In the early days of computing, many languages such as
FORTRAN IV and
BASIC lacked block syntax except sometimes for rudimentary loop constructs. A conditional was encoded using the
goto statement. In the following FORTRAN 66 code, the logical structure is less than clear due to the lack of blocks. C INITIALIZE VALUES TO BE CALCULATED PAYSTX = .FALSE. PAYSST = .FALSE. TAX = 0.0 SUPTAX = 0.0 C SKIP TAX DEDUCTION IF EMPLOYEE EARNS LESS THAN TAX THRESHOLD IF (WAGES .LE. TAXTHR) GOTO 100 PAYSTX = .TRUE. TAX = (WAGES - TAXTHR) * BASCRT C SKIP SUPERTAX DEDUCTION IF EMPLOYEE EARNS LESS THAN SUPERTAX THRESHOLD IF (WAGES .LE. SUPTHR) GOTO 100 PAYSST = .TRUE. SUPTAX = (WAGES - SUPTHR) * SUPRAT 100 TAXED = WAGES - TAX - SUPTAX Blocks allow the programmer to treat a group of statements as a unit, and the default values which had to appear in initialization in this style of programming can, with a block structure, be placed closer to the decision. The following code in Jensen and Wirth Pascal shows that block structure makes it easier to see how the code could be refactored for clarity, and also makes it easier to do, because the structure of the inner conditional can easily be moved out of the outer conditional altogether and the effects of doing so are easily predicted. Use of blocks in the fragment below clarifies the programmer's intent, and enables combining the resulting blocks into a nested hierarchy of
conditional statements. The structure of the code reflects the programmer's thinking more closely, making it easier to understand and modify. The code can be made even clearer by taking the inner if statement out of the outer one altogether, placing the two blocks one after the other to be executed consecutively. Semantically, there is little difference in this case, and the use of block structure, supported by indenting for readability, makes it easy for the programmer to refactor the code. if wages > tax_threshold then begin paystax := true; tax := (wages - tax_threshold) * tax_rate if wages > supertax_threshold then begin pays_supertax := true; supertax := (wages - supertax_threshold) * supertax_rate end else begin pays_supertax := false; supertax := 0 end end else begin paystax := false; pays_supertax := false; tax := 0; supertax := 0 end; taxed := wages - tax - supertax; In primitive languages, variables had broad scope. For instance, an integer variable called IEMPNO might be used in one part of a Fortran subroutine to denote an employee social security number (SSN), but during maintenance work on the same subroutine, a programmer might accidentally use the same variable, IEMPNO, for a different purpose, and this could result in a bug that was difficult to trace. Block structure makes it easier for programmers to control scope. In the following R5RS Standard
Scheme fragment, empno is used to identify both the manager and their underlings each by their respective SSN, but because the underling SSN is declared within an inner block it does not interact with the variable of the same name that contains the manager's SSN. In practice, considerations of clarity would probably lead the programmer to choose distinct variable names, but they have the choice and it is more difficult to introduce a bug inadvertently. Within the lambda expression, the variable empno refers to the SSN of an underling. The variable empno in the outer expression, referring to the manager's SSN, is shadowed. (let ((empno (ssn-of employee-name))) (while (is-manager empno) (let ((employees (length (underlings-of empno)))) (printf "~a has ~a employees working under him:~%" employee-name employees) (for-each (lambda (empno) (printf "Name: ~a, role: ~a~%" (name-of empno) (role-of empno))) (underlings-of empno))))) ==Hoisting==