Programmer's ROM - The Computer Language Library

home *** CD-ROM | disk | FTP | other *** search

/ Programmer's ROM - The Computer Language Library / programmersrom.iso / ada / metric / halstead.doc < prev next >

Wrap

Text File | 1988-05-03 | 34.1 KB | 1,027 lines

:::::::::::::: halstead.cnt :::::::::::::: [NOSC.RELEASES.V0302.HALSTEAD.SOURCE]IHSUBTYPE.DAT 6 11 [NOSC.RELEASES.V0302.HALSTEAD.SOURCE]IHSERIES_.DAT 6 11 [NOSC.RELEASES.V0302.HALSTEAD.SOURCE]IHAGG_NAM.DAT 6 11 [NOSC.RELEASES.V0302.HALSTEAD.SOURCE]IHCASE_AL.DAT 6 11 [NOSC.RELEASES.V0302.HALSTEAD.SOURCE]IHHANDLER.DAT 6 11 [NOSC.RELEASES.V0302.HALSTEAD.SOURCE]IHBLOCK_S.DAT 6 11 [NOSC.RELEASES.V0302.HALSTEAD.SOURCE]IHINNER_R.DAT 6 11 [NOSC.RELEASES.V0302.HALSTEAD.SOURCE]IHVARIABL.DAT 8 12 [NOSC.RELEASES.V0302.HALSTEAD.SOURCE]IHTYPE_DE.DAT 8 12 [NOSC.RELEASES.V0302.HALSTEAD.SOURCE]IHGENERIC.DAT 6 11 [NOSC.RELEASES.V0302.HALSTEAD.SOURCE]IHTASK_DE.DAT 6 11 [NOSC.RELEASES.V0302.HALSTEAD.SOURCE]HALSTEAD.ADA 70 252 [NOSC.RELEASES.V0302.HALSTEAD.TEST]SCANNERS.SPC 15 245 [NOSC.RELEASES.V0302.HALSTEAD.TEST]SCANNERS.BDY 85 283 :::::::::::::: halstead.mem :::::::::::::: The Halstead Tool COMMAND FORMAT -- Halstead : Computes HALSTEAD's software science formulas for Ada -- program units -- Version 3.01-1.01 subtype STRING_TYPE is STRING; type STRING_LIST is array (POSITIVE range <>) of STRING_TYPE; procedure HALSTEAD( UNITS : in STRING_LIST; OUTPUT : in STRING := ""; LIBRARY : in STRING := "_AIELIB"; ); -- UNITS : Names of the Ada compilation units -- OUTPUT : Name of the report file (defaults to standard output) -- LIBRARY : Name of the program library PARAMETERS UNITS This parameter is a list of string values, each of which is the name of a unit which is to be analyzed. This parameter is required and the list must have at least one value. A subunit may be specified using a qualified name, as in ("PARENT.SUBUNIT"), which would analyze the subunit called SUBUNIT of the compilation unit PARENT. OUTPUT This is the name of the output file. It is optional and if it is not used, output will be directed to standard output. LIBRARY This parameter is a string which is the program library containing the compilation units to be analyzed. This This parameter is optional and defaults to "_AIELIB" (WARNING: in the current release, this parameter is ignored) As an example assume that there exists a program library called [.AIELIB] and that it is the current directory. Also assume that in the program library there exists a compilation unit which is a procedure called MAIN. The objective is to run the HALSTEAD tool on the unit MAIN. The following command can be used: $ HALSTEAD(UNITS => ("MAIN")); This will compute the Halstead metrics on the unit MAIN and write The Halstead Tool Page 2 the report to the default output file. Note that an implicit specification is constructed by the Byron/AIE compiler front end for any compilation unit that does not have an explicit specification. The HALSTEAD tool does not distinguish these implicit specifications from explicit specifications. In the above example, the metrics will be computed for both the specification and body of MAIN, regardless of whether or not a source file containing a specification for MAIN was compiled into the library. DESCRIPTION The Halstead tool uses the number of operands and operators in the program to compute the expected length of the program. Each symbol in the program is classified as an operand or an operator. Operators are mathematical functions such as +, -, *, and /, as well as parenthesis and subprogram calls. Operands are constants and variables. From the number of operands and operators, Halstead makes predictions about the program such as intelligence content, number of expected bugs, programming effort, and the level at which the program is written. See [Els]. The metrics are computed on a compilation unit by breaking it down into its fundamental units. These are the basic units out of which Ada programs are built. These are subprograms, packages, and tasks. The metrics are computed in a lexically nested fashion. For example inside a package body which contains subprogram bodies in it each of the subprograms has the metrics computed for it. PREPARATION HALSTEAD computes the metrics for an ADA program by using the source program's DIANA representation. Therefore it is necessary to have the source program's DIANA in an ADA program library before running HALSTEAD. This can be accomplished in a number of steps. 1. First it is necessary to create an ADA program library which will contain the DIANA representations of the source program which HALSTEAD will analyze. This is accomplished by the command: AIEPLIB This command creates a directory called [._AIELIB] in the directory it is executed from. 2. Following this it is necessary to successfully run the source The Halstead Tool Page 3 program through ADASEM, which is semantics phase of the ADA compiler. This generates the DIANA representation for the source program. In order to run the semantics phase execute the command: ADASEM SOURCE.ADA RESTRICTIONS It is important to note that Halstead places some restrictions on the program whose estimates are being computed. It must be a "pure" program. Halstead defines a pure program to be one that does not contain any of the following constructions. 1. No complementary operators. There cannot be successive applications of two comple- mentary operations to the same operand. For example: z - t + t -> r 2. There cannot be ambigous operands. A given operand cannot refer to different objects at different times in the program. For instance in the following example r represents first a sum and then a product. p + q -> r r * r -> r Should be written as: p + q -> t t * t -> r 3. There cannot be synonomous operands. Two different names for the same object. There can only exist one name for one object. p + q -> t1 p + q -> t2 t1 * t2 -> r Should be written as: p + q -> t1 t1 * t1 -> r 4. Common subexpressions cannot exist. Any expression which is used more than once must be The Halstead Tool Page 4 assigned to a variable. (p + q) * (p + q) -> r Should be written as: (p + q) -> t1 t1 ** 2 -> r 5. No unwarranted assignments. If an assignment is made and the value used only once the assignment is unnecessary. p + q -> t1 t1 ** 2 -> r Should be written as: (p + q) ** 2 -> r 6. Expressions should be factored. p*p + 2*p*q + q*q Should be written as: (p + q) ** 2 METRICS COMPUTED Halstead begins by defining the following symbols: 1. NUMBER OF UNIQUE OPERATORS n1 2. NUMBER OF UNIQUE OPERANDS n2 3. THE VOCABULARY n = n1 + n2 4. TOTAL NUMBER OF OPERATORS N1 The Halstead Tool Page 5 5. TOTAL NUMBER OF OPERANDS N2 6. LENGTH OF PROGRAM N = N1 + N2 7. ESTIMATED PROGRAM LENGTH: N^ = n1*log2(n1) + n2*log2(n2) Where n1 and n2 are the number of distinct operators and operands. N^ is an estimator for program length. 8. PROGRAM VOLUME: V = N * log2(n) This is the size of the program in bits. Where N is program length and n is the vocabulary. This is because we need log2(n) bits for each symbol and there are N symbols. 9. POTENTIAL VOLUME V* = (2 + n2*) * log2(2 + n2*) This represents the size of the algorithm if it existed as a built-in function or a subprogram. It is the most compact form of the algorithm. The 2 is the minimum number of opera- tors the procedure name and a grouping symbol needed to per- form the function. The n2* represents the minimum number of operands needed if the algorithm existed as a built-in func- tion. The n2* can be viewed as the minimum number of input/output parameters for the algorithm if it existed as a builtin function. 10. PROGRAM LEVEL L = (V* / V) This measures the level at which the program was written. The the closer the volume V is to the potential volume V* the higher the level. 11. PROGRAM LEVEL APPROXOMATION L^ = (n1* / n1) * (n2 * N2) The Halstead Tool Page 6 This is an approximation for program level. 12. INTELLIGENCE CONTENT I = L^ * V This is the amount of information contained in a program. 13. PROGRAMMING EFFORT E = ( V / L ) The effort of programming increases as the volume of the pro- gram increases and decreases as the level increases. This means the larger a program the more difficult the effort and the higher the level of implemenation the the easier the effort. This effort also can be viewed as the number of ele- mentary discriminations necessary to complete a program. [Fitz Lov] 14. PROGRAMMING TIME T^ = E / S = (V ** 2) / (S * V*) The amount of time required to implement an algorithm is directly proportional to the programing effort E divided by a constant S. S is the speed of the programmer, or the number of mental discriminations per second. Halstead calculated S equaled 18. This relates to the work of Stroud, a psycholo- gist, who estimates S between 5 and 20. [Str], [Hal] 15. LANGUAGE LEVEL lambda = L * V* lambda = L**2 * V This measure enables us to rank languages in terms of their expressive power. The higher the measure the higher the levle of the langusage. For example Pascal has a lambda greater than the lambda for Fortran. 16. NUMBER OF DELIVERED ERRORS B = (E ** 2/3) / E0 B^ = (E ** 2/3) / 3000 The Halstead Tool Page 7 B^ = V / 3000 E0 is the mean number of elementary discriminations between errors. E0 has been estimated to be 3000. 17. Modularization n log2 (n/2) = M((n-n*)/M + n*) log2 (( (n-n*) / M ) /2) M is the number of modules the program should be divided into. OUTPUT FORMAT The output format will be organized by Ada units contained in compilation unit specified. The metrics computed for each nested unit will be displayed in reverse linear elaboration order. This means the most deeply nested units will have the metrics computed for them appear first in the report. The output format below has each column representing a different metric. The numbers heading each column correspond to the metrics described above. An example output for library unit appears below. The following is a sample input and output from HALSTEAD: SAMPLE INPUT: ---------------------------------------------------------------- package SCANNERS is . . . procedure scan_Delimited( --| Scan string delimited by a single character Scanner: in out scanner_Type; S: in string ) is quote: character; begin Scanner.First := Scanner.Index; if Scanner.Index <= Scanner.Max_Index then quote := S(Scanner.Index); Scanner.Index := Scanner.Index + 1; Scanner.First := Scanner.Index; while Scanner.Index <= Scanner.Max_Index and then S(Scanner.Index) /= quote loop Scanner.Index := Scanner.Index + 1; The Halstead Tool Page 8 end loop; end if; Scanner.Length := Scanner.Index - Scanner.First; Scanner.Last := Scanner.Index - 1; if Scanner.Index <= Scanner.Max_Index and then S(Scanner.Index) = quote then -- Null string? Scanner.Index := Scanner.Index + 1; end if; end scan_Delimited; ---------------------------------------------------------------- SAMPLE OUTPUT: PROCEDURE BODY OF SCANNERS.SCAN_DELIMITED AT LINE 205 UNIQUE OPERATORS 17 UNIQUE OPERANDS 11 TOTAL OPERATORS 60 TOTAL OPERANDS 61 VOCABULARY 28 PROGRAM LENGTH 121 ESTIMATED LENGTH 107.54 PROGRAM VOLUME 579.68 POTENTIAL VOLUME 15.50 PROGRAM LEVEL .03 ESTIMATED LEVEL 78.94 INTELLIGENCE CONTENT 45760.71 PROGRAMMING EFFORT 21674.78 PROGRAMMING TIME 4334.96 LANGUAGE LEVEL .41 DELIVERED ERRORS 4.17 ESTIMATED ERRORS .19 REFERENCES [Els] Elshoff, James, L., "Measuring Commercial PL/1 Programs Using Halstead's Criteria", SIGPLAN Notices, May 76, pp38-46 [Fitz Lov] Fitzsimmons, A., Love, T. "A Review and Evaluation of Software Science", Computing Surveys, Vol. 10, No. 1, March 1978, pp3-18 [Hal] Halstead, M. H., Elements of Software Science, Elsevier North Holland, Inc., Operating and Proramming Systems Series, New York, 1977 [Str] Stroud, John, M., The Fine Structure of Psychological Time, "Annals of New York Academy of Sciences", 1966, pp623-631 The Halstead Tool Page 9 APPENDIX A: INSTALLATION AND MODIFICATION Note: This version of HALSTEAD ignores the LIBRARY parameter. The directory containing the AIE program library must be your default directory when you run HALSTEAD. This is a limitation of the AIE Program Library Manager which should be corrected at some time in the future. To use the HALSTEAD tool you must: 1. Be a "HIF User". See the Byron/AIE Programmers Reference Manual for details. This requires that certain logical symbols be defined and that you have run the ADDUSER tool to make yourself known to the HIF database. 2. Define a symbol HALSTEAD so that VMS knows what executable to run when you type the HALSTEAD command. For example: $ HALSTEAD :== $USER1:[NOSC.TOOLS]HALSTEAD.EXE NOTE: The full path name of the executable is required in the definition of the symbol. The pathname given here is just an example and will be different on your system. 3. Create a Byron/AIE program library and use the Byron/AIE compiler front end to analyze source code into it. The AIEPLIB and AIECOMP commands are used respectively. 4. Enter a command with appropriate parameters. For example, $ HALSTEAD(("MYPROG", "TEST.OUT")); Entering the command with no parameters gives a brief description of how to use the tool. To build the HALSTEAD tool: 1. Compile all the abstractions into a program library (see READ.ME in abstractions directory for details). 2. Compile everything named in the HALSTEAD.CO file into the program library containing the abstractions or a sublibrary whose parent library contains all the abstractions. HALSTEAD.CO lists file names in a correct compilation order. 3. Link HALSTEAD with the program library where everything was compiled. To do this using the DEC Ada compiler type: $ ACS LINK HALSTEAD Files contained in the HALSTEAD.SOURCE directory: BLOCKU.SPC BLOCKU.BDY COMLIN.SPC The Halstead Tool Page 10 COMLIN.BDY DEFS.SPC COUNTYPE.SPC COUNT.SPC COUNT.BDY COUNTYPE.BDY DEFS.BDY HDB.SPC HDB.BDY SCTYPESP.SPC SCNAMEEX.SPC SCCONSTRA.SPC IHSUBTYPE.DAT OBJ.SPC OBJ.BDY IHSERIES.DAT IHAGGNAM.DAT SCCHOICE.SPC SCAGGCOM.SPC SCAGGCOM.BDY IHCASEAL.DAT IHHANDLER.DAT SCSTM.SPC SCITEM.SPC SCALTERNA.SPC SCALTERNA.BDY IHBLOCKS.DAT SRCUTIL.SPC SCBLOCKS.SPC SCBLOCKS.BDY IHINNERR.DAT SCCHOICE.BDY SCCOMPUN.SPC SCCOMPUN.BDY SCGENERAL.SPC SCCONSTRA.BDY IHVARIABL.DAT IHTYPEDE.DAT IDUTILS.SPC IDUTILS.BDY SCDEFID.SPC SCDEFID.BDY SCGENERAL.BDY IHGENERIC.DAT SCGENERIC.SPC SCGENERIC.BDY SCHEADER.SPC SCHEADER.BDY SCVARIANT.SPC SCINNERR.SPC SCINNERR.BDY IHTASKDE.DAT SCPKGDEF.SPC SCOBJECT.SPC SCSUBPDE.SPC SCITEM.BDY SCITERATI.SPC SCITERATI.BDY SCNAMEEX.BDY SCOBJECT.BDY SCPKGDEF.BDY SCSTM.BDY SCSUBPDE.BDY SCTYPESP.BDY SCVARIANT.BDY SRCUTIL.BDY HALSTEAD.ADA To change the assignemnt of a token as either an operator, operand, or neither, see COUNT.BDY. To change the computation of any of the HALSTEAD measures, see HDB.BDY. The file HALSTEAD.ADA contains the driver program. To test the HALSTEAD tool, create an AIE program library using AIEPLIB, compile [.TEST]SCANNERS.SPC and [.TEST]SCANNERS.BDY into the AIE program library using AIECOMP, make that library your default directory, and run HALSTEAD, as follows: $ AIEPLIB [.DEMOLIB] $ AIECOMP SCANNERS.SPC [.DEMOLIB] $ AIECOMP SCANNERS.BDY [.DEMOLIB] $ SET DEFAULT [.DEMOLIB] $ HALSTEAD(("SCANNERS"), "SCANNERS.OUT"); Then, verify that the file SCANNERS.OUT is the same as that shipped with this release. :::::::::::::: halstead.rno :::::::::::::: .RM 65 ! right margin makes page ~65 chars and almost centered .PS ,65 ! lines page numbers up with right margin .STHL ,,,0,,,,, ! turns off section numbering .KEEP ! Keep blank lines in literal text .T The Halstead Tool .C The Halstead Tool .HL Command Format .LT -- Halstead : Computes HALSTEAD's software science formulas for Ada -- program units -- Version 3.01-1.01 subtype STRING_TYPE is STRING; type STRING_LIST is array (POSITIVE range <>) of STRING_TYPE; procedure HALSTEAD( UNITS : in STRING_LIST; OUTPUT : in STRING := ""; LIBRARY : in STRING := "_AIELIB"; ); -- UNITS : Names of the Ada compilation units -- OUTPUT : Name of the report file (defaults to standard output) -- LIBRARY : Name of the program library .EL .HL PARAMETERS .LT UNITS This parameter is a list of string values, each of which is the name of a unit which is to be analyzed. This parameter is required and the list must have at least one value. A subunit may be specified using a qualified name, as in ("PARENT.SUBUNIT"), which would analyze the subunit called SUBUNIT of the compilation unit PARENT. OUTPUT This is the name of the output file. It is optional and if it is not used, output will be directed to standard output. LIBRARY This parameter is a string which is the program library containing the compilation units to be analyzed. This This parameter is optional and defaults to "_AIELIB" (WARNING: in the current release, this parameter is ignored) .EL .P As an example assume that there exists a program library called [._AIELIB] and that it is the current directory. Also assume that in the program library there exists a compilation unit which is a procedure called MAIN. The objective is to run the HALSTEAD tool on the unit MAIN. The following command can be used: .lt $ HALSTEAD(UNITS => ("MAIN")); .el This will compute the Halstead metrics on the unit MAIN and write the report to the default output file. .p Note that an implicit specification is constructed by the Byron/AIE compiler front end for any compilation unit that does not have an explicit specification. The HALSTEAD tool does not distinguish these implicit specifications from explicit specifications. In the above example, the metrics will be computed for both the specification and body of MAIN, regardless of whether or not a source file containing a specification for MAIN was compiled into the library. .HL DESCRIPTION .p The Halstead tool uses the number of operands and operators in the program to compute the expected length of the program. Each symbol in the program is classified as an operand or an operator. Operators are mathematical functions such as +, -, *, and /, as well as parenthesis and subprogram calls. Operands are constants and variables. From the number of operands and operators, Halstead makes predictions about the program such as intelligence content, number of expected bugs, programming effort, and the level at which the program is written. See [Els]. .p The metrics are computed on a compilation unit by breaking it down into its fundamental units. These are the basic units out of which Ada programs are built. These are subprograms, packages, and tasks. The metrics are computed in a lexically nested fashion. For example inside a package body which contains subprogram bodies in it each of the subprograms has the metrics computed for it. .hl PREPARATION .p HALSTEAD computes the metrics for an ADA program by using the source program's DIANA representation. Therefore it is necessary to have the source program's DIANA in an ADA program library before running HALSTEAD. This can be accomplished in a number of steps. .lt 1. First it is necessary to create an ADA program library which will contain the DIANA representations of the source program which HALSTEAD will analyze. This is accomplished by the command: AIEPLIB This command creates a directory called [._AIELIB] in the directory it is executed from. 2. Following this it is necessary to successfully run the source program through ADASEM, which is semantics phase of the ADA compiler. This generates the DIANA representation for the source program. In order to run the semantics phase execute the command: ADASEM SOURCE.ADA .el .HL RESTRICTIONS .p It is important to note that Halstead places some restrictions on the program whose estimates are being computed. It must be a "pure" program. Halstead defines a pure program to be one that does not contain any of the following constructions. .LT 1. No complementary operators. There cannot be successive applications of two comple- mentary operations to the same operand. For example: z - t + t -> r 2. There cannot be ambigous operands. A given operand cannot refer to different objects at different times in the program. For instance in the following example r represents first a sum and then a product. p + q -> r r * r -> r Should be written as: p + q -> t t * t -> r 3. There cannot be synonomous operands. Two different names for the same object. There can only exist one name for one object. p + q -> t1 p + q -> t2 t1 * t2 -> r Should be written as: p + q -> t1 t1 * t1 -> r 4. Common subexpressions cannot exist. Any expression which is used more than once must be assigned to a variable. (p + q) * (p + q) -> r Should be written as: (p + q) -> t1 t1 ** 2 -> r 5. No unwarranted assignments. If an assignment is made and the value used only once the assignment is unnecessary. p + q -> t1 t1 ** 2 -> r Should be written as: (p + q) ** 2 -> r 6. Expressions should be factored. p*p + 2*p*q + q*q Should be written as: (p + q) ** 2 .el .hl METRICS COMPUTED .p Halstead begins by defining the following symbols: .lt 1. NUMBER OF UNIQUE OPERATORS n1 2. NUMBER OF UNIQUE OPERANDS n2 3. THE VOCABULARY n = n1 + n2 4. TOTAL NUMBER OF OPERATORS N1 5. TOTAL NUMBER OF OPERANDS N2 6. LENGTH OF PROGRAM N = N1 + N2 7. ESTIMATED PROGRAM LENGTH: N^ = n1*log2(n1) + n2*log2(n2) Where n1 and n2 are the number of distinct operators and operands. N^ is an estimator for program length. 8. PROGRAM VOLUME: V = N * log2(n) This is the size of the program in bits. Where N is program length and n is the vocabulary. This is because we need log2(n) bits for each symbol and there are N symbols. 9. POTENTIAL VOLUME V* = (2 + n2*) * log2(2 + n2*) This represents the size of the algorithm if it existed as a built-in function or a subprogram. It is the most compact form of the algorithm. The 2 is the minimum number of opera- tors the procedure name and a grouping symbol needed to per- form the function. The n2* represents the minimum number of operands needed if the algorithm existed as a built-in func- tion. The n2* can be viewed as the minimum number of input/output parameters for the algorithm if it existed as a builtin function. 10. PROGRAM LEVEL L = (V* / V) This measures the level at which the program was written. The the closer the volume V is to the potential volume V* the higher the level. 11. PROGRAM LEVEL APPROXOMATION L^ = (n1* / n1) * (n2 * N2) This is an approximation for program level. 12. INTELLIGENCE CONTENT I = L^ * V This is the amount of information contained in a program. 13. PROGRAMMING EFFORT E = ( V / L ) The effort of programming increases as the volume of the pro- gram increases and decreases as the level increases. This means the larger a program the more difficult the effort and the higher the level of implemenation the the easier the effort. This effort also can be viewed as the number of ele- mentary discriminations necessary to complete a program. [Fitz Lov] 14. PROGRAMMING TIME T^ = E / S = (V ** 2) / (S * V*) The amount of time required to implement an algorithm is directly proportional to the programing effort E divided by a constant S. S is the speed of the programmer, or the number of mental discriminations per second. Halstead calculated S equaled 18. This relates to the work of Stroud, a psycholo- gist, who estimates S between 5 and 20. [Str], [Hal] 15. LANGUAGE LEVEL lambda = L * V* lambda = L**2 * V This measure enables us to rank languages in terms of their expressive power. The higher the measure the higher the levle of the langusage. For example Pascal has a lambda greater than the lambda for Fortran. 16. NUMBER OF DELIVERED ERRORS B = (E ** 2/3) / E0 B^ = (E ** 2/3) / 3000 B^ = V / 3000 E0 is the mean number of elementary discriminations between errors. E0 has been estimated to be 3000. 17. Modularization n log2 (n/2) = M((n-n*)/M + n*) log2 (( (n-n*) / M ) /2) M is the number of modules the program should be divided into. .el .HL OUTPUT FORMAT .p The output format will be organized by Ada units contained in compilation unit specified. The metrics computed for each nested unit will be displayed in reverse linear elaboration order. This means the most deeply nested units will have the metrics computed for them appear first in the report. The output format below has each column representing a different metric. The numbers heading each column correspond to the metrics described above. An example output for library unit appears below. .P The following is a sample input and output from HALSTEAD: .LT SAMPLE INPUT: ---------------------------------------------------------------- package SCANNERS is . . . procedure scan_Delimited( --| Scan string delimited by a single character Scanner: in out scanner_Type; S: in string ) is quote: character; begin Scanner.First := Scanner.Index; if Scanner.Index <= Scanner.Max_Index then quote := S(Scanner.Index); Scanner.Index := Scanner.Index + 1; Scanner.First := Scanner.Index; while Scanner.Index <= Scanner.Max_Index and then S(Scanner.Index) /= quote loop Scanner.Index := Scanner.Index + 1; end loop; end if; Scanner.Length := Scanner.Index - Scanner.First; Scanner.Last := Scanner.Index - 1; if Scanner.Index <= Scanner.Max_Index and then S(Scanner.Index) = quote then -- Null string? Scanner.Index := Scanner.Index + 1; end if; end scan_Delimited; ---------------------------------------------------------------- SAMPLE OUTPUT: PROCEDURE BODY OF SCANNERS.SCAN_DELIMITED AT LINE 205 UNIQUE OPERATORS 17 UNIQUE OPERANDS 11 TOTAL OPERATORS 60 TOTAL OPERANDS 61 VOCABULARY 28 PROGRAM LENGTH 121 ESTIMATED LENGTH 107.54 PROGRAM VOLUME 579.68 POTENTIAL VOLUME 15.50 PROGRAM LEVEL .03 ESTIMATED LEVEL 78.94 INTELLIGENCE CONTENT 45760.71 PROGRAMMING EFFORT 21674.78 PROGRAMMING TIME 4334.96 LANGUAGE LEVEL .41 DELIVERED ERRORS 4.17 ESTIMATED ERRORS .19 .EL .HL REFERENCES .LT [Els] Elshoff, James, L., "Measuring Commercial PL/1 Programs Using Halstead's Criteria", SIGPLAN Notices, May 76, pp38-46 [Fitz Lov] Fitzsimmons, A., Love, T. "A Review and Evaluation of Software Science", Computing Surveys, Vol. 10, No. 1, March 1978, pp3-18 [Hal] Halstead, M. H., Elements of Software Science, Elsevier North Holland, Inc., Operating and Proramming Systems Series, New York, 1977 [Str] Stroud, John, M., The Fine Structure of Psychological Time, "Annals of New York Academy of Sciences", 1966, pp623-631 .el .PG .HL Appendix A: Installation and Modification .NF .RM 80 .REQUIRE "[-]read.me" :::::::::::::: release.nts :::::::::::::: 28-Feb-85 Initial release. Does not support named program libraries. The LIBRARY parameter on the command line is ignored.