XTRAN(ID:2630/xtr001)
Interactive extensible dialect of FORTRAN
for eXtensible forTRAN
Interactive extensible dialect of FORTRAN, which evolved over time into an interactive extensible dialect of ALGOL!
Bob Bemer IBM 1958 (extensible FORTRAN)
Green, Julien et al, Applied Programming Systems, IBM White PLains NY 1959 (Extensible ALGOL)
Influential (according to Bemer) on IAL
But obviously then became influenced by IAL and Algol 60 (which only goes to show you how family trees don't work for the history of ideas)
People:
Related languages
| XTRAN | => | IAL | Influence | |
| XTRAN | => | Wegstein string Algol | Influence |
References:
in [ACM] CACM 1(07) July 1958 view details
in [ACM] CACM 1(07) July 1958 view details
Introduction
We are presently constructing an experimental ALGOL processor for the IBM 709. Our work has given rise to several considerations which we present here for those interested in ALGOL. The discussion in this paper is in three parts:
1. External and internal representations
2. Modifications to ALGOL
3. Symbol manipulation additions to ALGOL
1. External and Internal Representations
The preliminary report on ALGOL defines the basic symbols of the language. A subset of these can be represented externally (now) only as words; e.g., go to, do, if, etc. Nevertheless, they stand for single characters which will have some internal representation. A good processor translates external representation to internal. The dictionary used in making this translation should be flexible enough to allow arbitrary changing of the external representation of an internal symbol. We can therefore say that the processing of internal symbols can be independent of the external language.1 The choice of external symbols is more important in designing a language for efficient communication between human beings than in designing the processor for that language. For can be translated as easily as vary, loop, or do.
In any language certain symbols have syntactical meanings determined by the symbols contiguous with them. For communication between people this is practical because they can scan a group of symbols at a time. In a computer the scan is usually performed one symbol at a time and therefore all implied (syntactical) operations are converted to explicit internal symbols. This is as it should be, because the language should be designed for the convenience of the user and not the processor. (In the IBM 709 ALGOL processor this is done by a pre-scan which rearranges to remove implied syntax and redundancy.)
2. Modifications to ALGOL
A. if Statements
B. Arithmetic Expressions
C. Precedence of Operators
D. do Statements
3. Symbol Manipulation Additions to ALGOL
In the development of an ALGOL processor we found it necessary to use some language to describe the functions of the processor. It seemed natural to use the string and symbol terminology of the ALGOL report in the December 1958 Communications. The translation process could then be described as a sequence of operations upon the ALGOL string of symbols that creates a new string in the object language. It was felt that if these operations were formalized into an algebra which could be added to ALGOL, the ALGOL could be used to describe its own processor. The tremendous advantage in being able to do this (aside from the obvious one of increasing the power of the language) is that it will enable us to write a crude ALGOL processor which we can use in a "bootstrap" operation to a final processor. The same technique obviously would have similar advantages for generating processors and simulators of all kinds. The practicality of implementing a symbol manipulation language was suggested by the work of Professor John McCarthy of MIT and of Newell, Shaw, and Simon of the IPL.
Since symbol manipulation has not yet been specified in ALGOL, we have incorporated the following additions as a practical basis for producing the processor.
in [ACM] CACM 1(07) July 1958 view details
GREEN, JULIEN. Symbol manipulation in XTRAN. Comm. ACM 3, 4 (April 1960), 213-214.
The symbol manipulation language incorporated in the XTRAN processor is described. The description is in the form of tables and includes the definition of a "string", string expressions, and string procedures. These functions and procedures require that symbols be treated as integers. James H. Brown, Norristown, Pa.
in ACM Computing Reviews, January-December 1960 view details
In the September, 1959 Communications, we briefly described the symbol manipulation abilities of our processor. The following is a more complete description of the symbol manipulation language incorporated in our current processor.
The language we are using assumes that symbols are integers (which is the case in a binary machine), and they are handled in the same fashion as integer values. This assumption was made because it provides a usable system and it seems more convenient than creating a new class of value for symbols.
In the description the following representations are used :
ST for a string name
SY for a symbol name
I for an integer name
L for a statement label.
1. Type Statement
ST String (ST1, ST2, ...)--'STi' are identifiers of strings.
A string is a sequence of symbols,
2. String Literal
A string may be represented in the language as follows:
S(n,aaaa)--Identifies the 'n' symbols following the comma as a string.
3. String Expressions (S)
S ~ (S)
S ~ S(n,aaa)
~ F( ) where F is any function defining a string vahie
~ ST
~ ST1 + ST2 Concatenate string ST1 to end of string ST2
~ ST1 - ST2 Remove all ST1 strings from string ST2
~ ST + SY Add symbol SY to end of string ST
~ SY + ST Add symbol SY to beginniiig of string ST
4. String Functions
| Representation | Name | Class of Value Defined | Description of Value |
| Norm(ST) | Norm | I | Positive integer equal to the number of symbols in the string ST |
| GETNS(ST, I) | Get nth symbol | SY | Ith symbol of string |
| FIRST(ST) | First symbol of string | SY | GETNS (ST, 1) |
| SUB(ST, I1, I2) | Substring | ST | ~ as defined in [1] |
| FOS(ST, SY) | First occurrence of symbol | I | If the symbol SY is in the string then the value defined is an integer equal to the position which the SY occupies; otherwise the value is zero. |
| AI(I) | Transform an integer to a string | ST | |
| EI(ST) | Transform an string to a integer | ST |
5. Procedures
| Call | Name | Description |
| FREES(ST) | Free string | ST := NULL STRING |
| REMNS(ST, I, SY) | Remove I-lh symbol | Set SYequal to the I-th symbol of string ST, Remove I-th symbol of stringST |
| POINT(ST, I1, I2) | Set pointer | Set integer I2 to designate the I1-th symbol of string ST |
| SEQ(l1, SY, 12) | Sequence | Set symbol SY equal to the syrnbol which I2 designates and advance I2 to the next symbol in the string If the proeedure is called after I2 has passed the last symbol of the string transfer to statement L |
| INSRT(ST, I, SY) | Insert | Insert the symbol SY between the Ith and the (I+1)th symbols of the string ST so that SY becomes the (I+l)th symbol of string. |
| REPLS(ST, SY1, SY2) | Replace symbol with symbol | Wherever the symbol SY1 occurs in string ST replace with symbol SY2 |
| RPLST(ST1, SY1, ST2) | Replace symbol by string | Wherever replace symbol SY1 occurs in string ST1 replace by the string ST2 |
in [ACM] CACM 3(04) April 1960 view details
in [ACM] CACM 3(04) April 1960 view details
in [ACM] CACM 3(04) April 1960 view details
in [ACM] CACM 4(01) (Jan 1961) view details
in Artificial languages view details
in [ACM] CACM 15(06) (June 1972) view details