-ARTICLE: "cleave-shuffle-equivalence" "Expressing shuffle words with cleave combinators"
-"Cleave combinators are defined in terms of shuffle words, and mappings from certain shuffle idioms to cleave combinators are discussed in the documentation for " { $link bi } ", " { $link 2bi } ", " { $link 3bi } ", " { $link tri } ", " { $link 2tri } " and " { $link 3tri } "."
-$nl
-"Certain shuffle words can also be expressed in terms of the cleave combinators. Internalizing such identities can help with understanding and writing code using cleave combinators:"
-"The cleave combinators apply multiple quotations to a single value."
-$nl
-"Two quotations:"
-{ $subsection bi }
-{ $subsection 2bi }
-{ $subsection 3bi }
-"Three quotations:"
-{ $subsection tri }
-{ $subsection 2tri }
-{ $subsection 3tri }
-"Technically, the cleave combinators are redundant because they can be simulated using shuffle words and other combinators, and in addition, they do not reduce token counts by much, if at all. However, they can make code more readable by expressing intention and exploiting any inherent symmetry. For example, a piece of code which performs three operations on the top of the stack can be written in one of two ways:"
-{ $code
- "! First alternative; uses keep"
- "[ 1 + ] keep"
- "[ 1 - ] keep"
- "2 *"
- "! Second alternative: uses tri"
- "[ 1 + ]"
- "[ 1 - ]"
- "[ 2 * ] tri"
-}
-"The latter is more aesthetically pleasing than the former."
-$nl
-"A generalization of the above combinators to any number of quotations can be found in " { $link "combinators" } "."
-{ $subsection "cleave-shuffle-equivalence" } ;
-
-ARTICLE: "spread-shuffle-equivalence" "Expressing shuffle words with spread combinators"
-"Spread combinators are defined in terms of shuffle words, and mappings from certain shuffle idioms to spread combinators are discussed in the documentation for " { $link bi* } ", " { $link 2bi* } ", " { $link tri* } ", and " { $link 2tri* } "."
-$nl
-"Certain shuffle words can also be expressed in terms of the spread combinators. Internalizing such identities can help with understanding and writing code using spread combinators:"
-"The spread combinators apply multiple quotations to multiple values. The " { $snippet "*" } " suffix signifies spreading."
-$nl
-"Two quotations:"
-{ $subsection bi* }
-{ $subsection 2bi* }
-"Three quotations:"
-{ $subsection tri* }
-{ $subsection 2tri* }
-"Technically, the spread combinators are redundant because they can be simulated using shuffle words and other combinators, and in addition, they do not reduce token counts by much, if at all. However, they can make code more readable by expressing intention and exploiting any inherent symmetry. For example, a piece of code which performs three operations on three related values can be written in one of two ways:"
-{ $code
- "! First alternative; uses dip"
- "[ [ 1 + ] dip 1 - ] dip 2 *"
- "! Second alternative: uses tri*"
- "[ 1 + ] [ 1 - ] [ 2 * ] tri*"
-}
-"A generalization of the above combinators to any number of quotations can be found in " { $link "combinators" } "."
-{ $subsection "spread-shuffle-equivalence" } ;
-
-ARTICLE: "apply-combinators" "Apply combinators"
-"The apply combinators apply a single quotation to multiple values. The " { $snippet "@" } " suffix signifies application."
-$nl
-"Two quotations:"
-{ $subsection bi@ }
-{ $subsection 2bi@ }
-"Three quotations:"
-{ $subsection tri@ }
-{ $subsection 2tri@ }
-"A pair of utility words built from " { $link bi@ } ":"
-"Sometimes an additional storage area is needed to hold objects. The " { $emphasis "retain stack" } " is an auxilliary stack for this purpose. Objects can be moved between the data and retain stacks using a set of combinators."
-$nl
-"The dip combinators invoke the quotation at the top of the stack, hiding the values underneath:"
-{ $subsection dip }
-{ $subsection 2dip }
-{ $subsection 3dip }
-{ $subsection 4dip }
-"The slip combinators invoke a quotation further down on the stack. They are most useful for implementing other combinators:"
-{ $subsection slip }
-{ $subsection 2slip }
-{ $subsection 3slip }
-"The keep combinators invoke a quotation which takes a number of values off the stack, and then they restore those values:"
-"if we wanted to abstract out the message into a parameter, we could keep it on the stack between iterations:"
-{ $code ": print-10 ( message -- ) 10 [ dup print ] times drop ;" }
-"However, keeping loop-invariant values on the stack doesn't always work out nicely. For example, a word to subtract a value from each element of a sequence:"
-{ $code ": subtract-n ( seq n -- seq' ) swap [ over - ] map nip ;" }
-"Three shuffle words are required to pass the value around. Instead, the loop-invariant value can be partially applied to a quotation using " { $link curry } ", yielding a new quotation that is passed to " { $link map } ":"
-"Now consider the word that is dual to the one above; instead of subtracting " { $snippet "n" } " from each stack element, it subtracts each element from " { $snippet "n" } "."
-$nl
-"One way to write this is with a pair of " { $link swap } "s:"
-"Certain combinators transform quotations to produce a new quotation."
-{ $subsection "compositional-examples" }
-"Fundamental operations:"
-{ $subsection curry }
-{ $subsection compose }
-"Derived operations:"
-{ $subsection 2curry }
-{ $subsection 3curry }
-{ $subsection with }
-{ $subsection prepose }
-"These operations run in constant time, and in many cases are optimized out altogether by the " { $link "compiler" } ". " { $link "fry" } " are an abstraction built on top of these operations, and code that uses this abstraction is often clearer than direct calls to the below words."
-$nl
-"Curried dataflow combinators can be used to build more complex dataflow by combining cleave, spread and apply patterns in various ways."
-{ $subsection "curried-dataflow" }
-"Quotations also implement the sequence protocol, and can be manipulated with sequence words; see " { $link "quotations" } ". However, such runtime quotation manipulation will not be optimized by the optimizing compiler." ;
-"The following pair of words invoke words and quotations reflectively:"
-{ $subsection call }
-{ $subsection execute }
-"These words are used to implement combinators. Note that combinator definitions must be followed by the " { $link POSTPONE: inline } " declaration in order to compile in the optimizing compiler; for example:"
-{ $code
- ": keep ( x quot -- x )"
- " over [ call ] dip ; inline"
-}
-"Word inlining is documented in " { $link "declarations" } "." ;
-
-ARTICLE: "booleans" "Booleans"
-"In Factor, any object that is not " { $link f } " has a true value, and " { $link f } " has a false value. The " { $link t } " object is the canonical true value."
-{ $subsection f }
-{ $subsection t }
-"The " { $link f } " object is the unique instance of the " { $link f } " class; the two are distinct objects. The latter is also a parsing word which adds the " { $link f } " object to the parse tree at parse time. To refer to the class itself you must use " { $link POSTPONE: POSTPONE: } " or " { $link POSTPONE: \ } " to prevent the parsing word from executing."
-$nl
-"Here is the " { $link f } " object:"
-{ $example "f ." "f" }
-"Here is the " { $link f } " class:"
-{ $example "\\ f ." "POSTPONE: f" }
-"They are not equal:"
-{ $example "f \\ f = ." "f" }
-"Here is an array containing the " { $link f } " object:"
-{ $example "{ f } ." "{ f }" }
-"Here is an array containing the " { $link f } " class:"
-{ $example "{ POSTPONE: f } ." "{ POSTPONE: f }" }
-"The " { $link f } " object is an instance of the " { $link f } " class:"
-{ $example "USE: classes" "f class ." "POSTPONE: f" }
-"The " { $link f } " class is an instance of " { $link word } ":"
-{ $example "USE: classes" "\\ f class ." "word" }
-"On the other hand, " { $link t } " is just a word, and there is no class which it is a unique instance of."
-{ $example "t \\ t eq? ." "t" }
-"Many words which search collections confuse the case of no element being present with an element being found equal to " { $link f } ". If this distinction is imporant, there is usually an alternative word which can be used; for example, compare " { $link at } " with " { $link at* } "." ;
-
-ARTICLE: "conditionals-boolean-equivalence" "Expressing conditionals with boolean logic"
-"Certain simple conditional forms can be expressed in a simpler manner using boolean logic."
-$nl
-"The following two lines are equivalent:"
-{ $code "[ drop f ] unless" "swap and" }
-"The following two lines are equivalent:"
-{ $code "[ ] [ ] ?if" "swap or" }
-"The following two lines are equivalent, where " { $snippet "L" } " is a literal:"
-{ $code "[ L ] unless*" "L or" } ;
-
-ARTICLE: "conditionals" "Conditionals and logic"
-"The basic conditionals:"
-{ $subsection if }
-{ $subsection when }
-{ $subsection unless }
-"Forms abstracting a common stack shuffle pattern:"
-{ $subsection if* }
-{ $subsection when* }
-{ $subsection unless* }
-"Another form abstracting a common stack shuffle pattern:"
-{ $subsection ?if }
-"Sometimes instead of branching, you just need to pick one of two values:"