RTL/2 (Real-Time Language) is a discontinued high-level programming language for use in real-time computing, developed at Imperial Chemical Industries, Ltd. (ICI), by J.G.P. Barnes. It was originally used internally in ICI but was distributed by SPL International in 1974.^[1] It was based on concepts from ALGOL 68, and intended to be small and simple.^[2][3][4] RTL/2 was standardised in 1980 by the British Standards Institution.^[5]

The data types in RTL/2 were strongly typed, with separate compiling. The compilation units contained one or more items named bricks, i.e.:

• Procedure bricks
• Data bricks
• Stack bricks

A procedure brick was a procedure, which may or may not return a (scalar) value, have (scalar) parameters, or have local (scalar) variables. The entry mechanism and implementation of local variables was reentrant. Non-scalar data could only be accessed via reference (so-called REF variables were considered scalar).

A data brick was a named static collection of scalars, arrays and records. There was no heap or garbage collection, so programmers had to implement memory management manually.

A stack brick was an area of storage reserved for running all the procedures of a single process and contained the call stack, local variables and other housekeeping items. The extent to which stack bricks were used varied depending on the host environment in which RTL/2 programs ran.

Access to the host environment of an RTL/2 program was provided via special procedure and data bricks called SVC procedures and SVC data. These were accessible in RTL/2 but implemented in some other language in the host environment.

TITLE Hello World;

LET NL=10;

EXT PROC(REF ARRAY BYTE) TWRT;

ENT PROC RRJOB() INT;
TWRT("Hello World#NL#");
RETURN(1);
ENDPROC;

RTL/2 compiles to assembly language and provides the CODE statement to allow including assembly language in RTL/2 source code. This is only available when compiled with a systems programming option (CN:F)

The CODE statement takes two operands: the number of bytes used by the code insert and the number of bytes of stack used.

Within code statements two trip characters are used to access RTL/2 variables. These vary between different operating systems. On a Digital Equipment Corporation (DEC) PDP-11 running RSX-11M, and a VAX running VMS, the trip characters are * and /.

While the specifics varied by operating system the following is an example of a code insert on VAX/VMS:

CODE 6,0;
JMP CODE_ENT ; This code insert can be set to a fixed length as it jumps to a new psect.
; this method is especially useful on systems such as VMS where the length of
; instructions is variable
.SAVE_PSECT ; Save current program section
.PSECT ASMB_CODE,EXE,NOWRT,LONG
CODE_ENT:
MOVL *PARAM1(AP),*COUNTER/MYDATA
JMP CODE_EX
.RESTORE_PSECT
CODE_EX:
*RTL

This code insert moves the value of a variable passed into the RTL/2 procedure into a variable named COUNTER in a data brick named MYDATA.

SPL published a range of documentation for RTL/2. Each such document was assigned a reference number. The following is an incomplete list.

RTL/2 Ref 1  – RTL/2 Language Specification
RTL/2 Ref 2  – Introduction to RTL/2
RTL/2 Ref 3  – RTL/2 Training Manual
RTL/2 Ref 4  – System Standards
RTL/2 Ref 5  – Stream I/O
RTL/2 Ref 18 – Hints on writing RTL/2 Programs
RTL/2 Ref 26 – Language Reference Card
RTL/2 Ref 39 – Run time environment on the PDP-11
RTL/2 Ref 63 – User Manual for the PDP-11 under RSX-11M
RTL/2 Ref 107 – VAX/VMS RTL/2 User Manual
RTL/2 REF 130 – The RTL/2 32-bit run time environment on the VAX