Chapter 2

Interface Essentials

The DrScheme window has three parts: a row of buttons at the top, two editing panels in the middle, and a status line at the bottom.

The top editing panel, called the definitions window, is for defining Scheme programs. The above figure shows a program that defines the function square.

The bottom panel, called the interactions window, is for evaluating Scheme expressions interactively. The Language line in the interactions window indicates which primitives are available in the definitions and interactions windows. In the above figure, the language is Beginning Student, which is the default language.

Clicking the Run button evaluates the program in the definitions window, making the program's definitions available in the interactions window. Given the definition of square as in the figure above, typing (square 2) in the interactions window produces the result 4.

The status line at the bottom of DrScheme's window provides information about the current line and position of the editing caret, whether the current file can be modified, and whether DrScheme is currently evaluating any expression. The recycling icon flashes while DrScheme is ``recycling'' internal resources, such as memory.

2.1  Buttons

The left end of the row of buttons in DrScheme contains a miniature button with the current file's name. Clicking the button opens a menu that shows the file's full pathname. Selecting one of the menu entries produces an open-file dialog starting in the corresponding directory.

Below the filename button is a (define ...) button for a popup menu of names that are defined in the definitions window. Selecting an item from the menu moves the blinking caret to the corresponding definition.

The Save button appears whenever the definitions window is modified. Clicking the button saves the contents of the definitions window to a file. The current name of the file appears to the left of the Save button, but a file-selection dialog appears if the file has never been saved before.

The Step button -- which appears only for the How to Design Programs teaching languages ``Beginning Student'' through ``Intermediate Student with Lambda'' -- starts the Stepper, which shows the evaluation of a program as a series of small steps. Each evaluation step replaces an expression in the program with an equivalent one using the evaluation rules of DrScheme. For example, a step might replace (+ 1 2) with 3. These are the same rules used by DrScheme to evaluate a program. Clicking Step opens a new window that contains the program from the definitions window, plus several new buttons: these buttons allow navigation of the evaluation as a series of steps.

The Debug button -- which does not appear for the How to Design Programs teaching languages -- starts a more conventional stepping debugger. It runs the program in the definitions window like the Run button, but also opens a debugging panel with Pause, Continue, and Step buttons. A newly started program is paused the program's first possible pause point, and the current pause location is shown with a green arrow. Click the Continue button to continue running the program, click Step to run until the next possible pause point, and right-click on an expression's open or close parenthesis to set or remove an explicit pause. (Valid pause points are highlighted with a pink dot as you mouse over the program text.) When the program is paused, move the mouse over a variable to display its value in the debugging panel to the right of the buttons. When pausing at an expression's result, the result is shown to the left of the debugging panel's buttons, and the result can be changed by right-clicking the pause point. Click the Stop button to stop debugging so that the program in the definitions window can be edited. Debugging also stops when all expressions in the definition window have been evaluated.

Clicking the Check Syntax button annotates the program text in the definitions window. It add these annotations:

• Syntactic Highlighting Imported variables and locally defined variables are highlighted with color changes.

• Lexical Structure The lexical structure is shown with arrows overlaid on the program text. When the mouse cursor passes over a variable, DrScheme draws an arrow from the binding location to the variable, or from the binding location to every bound occurrence of the variable.

  Additionally, right-button clicking (or control-clicking under Mac OS X) on a variable activates a popup menu that lets you jump from binding location to bound location and vice versa, alpha-rename the variable, or tack the arrows so they do not disappear.

• Tail Calls Any subexpression that is (syntactically) in tail-position with respect to its enclosing context is annotated by drawing a light purple arrow from the tail expression to its surrounding expression.

• Require Annotations Right-button clicking (or control-clicking under Mac OS X) on the argument to require activates a popup menu that lets you open the file that contains the required module.

  Passing the mouse cursor over a require expression inside a module shows all of the variables that are used from that require expression. Additionally, if no variables are used from that require expression, it is colored like an unbound variable.

  Also, passing the mouse cursor over a variable that is imported from a module shows the module that it is imported from in a status line at the bottom of the frame.

The Run button evaluates the program in the definitions window and resets the interactions window.

The Break button interrupts an evaluation, or beeps if DrScheme is not evaluating anything. For example, after clicking Run or entering an expression into the interactions window, click Break to cancel the evaluation. Click the Break button once to try to interrupt the evaluation gracefully; click the button twice to kill the evaluation immediately.

2.2  The Editor

DrScheme's editor provides special support for managing parentheses in a program. When the blinking caret is next to a parenthesis, DrScheme shades the region between the parenthesis and its matching parenthesis. This feature is especially helpful when for balancing parentheses to complete an expression.

Although whitespace is not significant in Scheme, DrScheme encourages a particular format for Scheme code. When you type Enter or Return, the editor inserts a new line and automatically indents it. To make DrScheme re-indent an existing line, move the blinking caret to the line and hit the Tab key. (The caret can be anywhere in the line.) You can re-indent an entire region by selecting the region and typing Tab.

DrScheme also rewrites parenthesis as you type them, in order to make them match better. If you type a closing parenthesis ``)'', a closing square bracket ``['', or a closing curley brace ``}'', and DrScheme can match it back to some earlier opening parenthesis, bracket, or brace, DrScheme will change what you type to match. DrScheme also rewrites open square brackets, usually to an open parenthesis. There are some exceptions, where opening square brackets are not automatically changed to parentheses:

• if the square bracket is after the cond or provide/contract keywords,

• if the square bracket is after a parenthesis that is after a let, let*, letrec (or similar) keyword,

• if the square bracket is in a new expression beginning the keyword arguments,

• if the square bracket is in a sequence and the s-expression before in the sequence is a compound expression, DrScheme uses the same kind parenthesis, brace, or bracket as before, or

• if the square bracket is in the middle of string, comment, character, or symbol.

The upshot of DrScheme's trickery: always use the (presumably unshifted) square brackets on your keyboard to type parenthesis. For example, when typing this:

(define (length l)
  (cond
   [(empty? l) 0]
   [else (+ 1 (length (rest l)))]))

If you always type ``['' and ``]'' where any of the square brackets or parentheses appear, DrScheme will change the square brackets to match the code above.

Of course, these features can be disabled in the preferences dialog (see section 3.2). Also, in case DrScheme is only being too clever occasionally, holding down the control key while typing a parenthesis prevents DrScheme from changing it.

2.3  The Interactions Window

The interactions window lets you type an expression after the > prompt for immediate evaluation. You cannot modify any text before the last > prompt. To enter an expression, the blinking caret must appear after the last prompt, and also after the space following the prompt.

When you type a complete expression and hit Enter or Return, DrScheme evaluates the expression and prints the result. After printing the result, DrScheme creates a new prompt for another expression. Some expressions return a special ``void'' value; DrScheme never prints void, but instead produces a new prompt immediately.

If the expression following the current prompt is incomplete, then DrScheme will not try to evaluate it. In that case, hitting Enter or Return produces a new, auto-indented line. You can force DrScheme to evaluate the expression by typing Alt-Return or Command-Return (depending on your platform).

To copy the previous expression to the current prompt, type ESC-p (i.e., type Escape and then type p). Type ESC-p multiple times to cycle back through old expressions. Type ESC-n to cycle forward through old expressions.

Clicking the Run button evaluates the program in the definitions window and makes the program's definitions available in the interactions window. Clicking Run also resets the interactions window, erasing all old interactions and removing old definitions from the interaction environment. Although Run erases old > prompts, ESC-p and ESC-n can still retrieve old expressions.

2.4  Tabbed Editing

DrScheme's allows you to edit multiple files in a single window, via tabs. The File|New Tab creates a new tab to show a new file. Each tab has its own interactions window.

In the General sub-pane of the Editing pane in the preferences window, a checkbox labelled Open files in separate tabs causes DrScheme to open files in new tabs in the frontmost window, rather than opening a new window for the file.

The key bindings Control-Pageup and Control-Pagedown move between tabs. Under Mac OS X, Command-Shift-Left and Command-Shift-Right also move between tabs.

2.5  Errors

Whenever DrScheme encounters an error while evaluating an expression, it prints an error message in the interactions window and highlights the expression that triggered the error. The highlighted expression might be in the definitions window, or it might be after an old prompt in the interactions window.

For certain kinds of errors, DrScheme turns a portion of the error message into a hyperlink. Click the hyperlink to get help regarding a function or keyword related to the error.

2.6  Languages

DrScheme supports multiple dialects of Scheme. The name of the current evaluation language always appears in in the top of the interactions window. To choose a different language, select the Language|Choose Language... menu item. After changing the language, click Run to reset the language in the interactions window.

Five of DrScheme's languages are specifically designed for teaching:

• Beginning Student is a small version of Scheme that is tailored for beginning computer science students.

• Beginning Student with List Abbreviations is an extension to Beginning Student that prints lists with list instead of cons, and accepts quasiquoted input.

• Intermediate Student adds local bindings and higher-order functions.

• Intermediate Student with Lambda adds anonymous functions.

• Advanced Student adds mutable state.

The teaching languages different from conventional Scheme in a number of ways, described below.

DrScheme also supports several languages for experienced programmers:

• Standard (R5RS) contains those primitives and syntax defined in the R5RS Scheme standard. See the Revised5 Report on the Algorithmic Language Scheme for details.

• PLT Textual (MzScheme) extends R5RS with exceptions, threads, objects, modules, components, regular expressions, TCP support, filesystem utilities, and process control operations. See PLT MzScheme: Language Manual for details.

• PLT Graphical (MrEd) extends MzScheme with a graphical toolbox for creating GUI applications (with special support for editor applications, hence the ``Ed'' in ``MrEd''). See also PLT MrEd: Graphical Toolbox Manual.

• PLT Pretty Big extends MrEd with the forms of the Advanced Student teaching language and more.¹ It is useful as a step past Advanced Student, or for implementing MrEd programs with a richer base syntax and set of primitives.

  Note that there some forms (for example, define-struct) that appear in both Advanced and MrEd, but with slightly different semantics. In all such cases, the PLT Pretty Big language uses the forms from the MrEd language (In the case of define-struct, Advanced makes the structure transparent, whereas MrEd does not).

• module requires that the definitions window contain only a single module declaration, as defined in PLT MzScheme: Language Manual. The module explicitly declares the language for the module's body.

The Language|Choose Language... dialog contains a Show Details button for configuring certain details of the language specification. Whenever the selected options do not match the default language specification, a Custom indicator appears next to the language-selection control at the top of the dialog.

The teaching languages differ from conventional Scheme in a number of ways:

• Case-sensitive identifiers and symbols -- In a case-sensitive language, the variable names x and X are distinct, and the symbols 'x and 'X are also distinct. In a case-insensitive language, x and X are equivalent and 'x and 'X represent the same value. The teaching languages are case-sensitive by default, and other languages are usually case-insensitive. Case-sensitivity can be adjusted through the detail section of the language-selection dialog.

• All numbers are exact unless #i is specified -- In the Beginning Student through Intermediate Student with Lambda languages, numbers containing a decimal point are interpreted as exact numbers. This interpretation allows students to use familiar decimal notation without inadvertently triggering inexact arithmetic. Exact numbers with decimal representations are also printed in decimal. Inexact inputs and results are explicitly marked with #i.

• Procedures must take at least one argument -- In the Beginning Student through Intermediate Student languages, defined procedures must consume at least one argument. Since the languages have no side-effects, zero-argument functions are not useful, and rejecting such function definitions helps detect confusing syntactic mistakes.

• Identifier required at function call position -- In the Beginning Student through Intermediate Student languages, procedure calls must be of the form (identifier ...). This restriction helps detect confusing misuses of parentheses, such as (1) or ((+ 3 4)), which is a common mistake among beginners who are used to the optional parentheses of algebra.

• Top-level required at function call position -- In the Beginning Student languages, procedure calls must be of the form (top-level-identifier ...), and the number of actual arguments must match the number of formal arguments if top-level-identifier is defined. This restriction helps detect confusing misuses of parentheses, such as (x) where x is a function argument. DrScheme can detect such mistakes syntactically because Beginning Student does not support higher-order procedures.

• Primitive and defined functions allowed only in function call position -- In Beginning Student languages, the name of a primitive operator or of a defined function can be used only after the open-parenthesis of a function call (except where teachpack extensions allow otherwise, as in the convert-gui extension). Incorrect uses of primitives trigger a syntax error. Incorrect uses of defined names trigger a run-time error. DrScheme can detect such mistakes because Beginning Student does not support higher-order procedures.

• lambda allowed only in definitions -- In the Beginning Student through Intermediate Student languages, lambda (or case-lambda) may appear only in a definition, and only as the value of the defined variable.

• Free variables are not allowed -- In the Beginning Student through Advanced Student languages, every variable referenced in the definitions window must be defined, pre-defined, or the name of a local function argument.

• quote works only on symbols, quasiquote disallowed -- In the Beginning Student language, quote and ' can specify only symbols. This restriction avoids the need to explain to beginners why 1 and '1 are equivalent in standard Scheme. In addition, quasiquote, `, unquote, ,, unquote-splicing, and ,@ are disallowed.

• Unmatched cond/case is an error -- In the Beginning Student through Advanced Student languages, falling through a cond or case expression without matching a clause signals a run-time error. This convention helps detect syntactic and logical errors in programs.

• Conditional values must be true or false -- In the Beginning Student through Advanced Student languages, an expression whose value is treated as a boolean must return an actual boolean, true or false. This restriction, which applies to if, cond, and, or, nand, and nor expressions, helps detect errors where a boolean function application is omitted.

• +, *, and / take at least two arguments -- In the Beginning Student through Advanced Student languages, mathematical operators that are infix in algebra notation require at least two arguments in DrScheme. This restriction helps detect missing arguments to an operator.

• and, or, nand, and nor require at least 2 expressions -- In the Beginning Student through Advanced Student languages, the boolean combination forms require at least two sub-expressions. This restriction helps detect missing or ill-formed sub-expressions in a Boolean expression.

• set! disallowed on arguments -- In the Advanced Student language, set! cannot be used to mutate variables bound by lambda. This restriction ensures that the substitution model of function application is consistent with DrScheme's evaluation.

• Improper lists disallowed -- A proper list is either an empty list or a list created by consing onto a proper list. In the Beginning Student through Advanced Student languages, cons constructs only proper lists, signaling an error if the second argument is not a proper list. Since beginning students do not need improper lists, this restriction help detect logical errors in recursive functions.

• Dot is disallowed -- In the Beginning Student through Advanced Student languages, a delimited period is disallowed, (e.g., as an improper-list constructor in a quoted form, or for defining multi-arity procedures).

• Keywords disallowed as variable names -- In the Beginning Student through Advanced Student languages, all syntactic form names are keywords that cannot be used as variable names.

• Re-definitions are disallowed -- In the Beginning Student through Advanced Student languages, top-level names can never be re-defined.

• Function definitions are allowed only in the definitions window -- In the Beginning Student languages, function definitions are not allowed in the interactions window.

The teaching languages also deviate from traditional Scheme in printing values. Different printing formats can be selected for any language through the detail section of language-selection dialog.

• Constructor-style output -- See Constructor-style Output.

• Quasiquote-style output -- See Quasiquote-style Output.

• Rational number printing - In the teaching languages, all numbers that have a finite decimal expansion are printed in decimal form. For those numbers that do not have a finite decimal expansion (such as 4/3) DrScheme offers a choice. It either prints them as mixed fractions or as repeating decimals, where the repeating portion of the decimal expansion is shown with an overbar. In addition, DrScheme only shows the first 25 digits of the number's decimal expansion. If there are more digits, the number appears with an ellipses at the end. Click the ellipses to see the next 25 digits of the expansion.

  This setting controls only the initial display of a number. Right-clicking or control-clicking (Mac OS X) on the number lets you change from the fraction representation to the decimal representation.

• write output -- Prints values with write.

• Show sharing in values -- Prints interaction results using the shared syntax, which exposes shared structure within a value. For example, the list created by (let ([lt (list 0)]) (list lt lt)) prints as

  (shared ((-1- (list 0))) (list -1- -1-))

  instead of

  (list (list 0) (list 0)).

2.7  Executables

DrScheme's Create Executable... menu item lets you create an executable for your program that you can start without first starting DrScheme. To create an executable, first save your program to a file and set the language and teachpacks. Click Run, just to make sure that the program is working as you expect. The executable you create will not have a read-eval-print-loop, so be sure to have an expression that starts your program running in the definitions window before creating the executable.

Once you are satisfied with your program, choose the Create Executable... menu item from the Scheme menu. You will be asked to choose an executable file name or an archive file name. In the latter case, unpack the generated archive (on this machine or another one) to access the executable. In either case, you will be able to start the executable in the same way that you start any other program on your computer.

The result of Create Executable... is either a launcher executable, a stand-alone executable, or a distribution archive, and it uses either a MzScheme (textual) or MrEd (graphical) engine. For programs implemented with certain languages, Create Executable... will prompt you to choose the executable type and engine, while other languages support only one type or engine.

Each type has advantages and disadvantages:

• A launcher executable uses the latest version of your program source file when it starts. It also accesses library files from your DrScheme installation when it runs. Since a launcher executable contains specific paths to access those files, launchers usually cannot be moved from one machine to another.

• A stand-alone executable embeds a compiled copy of your program and any Scheme libraries that your program uses. When the executable is started, it uses the embedded copies and does not need your original source file. It may, however, access your DrScheme installation for DLLs, frameworks, shared libraries, or helper executables. Consequently, a stand-alone executable usually cannot be moved from one machine to another.

• A distribution archive packages a stand-alone executable together with any needed DLLs, frameworks, shared libraries, and helper executables. A distribution archive can be unpacked and run on any machine with the same operating system as yours.

TIP: Disable debugging in the language dialog before creating your executable. With debugging enabled, you will see a stack trace with error messages, but your program will run more slowly. To disable debugging, open the language dialog, click the Show Details button, and select No debugging or profiling, if it is available.

2.8  Printed Results

This section describes the different formats that DrScheme uses for printing results in the interactions window. Each of the different settings here also apply to the print primitive. That is, printing in the interactions window is identical to output produced by the print primitive.

2.8.1  Constructor-style Output

DrScheme's constructor-style output treats cons, vector, and similar primitives as value constructors, rather than functions. It also treats list as shorthand for multiple cons's ending with the empty list. Constructor-style printing is valuable for beginning computer science students, because output values look the same as input values.

Results printed in DrScheme's interactions window using constructor-style printing look different than results printed in traditional Scheme implementations, which use write to print results. The table in Figure 1 shows the differences between values printed in constructor style and values printed with write.

Figure 1:  Comparison of constructor-style output to write

2.8.2  Quasiquote-style Output

Constructor-style output is inconvenient for printing S-expression results that represent programs. For example, the value '(lambda (x) (lambda (y) (+ x y))) prints as

DrScheme's quasiquote-style output combines the input-output invariance of constructor-style printing with the S-expression readability of write. It uses quasiquote to print lists, and uses unquote to escape back to constructor style printing for non-lists and non-symbols.

With quasiquote-style printing, the above example prints as:

This example:

2.9  Input and Output

Many Scheme programs avoid explicit input and output operations, obtaining input via direct function calls in the interactions window, and producing output by returning values. Other Scheme programs explicitly print output for the user during evaluation using write or display, or explicitly request input from the user using read or read-char.

Explicit input and output appear in the interactions window, but within special boxes that separate explicit I/O from normal expressions and results. For example, evaluating

in the interactions window produces a special box for entering input:

(The underscore indicates the location of the blinking caret.) Type a number into the box and hit Enter, and that number becomes the result of the (read) expression. Once text is submitted for an input box, it is moved outside the input box, and when DrScheme shows a new prompt, it hides the interaction box. Thus, if you type 5 in the above input box and hit Return, the result appears as follows:

In this case, the first 5 is the input, and the second 5 is the result of the (read) expression. The second 5 is colored blue, as usual for a result printed by DrScheme.

Output goes into the interaction window directly. If you run the program

(define v (read))
(display v)
v

and provide the input S-expression (1 2), the interactions window ultimately appears as follows:

In this example, display produces output immediately beneath the input you typed, and the final result is printed last. The displayed output is drawn in purple. (The above example assumes constructor-style printing. With traditional value printing, the final line outside the box would be (1 2).)

Entering the same program line-by-line in the interactions window produces a different-looking result:

Depending on the input operation, you may enter more text into an input box than is consumed. In that case, the leftover text remains in the input stream for later reads. For example, in the following interaction, two values are provided in response to the first (read), so the second value is returned immediately for the second (read):

The following example illustrates that submitting input with Return inserts a newline character into the input stream:

2.10  XML

DrScheme has special support for XML concrete syntax. The Special menu's Insert XML Box menu inserts an embedded editor into your program. In that embedded editor, you type XML's concrete syntax. When a program containing an XML box is evaluated, the XML box is translated into an x-expression (or xexpr). Xexprs are s-expression representation for XML expressions. Each xexpr is a list whose first element is a symbol naming the tag, second element is an association list representing attributes and remaining elements are the nested XML expressions.

XML boxes have two modes for handling whitespace. In one mode, all whitespace is left intact in the resulting xexpr. In the other mode, any tag that only contains nested XML expressions and whitespace has the whitespace removed. You can toggle between these modes by right-clicking or control-clicking (Mac OS X) on the top portion of the XML box.

In addition to containing XML text, XML boxes can also contain Scheme boxes. Scheme boxes contain Scheme expressions. These expressions are evaluated and their contents are placed into the containing XML box's xexpr. There are two varieties of Scheme box: the standard Scheme box and the splicing Scheme box. The standard Scheme box inserts its value into the containing xexpr. The contents of the splice box must evaluate to a list and the elements of the list are ``flattened'' into the containing xexpr. Right-clicking or control-clicking (Mac OS X) on the top of a Scheme box opens a menu to toggle the box between a Scheme box and a Scheme splice box.

2.11  Test Cases

DrScheme also includes support for creating test cases as part of the program text. This support is designed as an aid for students building tests as part of the How to Design Programs design recipes.

Test cases in DrScheme are written in special boxes. To create one, choose Insert Test Case from the Scheme menu in the DrScheme window. The test cases consists of three editable areas. From the top, the first is a comment that names the test case. The second is an expression to test. The third is the expected result. Below the expected result is a light-blue box that will contain the actual result of the expression to be tested. Clicking on the triangle in the top-right hides the expression to test and the expected and actual results.

To run test cases, simply click Run. The top-right corner of the test case will change to either be a check box or a red X, indicating success or failure of the test case.

To disable all of the test cases in the definitions window, choose the Disable All Test Cases menu item in the Scheme menu.