RFC 138

Network Working Group                                       Bob Anderson
Request for Comments: 138                                           Rand
NIC 6715                                                       Vint Cerf
                                                            Eric Harslem
                                                            John Heafner
                                                              Jim Madden
                                                          U. of Illinois
                                                            Bob Metcalfe
                                                           Arie Shoshani
                                                               Jim White
                                                              David Wood
                                                           28 April 1971


     I.  INTRODUCTION .................................    2

         Purpose of this RFC ..........................    2
         Motivation ...................................    2


         Elements of Data Reconfiguration Service .....    3
         Conceptual Network Connections ...............    3
         Connection Protocols and Message Formats .....    4
         Example Connection Configurations ............    6

   III.  THE FORM MACHINE .............................    7

         Input/Output Stream and Forms ................    7
         Form Machine BNF Syntax ......................    7
         Alternate Specification of Form Machine Syntax    8
         Forms ........................................    9
         Rules ........................................   10
         Terms ........................................   10

           Term Format 1 ..............................   11
           Term Format 2 ..............................   11
           Term Format 3 ..............................   13
           Term Format 4 ..............................   13
           Application of a Term ......................   14

Anderson, et al.                                                [Page 1]

RFC 138               Data Reconfiguration Service          April 1971

           Restrictions and Interpretations of
             Term Functions ...........................   14
         Term and Rule Sequencing .....................   16

    IV.  EXAMPLES .....................................   16

         Remarks ......................................   16
         Field Insertion ..............................   17
         Deletion .....................................   17
         Variable Length Records ......................   17
         String Length Computation ....................   18
         Transposition ................................   18
         Character Packing and Unpacking ..............   18


    VI.  IMPLEMENTATION PLANS .........................   20

   Appendix A .........................................   21

         Note 1 to the DRS Working Group ..............   21
         Note 2 to the DRS Working Group ..............   22



   The purpose of this RFC is to describe, in part, a proposed Network
   experiment and to solicit comments on any aspect of the experiment.
   The experiment involves a software mechanism to reformat Network data
   streams.  The mechanism can be adapted to numerous Network
   application programs.  We hope that the results of the experiment
   will lead to a further standard service that embodies the principles
   described in this RFC.   We would like comments on the
   appropriateness of this work as a Network experiment and also
   comments on particular Network data reformatting needs that could not
   easily be accomplished using these techniques.


   Application programs require specific data I/O formats yet the
   formats are different from program to program.  We take the position
   that the Network should adapt to the individual program requirements
   rather than changing each program to comply with a standard.  This
   position doesn't preclude the use of standards that describe the
   formats of regular message contents; it is merely an interpretation
   of a standard as being a desirable mode of operation but not a
   necessary one.

Anderson, et al.                                                [Page 2]

RFC 138               Data Reconfiguration Service          April 1971

   In addition to differing program requirements, a format mismatch
   problem occurs where users wish to employ many different kinds of
   consoles to attach to a single service program.  It is desirable to
   have the Network adapt to individual console configurations rather
   than requiring unique software packages for each console

   One approach to providing adaptation is for those sites with
   substantial computing power to offer a data reconfiguration service;
   a proposed example of such a service is described here.

   The envisioned modus operandi of the service is that an applications
   programmer defines _forms_ that describe data reconfigurations.  The
   service stores the forms by name.  At a later time, a user (perhaps a
   non-programmer) employs the service to accomplish a particular
   transformation of a Network data stream, simply by calling the form
   by name.

   We have attempted to provide a notation tailored to some specifically
   needed instances of data reformatting while keeping the notation and
   its underlying implementation within some utility range that is
   bounded on the lower end by a notation expressive enough to make the
   experimental service useful, and that is bounded on the upper end by
   a notation short of a general purpose programming language.



   An implementation of the Data Reconfiguration Service (DRS) includes
   modules for connection protocols, a handler of some requests that can
   be made of the service, a compiler and/or interpreter (called the
   Form Machine) to act on those requests, and a file storage module for
   saving and retrieving definitions of data reconfigurations (forms).
   This section highlights connection protocols and requests.  The next
   section covers the Form Machine language in some detail.  File
   storage is not described in this document because it is transparent
   to the use of the service and its implementation is different at each
   DRS host.


   There are three conceptual Network connections to the DRS, see Fig.

         1)  The control connection (CC) is between an originating user
             and the DRS.  A form specifying data reconfiguration is

Anderson, et al.                                                [Page 3]

RFC 138               Data Reconfiguration Service          April 1971

             defined over this connection and is applied to data passing
             over the two connections described below.
         2)  The user connection (UC) is between a user process and the
         3)  The server connection (SC) is between the DRS and the
             serving process.

   Since the goal is to adapt the Network to user and server processes,
   a minimum of requirements are imposed on the UC and SC.

      +-------------+   CC   +-----------+   SC   +-----------+
      | ORIGINATING +--------+    DRS    +--------+ SERVER    |
      |    USER     |   ^    |           |    ^   | PROCESS   |
      +-------------+   |    +------+----+    |   +-----------+
                        |          /          |
                     Telnet       / <------ Simplex or Duplex
                    Protocol   UC/            Connections
                   Connection   /
                        | USER      |
                        | PROCESS   |

                Figure 1.  DRS Network Connections


   Over a control connection the dialog is directly between an
   originating user and the DRS.  Here the user is defining forms or
   assigning forms to connections for reformatting.

   The user connects to the DRS via the initial connection protocol
   (ICP) specified in NWG/RFC #80, version 1.  Rather than going through
   a logger, the user calls on a particular socket on which the DRS
   always listens.  DRS switches the user to another socket pair.

   Messages sent over a control connection are of the types and formats
   to be specified for TELNET.  Thus, a user at a terminal should be
   able to connect to a DRS via his local TELNET, for example, as shown
   in Fig.  2.

Anderson, et al.                                                [Page 4]

RFC 138               Data Reconfiguration Service          April 1971

                         +--------+  CC   |              |
                 +-------+ TELNET +-------+     DRS      |
                 |       +--------+       |              |
                 |                        +--------------+
      |      USER          |

           Figure 2.  A TELNET Connection to DRS

   When a user connects to DRS he supplies a six-character user ID (UID)
   as a qualifier to guarantee the uniqueness of his form names.  He
   will have (at least) the following commands:

         1.  DEFFORM (name)
         2.  ENDFORM (name)

             These two commands define a form, the text of which is
             chronologically entered between them.  The (name) is six
             characters.  The form is stored in the DRS local file

         3.  PURGE (name)

             The named form, as qualified by the current UID, is purged
             from the DRS file system.

         4.  LISTNAMES (UID)

             The unqualified names of all forms assigned to UID are

         5.  LISTFORM (name)

             The source text of a named form is returned.

         6.  DUPLEXCONNECT (user site, user send, user receive,
                        user method, server site, server
                        send, server receive, server method,
                        user-to-server form, server-to-user form)

         7.  SIMPLEXCONNECT (send site, send socket, send
                          method, receive site, receive
                          socket, receive method, form)

Anderson, et al.                                                [Page 5]

RFC 138               Data Reconfiguration Service          April 1971

   Either one, both, or neither of the two parties specified in 6 or 7
   may be at the same host as the party issuing the request.  Sites and
   sockets specify user and server for the connection.  Method indicates
   the way in which the connection is established.  Three options are
        1)  Site/socket already connected to DRS as a dummy
            control connection.  (A dummy control connection
            should not also be the real control connection.)
        2)  Connect via standard ICP. (Only for command no. 6.)
        3)  Connect directly via STR, RTS.


   There are basically two modes of DRS operation: 1) the user wishes to
   establish a DRS UC/SC connection(s) between two programs and 2) the
   user wants to establish the same connection(s) where he (his
   terminal) is at the end of the UC or the SC.  The latter case is
   appropriate when the user wishes to interact from his terminal with
   the serving process (e.g., a logger).

   In the first case (Fig. 1, where the originating user is either a
   terminal or a program) the user issues the appropriate CONNECT
   command.  The UC/SC can be simplex or duplex.

   The second case has two possible configurations, shown in Figs. 3 and

               +--------+  CC  +--------+      +------+
               |        +------+        |  SC  |      |
     +------+ /| TELNET |  UC  |  DRS   +------+  SP  |
     |      |/ |        +------+        |      |      |
     | USER | /+--------+      +--------+      +------+
     |      |/

            Figure 3.  Use of Dummy Control Connection

     +------+ /|  USER  |  CC  +--------+      +------+
     |      |/ |  SIDE  +------+        |  SC  |      |
     | USER |  +--------+  UC  |  DRS   +------+  SP  |
     |      |\ | SERVING+------+        |      |      |
     +------+ \|  SIDE  |      +--------+      +------+

            Figure 4.  Use of Server TELNET

Anderson, et al.                                                [Page 6]

RFC 138               Data Reconfiguration Service          April 1971

   In Fig. 3 the user instructs his TELNET to make two duplex
   connections to DRS.  One is used for control information (the CC) and
   the other is a dummy.  When he issues the CONNECT he references the
   dummy duplex connection (UC) using the "already connected" option.

   In Fig. 4 the user has his TELNET (user side) call the DRS.  When he
   issues the CONNECT the DRS calls the TELNET (server side) which
   accepts the call on behalf of the console.  This distinction is known
   only to the user since to the DRS the configuration in Fig. 4 appears
   identical to that in Fig. 1.  Two points should be noted:

        1)  TELNET protocol is needed only to define forms and direct
            connections.  It is not required for the using and serving
        2)  The using and serving processes need only a minimum of
            modification for Network use, i.e., an NCP interface.



   This section describes the syntax and semantics of forms that specify
   the data reconfigurations.  The Form Machine gets an input stream,
   reformats the input stream according to a form describing the
   reconfiguration, and emits the reformatted data as an output stream.

   In reading this section it will be helpful to envision the
   application of a form to the data stream as depicted in Fig. 5.  An
   input stream pointer identifies the position of data (in the input
   stream) that is being analyzed at any given time by a part of the
   form.  Likewise, an output stream pointer locates data being emitted
   in the output stream.

       /\/\                                                  /\/\
  ^    |  |                     FORM                         |  |   ^
  |    |  |                -----------------                 |  |   |
  |    |  |            +-  -----------------  -+             |  |   |
  |    |  |            |   CURRENT PART OF     |             |  |   |
INPUT  |  |<= CURRENT <    -----------------    > CURRENT => |  | OUTPUT
       |  |            +-  -----------------  -+             |  |
       |  |                -----------------                 |  |
       |  |                -----------------                 |  |
       |  |                -----------------                 |  |
       \/\/                                                  \/\/
              Figure 5.  Application of Form to Data Streams

Anderson, et al.                                                [Page 7]

RFC 138               Data Reconfiguration Service          April 1971


   form           ::=  rule | rule form

   rule           ;;=  label  inputstream  outputstream ;

   label          ::=  INTEGER | <null>

   inputstream    ::=  terms | <null>

   terms          ::=  term | terms , term

   outputstream   ::=  : terms | <null>

   term           ::=  identifier | identifier  descriptor |
                       descriptor | comparator

   identifier     ::=  an alpha character followed by 0 to 3

   descriptor     ::=  (replicationexpression , datatype ,
                       valueexpression , lengthexpression  control)

   comparator     ::=  (value  connective  value  control)  |
                       (identifier .<=>. control)

   replicationexpression  ::=  arithmeticexpression | <null>

   datatype       ::=  B | O | X | E | A

   valueexpression  ::=  value | <null>

   lengthexpression  ::=  # | arithmeticexpression | <null>

   connective     ::=  .LE. | .LT. | .GE. | .GT. | .EQ. | .NE.

   value          ::=  literal | arithmeticexpression

   arithmeticexpression  ::=  primary | primary operator

   primary        ::=  identifier | L(identifier) | V(identifier) |

   operator       ::=  + | - | * | /

   literal        ::=  literaltype "string"

Anderson, et al.                                                [Page 8]

RFC 138               Data Reconfiguration Service          April 1971

   literaltype    ::=  B | O | X | E | A

   string         ::=  from 0 to 256 characters

   control        ::=  :  options | <null>

   options        ::=  S(where) | F(where) | U(where) |
                       S(where) , F(where) |
                       F(where) , S(where)

   where          ::=  arithmeticexpression | R(arithmeticexpression)


form                    ::=  {rule}
                                      1         1          1
rule                    ::=  {INTEGER}   {terms}   {:terms} ;
                                      0         0          0
terms                   ::=  term {,term}
term                    ::=  identifier | {identifier}   descriptor
                             | comparator
descriptor              ::=  ({arithmeticexpression}  , datatype ,
                                    1                     1           1
                             {value} ,  {lengthexpression}  {:options}
                                    0                     0           0
comparator              ::=  (value  connective  value {:options} ) |
                             (identifier .<=. value {:options} )
connective              ::=  .LE. | .LT. | .GE. | .GT. | .EQ. | .NE.

lengthexpression        ::=  # | arithmeticexpression

datatype                ::=  B | O | X | E | A

value                   ::=  literal | arithmeticexpression

Anderson, et al.                                                [Page 9]

RFC 138               Data Reconfiguration Service          April 1971

arithmeticexpression    ::=  primary  {operator  primary}
operator                ::= + | - | * | /

primary                 ::=  identifier | L(identifier) |
                             V(identifier) | INTEGER
literal                 ::=  literaltype  "{CHARACTER}   "
literaltype             ::=  B | O | X | A | E
options                 ::=  S(where) {,F(where)}  |
                             F(where) {,S(where)}  | U(where)

where                   ::=  arithmeticexpression |
identifier              ::=  ALPHABETIC  {ALPHAMERIC}


   A form is an ordered set of rules.

         form ::=  rule | rule form

   The current rule is applied to the current position of the input
   stream.  If the (input stream part of a) rule fails to correctly
   describe the contents of the current input then another rule is made
   current and applied to the current position of the input stream.  The
   next rule to be made current is either explicitly specified by the
   current term in the current rule or it is the next sequential rule by
   default.  Flow of control is more fully described under TERM AND RULE

   If the (input stream part of a) rule succeeds in correctly describing
   the current input stream, then some data may be emitted at the
   current position in the output stream according to the rule.  The
   input and output stream pointers are advanced over the described and
   emitted data, respectively, and the next rule is applied to the now
   current position of the input stream.

   Application of the form is terminated when an explicit return
   (R(arithmeticexpression)) is encountered in a rule.  The user and

Anderson, et al.                                               [Page 10]

RFC 138               Data Reconfiguration Service          April 1971

   server connections are closed and the return code
   (arithmeticexpression) is sent to the originating user.


   A rule is a replacement, comparison, and/or an assignment operation
   of the form shown below.

         rule ::= label  inputstream  outputstream ;

   A label is the name of a rule and it exists so that the rule may be
   referenced elsewhere in the form for explicit rule transfer of
   control.  Labels are of the form below.

         label ::=  INTEGER | <null>

   The optional integer labels are in the range 0 >= INTEGER >= 9999.
   The rules need not be labeled in ascending numerical order.


   The inputstream (describing the input stream to be matched) and the
   outputstream (describing data to be emitted in the output stream)
   consist of zero or more terms and are of the form shown below.

         inputstream   ::=  terms | <null>
         outputstream  ::=  :terms | <null>
         terms         ::=  term | terms , term

   Terms are of one of four formats as indicated below.

         term ::=  identifier | identifier  descriptor |
                   descriptor | comparator

Term Format 1

   The first term format is shown below.


   The identifier is a symbolic reference to a previously identified
   term (term format 2) in the form.  It takes on the same attributes
   (value, length, type) as the term by that name.  Term format 1 is
   normally used to emit data in the output stream.

   Identifiers are formed by an alpha character followed by 0 to 3
   alphameric characters.

Anderson, et al.                                               [Page 11]

RFC 138               Data Reconfiguration Service          April 1971

Term Format 2

   The second term format is shown below.

         identifier descriptor

   Term format 2 is generally used as an input stream term but can be
   used as an output stream term.

   A descriptor is defined as shown below.

         descriptor ::= (replicationexpression, datatype,
                        valueexpression, lengthexpression

   The identifier is the symbolic name of the term in the usual
   programming language sense.  It takes on the type, length, and value
   attributes of the term and it may be referenced elsewhere in the

   The replication expression is defined below.

         replicationexpression ::= arithmeticexpression | <null>
         arithmeticexpression ::= primary | primary operator
         operator ::= + | - | * | /
         primary ::= identifier | L(identifier) | V(identifier) |

   The replication expression is a repeat function applied to the
   combined data type and value expression.  It expresses the number of
   times that the value (of the data type/value expression) is to be
   repeated within the field length denoted by the data type/length

   A null replication expression has the value of one.  Arithmetic
   expressions are evaluated from left-to-right with no precedence.  (It
   is anticipated that this interpretation might be changed, if

   The L(identifier) is a length operator that generates a 32-bit binary
   integer corresponding to the length of the term named.  The
   V(identifier) is a value operator that generates a 32-bit binary
   integer corresponding to the value of the term named.  (See
   Restrictions and Interpretations of Term Functions.)  The value
   operator is intended to convert character strings to their numerical

Anderson, et al.                                               [Page 12]

RFC 138               Data Reconfiguration Service          April 1971

   The data type is defined below.

             datatype ::= B | O | X | E | A

   The data type describes the kind of data that the term represents.
   (It is expected that additional data types, such as floating point
   and user-defined types, will be added as needed.)

        Data Type         Meaning              Unit Length

            B             Bit string              1 bit
            O             Bit string              3 bits
            X             Bit string              4 bits
            E             EBCDIC character        8 bits
            A             Network ASCII character 8 bits

        The value expression is defined below.

                 valueexpression ::= value | <null>
                 value ::= literal | arithmeticexpression
                 literal ::= literaltype "string"
                 literaltype ::= B | O | X | E | A

   The value expression is the unit value of a term expressed in the
   format indicated by the data type.  It is repeated according to the
   replication expression, in a field whose length is constrained by the
   length expression.

   A null value expression in the input stream defaults to the data
   present in the input stream.  The data must comply with the datatype
   attribute, however.

   A null value expression generates padding according to Restrictions
   and Interpretations of Term Functions.

   The length expression is defined below.

         lengthexpression ::= # | arithmeticexpression | <null>

   The length expression states the length of the field containing the
   value expression as expanded by the replication expression.  If the
   value of the length expression is less then the length implied by the
   expanded value expression, then the expanded value expression is
   truncated, see Restrictions and Interpretations of Term Functions.

   The terminal symbol # means an arbitrary length, explicitly
   terminated by the value of the next term.  The # is legal only in the
   input stream and not in the output stream.

Anderson, et al.                                               [Page 13]

RFC 138               Data Reconfiguration Service          April 1971

   If the length expression is less than or equal to zero, the term
   succeeds but the appropriate stream pointer is not advanced.
   Positive lengths cause the appropriate stream pointer to be advanced
   if the term otherwise succeeds.

   Control is defined under TERM AND RULE SEQUENCING.

Term Format 3

   Term format 3 is shown below.


   It is identical to term format 2 with the omission of the identifier.
   Term format 3 is generally used in the output stream.  It is used in
   the input stream where input data is to be passed over but not
   retained for emission or later reference.

Term Format 4

   The fourth term format is shown below.

         comparator    ::= (value connective value control) |
                           (identifier .<=. value control)
         value         ::= literal | arithmeticexpression
         literal       ::= literaltype "string"
         literaltype   ::= B | O | X | E | A
         string        ::= from 0 to 256 characters
         connective    ::= .LE. | .LT. | .GE. | .GT. | .EQ. | .NE.

   The fourth term format is used for assignment and comparison.

   The assignment operator .<=. assigns the value to the identifier.
   The connectives have their usual meaning.  Values to be compared must
   have the same type and length attributes or an error condition arises
   and the form fails.

The Application of a Term

   The elements of a term are applied by the following sequence of

         1.  The data type and value expression together specify a unit
             value, call it x.

         2.  The replication expression specifies the number of times x
             is to be repeated (or copied) in concatenated fashion.  The
             value of the concatenated xs becomes, say, y of length L1.

Anderson, et al.                                               [Page 14]

RFC 138               Data Reconfiguration Service          April 1971

         3.  The data type and the length expression together specify a
             field length of the input area (call it L2) that begins at
             the current stream pointer position.

         4.  The value of y is truncated to y' if L1 > L2.  Call the
             truncated length L1'.

         5.  If the term is an input stream term, then the value y' of
             length L1' is compared to the input value beginning at the
             current input pointer position.

         6.  If the values are identical over the length L1' then the
             input value of length L2 (may be greater than L1') starting
             at the current pointer position in the input area, becomes
             the value of the term.

   In an output stream term, the procedure is the same except that the
   source of input is the value of the term(s) named in the value
   expression and the data is emitted in the output stream.

   The above procedure is modified to include a one term look-ahead
   where lengths are indefinite because of the arbitrary symbol, #.

Restrictions and Interpretations of Term Functions

   1.  Terms specifying indefinite lengths, through the use of the #
         symbol must be separated by some type-specific data such as a
         literal.  (A literal isn't specifically required, however.  An
         arbitrary number of ASCII characters could be terminated by a
         non-ASCII character.)  # is legal only in the input stream, not
         in the output stream.

   2.  Truncation and padding is as follows:
         a)  Character to character (A <--> E) conversion is left
             justified and truncated or padded on the right with blanks.
         b)  Character to numeric and numeric to numeric conversions are
             right-justified and truncated or padded on the left with
         c)  Numeric to character conversion is right-justified and
             left-padded with blanks.

   3.  The following are ignored in a form definition over the control
         a)  TAB, carriage return, etc.
         b)  blanks except within quotes.
         c)  /* string */ is treated as comments except within quotes.

   4.  The following defaults prevail where the term part is omitted.

Anderson, et al.                                               [Page 15]

RFC 138               Data Reconfiguration Service          April 1971

         a)  The replication expression defaults to one.
         b)  The data type defaults to type B.
         c)  The value expression of an input stream term defaults to
             the value found in the input stream, but the input stream
             must conform to data type and length expression.  The value
             expression of an output stream term defaults to padding
         d)  The length expression defaults to the size of the quantity
             determined by replication expression, data type, and value
         e)  Control defaults to the next sequential term if a term is
             successfully applied; else control defaults to the next
             sequential rule.  If _where_ evaluates to an undefined
             _label_ the form fails.

   5.  Arithmetic expressions are evaluated left-to-right with no

   6.  The following limits prevail.

         a)  Binary lengths are <= 32 bits
         b)  Character strings are <= 256 8-bit characters
         c)  Identifier names are <= 4 characters
         d)  Maximum number of identifiers is <= 256
         e)  Label integers are >= 0 and <= 9999

   7.  Value and length operators product 32-bit binary integers.  The
         value operator is currently intended for converting A or E type
         decimal character strings to their binary correspondents.  For
         example, the value of E'12' would be 0......01100.  The value
         of E'AB' would cause the form to fail.

Anderson, et al.                                               [Page 16]

RFC 138               Data Reconfiguration Service          April 1971


   Sequencing may be explicitly controlled by including control in a

        control ::=  :options | <null>
        options ::=  S(where) | F(where) | U(where)
                     S(where) , F(where) |
                     F(where) , S(where)

        where   ::=  arithmeticexpression | R(arithmeticexpression)

   S, F, and U denote success, fail, and unconditional transfers,
   respectively.  _Where_ evaluates to a _rule_ label, thus transfer can
   be effected from within a rule (at the end of a term) to the
   beginning of another rule.  R means terminate the form and return the
   evaluated expression to the initiator over the control connection (if
   still open).

   If terms are not explicitly sequenced, the following defaults

   1)  When a term fails go to the next sequential rule.
   2)  When a term succeeds go to the next sequential
       term within the rule.
   (3) At the end of a rule, go to the next sequential

   Note in the following example, the correlation between transfer of
   control and movement of the input pointer.

        1   XYZ(,B,,8:S(2),F(3)) : XYZ ;
        2   . . . . . . .
        3   . . . . . . .

   The value of XYZ will never be emitted in the output stream since
   control is transferred out of the rule upon either success or
   failure.  If the term succeeds, the 8 bits of input will be assigned
   as the value of XYZ and rule 2 will then be applied to the same input
   stream data.  That is, since the complete rule 1 was not successfully
   applied, then the input stream pointer is not advanced.

Anderson, et al.                                               [Page 17]

RFC 138               Data Reconfiguration Service          April 1971



   The following examples (forms and also single rules) are simple
   representative uses of the Form Machine.  The examples are expressed
   in a term-per-line format only to aid the explanation.  Typically, a
   single rule might be written as a single line.


   To insert a field, separate the input into the two terms to allow the
   inserted field between them.  For example, to do line numbering for a
   121 character/line printer with a leading carriage control character,
   use the following form.

   (NUMB.<=>.1);       /*initialize line number counter to one*/
   1 CC(,E,,1:F(R(99))),  /*pick up control character and save
                            as CC*/
                          /*return a code of 99 upon exhaustion*/
   LINE(,E,,121 : F(R(98)))    /*save text as LINE*/
   :CC,               /*emit control character*/
   (,E,NUMB,2),       /*emit counter in first two columns*/
   (,E,E".",1),       /*emit period after line number*/
   (,E,LINE,117),     /*emit text, truncated in 117 byte field*/
   (NUMB.<=.NUMB+1:U(1));    /*increment line counter and go to
                               rule one*/;;


   Data to be deleted should be isolated as separate terms on the left,
   so they may be omitted (by not emitting them) on the right.

   (,B,,8),           /*isolate 8 bits to ignore*/
   SAVE(,A,,10)       /*extract 10 ASCII characters from
                        input stream*/
   :(,E,SAVE,);      /*emit the characters in SAVE as EBCDIC
                       characters whose length defaults to the
                       length of SAVE, i.e., 10, and advance to
                       the next rule*/

   In the above example, if either input stream term fails,
   the next sequential rule is applied.


   Some devices, terminals and programs generate variable length
   records.  To following rule picks up variable length EBCDIC records

Anderson, et al.                                               [Page 18]

RFC 138               Data Reconfiguration Service          April 1971

   and translates them to ASCII.

   CHAR(,E,,#),       /*pick up all (an arbitrary number of)
                        EBCDIC characters in the input stream*/
   (,X,X"FF",2)       /*followed by a hexadecimal literal,
                        FF (terminal signal)*/
   :(,A,CHAR,),       /*emit them as ASCII*/
   (,X,X"25",2);      /*emit an ASCII carriage return*/


   It is often necessary to prefix a length field to an arbitrarily long
   character string.  The following rule prefixes an EBCDIC string with
   a one-byte length field.

   Q(,E,,#),          /*pick up all EBCDIC characters*/
   TS(,X,X"FF",2)     /*followed by a hexadecimal literal, FF*/
   :(,B,L(Q)+2,8),    /*emit the length of the characters
                        plus the length of the literal plus
                        the length of the count field itself,
                        in an 8-bit field*/
   Q,                 */emit the characters*/
   TS;                */emit the terminal*/


   It is often desirable to reorder fields, such as the following

   Q(,E,,20), R(,E,,10) , S(,E,,15), T(,E,,5) : R, T, S, Q ;

   The terms are emitted in a different order.


   In systems such as HASP, repeated sequences of characters are packed
   into a count followed by the character, for more efficient storage
   and transmission.  The first form packs multiple characters and the

Anderson, et al.                                               [Page 19]

RFC 138               Data Reconfiguration Service          April 1971

   second unpacks them.
   /*form to pack EBCDIC streams*/
   /*returns 99 if OK, input exhausted*/
   /*returns 98 if illegal EBCDIC*/
   /*look for terminal signal FF which is not a legal EBCDIC*/
   /*duplication count must be 0-254*/
   1 (,X,X"FF",2 : S(R(99))) ;
   /*pick up the EBCDIC and initialize count/*
     CHAR(,E,,1 : F(R(98))) , (CNT .<=. 1) ;
   /*count consecutive EBCDICs like CHAR*/
   2 (,E,CHAR,1 : F(3)) , (CNT .<=. CNT+1 : U(2)) ;
   /*emit count and current character*/
   3 : (,B,CNT,8), CHAR, (:U(1));
   /*end of form*/;;

   /*form to unpack EBCDIC streams*/
   /*look for terminal*/
   1 (,X,X"FF",2 : S(R(99))) ;
   /*emit character the number of times indicated*/
   /*by the counter contents*/
   CNT(,B,,8), CHAR(,E,,1) : (CNT,E,CHAR,CNT:U(1));
   /*failure of form*/
   (:U(R(98))) ;;


   The following are some proposed uses of the DRS that were submitted
   by the sites indicated.

   1.  Pack/unpack text files.
   2.  Preprocessor to scan META compiler input.
   3.  Perhaps graphics.

   1.  Reformatting within file transfer service.
   2.  Character conversions.
   3.  Possible graphics service (Evans and Sutherland output
   4.  Reformat arguments of subroutines remote to each other.

   1.  Dependent upon remote use of DRS for many remote

   1.  Would be essential to data transfer in general.

Anderson, et al.                                               [Page 20]

RFC 138               Data Reconfiguration Service          April 1971

   2.  Could be used in relation to data management language.

   1.  Checkout of I/O formats of file system.
   2.  Debugging Network services in general.
   3.  Mapping their services into future standards.

   1.  To describe RJO/RJE message formats at UCSB.
   2.  To describe RJS message formats at UCLA.
   3.  To adapt Network to existing services, in general.

   1.  Character conversions.
   2.  Testing data formats going into data bases for correct
       field formatting.


   Four sites currently plan to implement and offer the service on an
   experimental basis.

   1.  MIT    Implementation of forms interpreter in MIDAS
              (assembly).  Perhaps Tree Meta compiler of
              forms.  Implementation on PDP-10.

   2.  UCLA   Implementation on SIGMA-7 employing META-7
              to compile forms.

   3.  UCSB   Implementation on 360/75.

   4.  RAND   Initial implementation on 360/65; compiler to be written
               in graphics CPS; compiled intermediate forms to be
               interpreted by assembler language subroutine.  Later
               implemented on PDP-10.

   In addition to the above sites, the University of Illinois and Mitre
   plan to experiment with the service.

Anderson, et al.                                               [Page 21]

RFC 138               Data Reconfiguration Service          April 1971

                                APPENDIX A

Note 1 to the DRS Working Group

   As you recall, we spent considerable time in discussing the use and
   meaning of the arbitrary symbol, #.  To summarize, it was clear that
   inclusion of the # in both replication and length expressions led to
   ambiguities.  We settled on its restricted use in the length
   expression of an input term, although no one was entirely satisfied
   with this definition.

   Recently, Jim White has again commented on the #.  Jim feels that it
   is curious that one can pick up an arbitrary number of EBCDIC
   characters, for example, but can't pick up an arbitrary number of
   specific EBCDIC characters such as EBCDIC A's.  Jim feels that a more
   natural way to interpret the length, value, and replication
   expressions would be as the IBM OS assembler interprets the
   attributes of the pseudo instruction, define constant (CD).

   The IBM OS assembler uses the following format.

        1             2              3           4
   duplication       type        modifiers   nominal value

   The duplication factor, if specified, causes the constant to be
   generated the number of times indicated by the factor.  The type
   defines the type of constant being specified.  Modifiers describe the
   length, scaling, and exponent of the constant.  Nominal value
   supplies the constant described by the subfields that precede it.

   Assume that we use the # only as a duplication factor (replication
   expression).  Hence, in the example of the form to pack EBCDIC
   characters, the counter and looping can be eliminated.

   CHAR(,E,,1) ;
   LEN(#,#,CHAR,1) : (,B,L(LEN)+1,*) , CHAR ;

   The interpretation is that the data type, length expression, and
   value expression make up the unit value.  This quantity can then be
   replicated.  As our document now stands, only the data type and value
   expression make up the unit value.

   The application of a term according to Jim's suggestion is as
   1.  The data type, value expression, and length expression together
       specify a unit value, call it x.

Anderson, et al.                                               [Page 22]

RFC 138               Data Reconfiguration Service          April 1971

   2.  The replication expression specifies the number of times x is to
       be repeated.  The value of the concatenated xs becomes y of
       length L.
   3.  If the term is an input stream term then the value beginning at
       the current input pointer position.
   4.  If the input value satisfies the constraints of y over length L
       then the input value of length L becomes the value of the term.

Note 2 to the DRS Working Group

   There has been recent debate of whether the input pointer should be
   advanced upon successful completion of a rule (as it now is defined)
   or upon successful completion of each term.  See the example on page
   22.  If the input pointer is advanced upon successful completion of a
   term, then rules become equivalent to terms.

   I would like to for us to discuss at the SJCC both the term
   attributes and the input pointer advance issues.


       [ This RFC was put into machine readable form for entry ]
       [ into the online RFC archives by Katsunori Tanaka 4/99 ]

Anderson, et al.                                               [Page 23]