6.1 Procedures, calls, and results🔗ℹ

Inputs are declared with : in the to header and referenced with : inside that procedure, including in calls it makes. as period names the result in the caller. A call using as needs an output in its procedure. Procedure names must be distinct; recursive calls are rejected. print is a teaching and display action, not a scientific observation or evidence record.

An optional about "Markdown text" belongs first inside a procedure’s opening bracket. It describes the reusable method and remains available through imports and includes. It does not run as a command or replace its input and result declarations. See Describe a workflow, stage, or procedure.

6.2 Local values and scope🔗ℹ

name = value creates a named value or updates a visible value previously created by assignment. The procedure can read its declared inputs, but does not see the caller’s local names. A nested block can read names from its parent and update a visible assigned name. A name first introduced inside do [ ... ], a loop, or a branch remains inside that block; its parent and sibling blocks cannot read it. There are no global variable declarations.

At command position, name = value is assignment. A dotted target such as trial.clock.time = "12:00" updates nested content. Inside a record or trial.[field = value], field = value names a field; inside an interface it supplies a field default. name += [field = value] updates or adds record fields; name.field += number adds to an existing numeric field. When the left side is a list, += appends one item. These updates require an existing value on the left. In a condition such as if count == 10 [ ... ], == compares values. A comparison can be passed directly to a procedure, as in do check count == 10. Parentheses remain available to group an expression.

Here attempts becomes 1; note is local to the do block. Procedure inputs, as results, and for each iteration names cannot be changed by assignment. An as result must introduce a new name in its current scope. A to procedure sees its own inputs, not values from the surrounding workflow. Stages are labels, not variable scopes.

6.3 Sequence and stage headers🔗ℹ

do [ commands ] runs a block once, in order. stage name labels following commands until another stage at the same nesting level or the block’s end. A stage is a checkpoint for the written method; it is neither a procedure nor a command to call. Stage names must be distinct across the assembled workflow, including its procedures. A procedure defined between stage headers remains available throughout its enclosing workflow.

The second stage can use plan because a stage does not create a new scope. do Review would call a procedure named Review; it does not call the stage. Stages and state explains the paired form that declares a stage’s incoming and outgoing state.

6.4 Conditions and repetition🔗ℹ

Each condition is a Boolean expression or comparison from Values and expressions. An unknown or nonboolean condition stops execution instead of choosing a branch. Square brackets delimit each command block.

repeat needs a nonnegative integer count. for each needs an ordered list; its iteration name exists only inside that iteration and cannot be updated by assignment. A repeat ... until condition may read a name introduced in its body. while and repeat ... until stop with an error after 10,000 iterations if they have not terminated.