and an original
      research project in resource sharing.  A recursive problem in
      pattern recognition and a triple integration were solved to
      demonstrate MACSYMA generated expressions into user programs on
      the OLS.  More information on this project is forthcoming.

   V.  Local User Guide

      A first pass network users manual was completed for UCSB users.
      In it are described console access and settings, character
      mappings, current servers, users and TIPs, and error conditions.
      Following minor revisions this guide will be distributed to local
      users.

   VI.  Local IMLAC Access to Network

      Access to network graphics programs was attempted with a local
      IMLAC.  Due to the non-uniformity of network IMLACs very little
      success was obtained.  However, a program to access SRI-ARC's NLS
      was compiled and loaded from NIC and attempts were made to iron
      out the bugs.  In addition a project was begun to maintain an
      IMLAC library and compiler locally for network usage.  As in the
      other projects, basic operating procedures were documented.

   VII.  Harvard Graphics

      Several attempts were made to learn of availability of graphics
      access to organic molecule synthesis programs but no response
      could be generated.  This project was eventually abandoned.

MINIMAN Composition

   As mentioned in a previous report, concise manuals are needed for
   network resources so that uniniated users may gain basic familiarity
   with foreign systems.  In addition, manuals which describe specific
   network wide functions, such as Fortran compilors, are needed if
   resource sharing is to become a real trait of the ARPANET.  For the
   resources evaluated, each group member submitted two reports
   analagous to the two types of MINIMANS needed in the network.  The
   headings and format of the reports will be included here to stimulate
   future discussion on MINIMAN composition.

   REPORT # 1: Online Help for [a specific host computer]

         I.  Connection, Login, and Optimal TELNET Settings

        II.  Help Files

       III.  Job Status

        IV.  Time of Day

         V.  Time/Money  Used/Left

        VI.  Interpersonal Communications

             A. Console Linking

             B. Location of Users

             C. Mail Facilities

             D. Access to Operator and/or Consultants

       VII.  Warnings or Unfriendly User Behavior

      VIII.  Useful References and Documentation

   REPORT #2: How to Use [a specific resource]

         I.  Table of Contents

        II.  Access and Usage (or How to Start and Stop)

       III.  Editing Commands and File Structures

        IV.  Documentation, Location, and Cost

         V.  Sample Solutions and Significant Problems

        VI.  Appendices

             A. Special Characters and Terminal Settings

             B. Similarities and Differences from Site to Site

Future Test Groups

   A number of projects are envisioned for future resource evaluators
   and include:

   1. Complete evaluation of APL at MULTICS and UCSD with comparison to
      the On-Line System (OLS).

   2. Investigate BASIC in depth, network wide.

   3. Evaluate other symbolic manipulation programs such as REDUCE.

   4. Summarize all games available in the network.

   5. Find and evaluate specific application programs such as ZOG or the
      weather data base at CCA.

   The projects undertaken will be determined in part by local
   interests.  But a serious effort is being made for reports to
   accompany each evaluation.

Conclusion

   Good results have been obtained from the two test groups thus far.
   Although composed of novices, as far as network familiarity is
   concerned, the groups have been able to produce data and reports
   which benefit the network community.  The reports run the gambit from
   poor to excellent, but even the poorer ones have generated results by
   motivating more knowledgeable system personnel to find time to write
   the report in the "right" way.  All data and reports compiled by
   these groups are available to interested network users.  In addition,
   any information or documentation or manuals which might fit into the
   framework of the MINIMAN is requested from the network community.  As
   this information begins to be collected, the network may truly start
   to become a resource sharing network.


         [ This RFC was put into machine readable form for entry ]
         [ into the online RFC archives by Nasser M. Akhtar 2/98 ]