Schism is a small, embedded, Scheme-like Lisp. Schism is strongly, dynamically typed, uses lexical scoping, evaluates eagerly and tends to avoid side effects. It's a Lisp-1. The two extant implementations (both to be found here) are interpreters. Schism has been designed to implement small pieces of business logic that need to be available and in-sync both client-side and server-side. Ideologically, Schism is unashamedly functional, expression-based, and largely single-assignment. Being a DSL, it does not have an I/O system. There are no constructs for imperative programming, nor any support for object-oriented programming. JavaScript-style object-based system would be easy to implement on top of it, however. Schism is Scheme-like, but it's not a Scheme. It's not a recursive acronym either. Schism depends on Picoparsec: https://github.com/pbl64k/Picoparsec See REPL demo at: http://ranselett.net/Schism/ (Works only in Chrome.) DIFFERENCES FROM SCHEME 1. Perhaps most conspicuously, Schism doesn't have macros of any sort, nor any other facilities to operate on its own AST. In fact, the internal representation is incompatible with Schism's lists, so as it is, Schism is not homoiconic. 2. Schism doesn't have quoting or quasiquoting. 3. Schism does not perform tail call elimination. 4. Schism doesn't have first-class continuations (call/cc in particular). 5. Schism's type system is simpler than Scheme's. Schism has six primitive types: boolean, numeric (aka number - and there's no numeric tower at all), string (a char is just a string that has length one), nil (aka empty or null), pair (a cons cell) and lambda (aka procedure). There's no support for vectors. Schism doesn't have symbols as first-class citizens, nor does it have anything like ports. 6. Schism only has seven special forms: lambda, define, letrec, if, cond, and & or. 7. Schism doesn't have named let or lambda forms. Use define, letrec or fix instead. 8. All values are immutable. Limited destructive assignment is supported (or rather, being treated as a lesser evil) at top level through the use of define. set! and its ilk are not included. 9. Schism doesn't support variadic functions, whether named or anonymous. Host functions exposed through the environment may be variadic. 10. Schism lacks many standard library functions, has no equality/equivalence hierarchy, doesn't have literals #t and #f, and does not support dot syntax for improper lists. CAVEATS 1. Schism hasn't been designed for performance. It shows. Both implementations are interpreters running on top of interpreters. Parser library was chosen for familiarity and simplicity rather than performance. In particular, the PHP version is atrociously slow at parse stage. In production it is essential to cache parsed ASTs to avoid the overhead. It should be easy enough to plug in an alternate parser (e.g., a straightforward S-expression parser with some postprocessing) if one so desires. Additionally, the handling of environments is highly inefficient, which is going to show on anything but the small, straightforward pieces of business logic it's designed to run. There's no static validation - most of it happens at run-time, which further affects performance. Schism is subject to host platform's limits on stack depth, which doesn't work well in absence of TCO (or imperative language constructs). 2. Diagnostics are extremely limited. Well, pretty much non-existent. Starting with the fact that line/character information is not preserved by the parser, and ending with the general observation that run-time errors are not very informative. 3. The parser is not very robust and at times idiosyncratic. (For example, while the parser is generally lenient on whitespace, the names of special forms must be followed by at least one whitespace character, otherwise the parser, being largely non-backtracking, would be unable to accept "orientation" as a valid symbol.) Numeric literals interfere with naming identifiers - you cannot define 1+ in current Schism implementations. Other quirks that I'm not aware of might very well be present. 4. While Schism is supposed to have the same evaluation semantics in both implementations, there might be differences in practice. Math is not isolated enough from the underlying platform to guarantee exact same behavior. Furthermore, Schism is not mature enough to guarantee there are no overlooked differences between the two implementations. You may want to check out the standard library and test suite (schism-stdlib.scm and schism-tests.scm correspondingly) for code samples. The best use I've found for Schism so far is as a replacement for a programmable calculator (as Schism/JS REPL works just fine in Android Chrome). LANGUAGE "REFERENCE" 1. Special forms (comment EXPR ...) (lambda [] EXPR-BODY) (lambda [ARG1 ...] EXPR-BODY) (define NAME EXPR-VAL) (define [NAME] EXPR-BODY) (define [NAME ARG1 ...] EXPR-BODY) (letrec ([NAME EXPR-VAL] ...) EXPR-VAL) (if EXPR-PRED EXPR-VAL EXPR-VAL) (cond [EXPR-PRED EXPR-VAL] ...) (and ) (and EXPR1 ...) (or ) (or EXPR1 ...) 2. Primitive functions (and values) true false nil (sys\raise-error STR) (boolean? ANY) (numeric? ANY) (string? ANY) (procedure? ANY) (nil? ANY) (cons? ANY) (equal? ANY) (repr ANY) (not BOOL) (+) (+ NUM1 ...) (- NUM) (- NUM1 NUM2) (*) (* NUM1 ...) (/ NUM) (/ NUM1 NUM2) (< NUM1 NUM2) (<= NUM1 NUM2) (= NUM1 NUM2) (/= NUM1 NUM2) (> NUM1 NUM2) (>= NUM1 NUM2) (floor NUM) (ceiling NUM) (round NUM) (expt NUM1 NUM2) (sin NUM) (cos NUM) (tan NUM) (asin NUM) (acos NUM) (atan NUM) (log NUM) (number->string NUM) (integer->string NUM) (string<? STR1 STR2) (string-ci<? STR1 STR2) (string<=? STR1 STR2) (string-ci<=? STR1 STR2) (string=? STR1 STR2) (string-ci=? STR1 STR2) (string/=? STR1 STR2) (string-ci/=? STR1 STR2) (string>? STR1 STR2) (string-ci>? STR1 STR2) (string>=? STR1 STR2) (string-ci>=? STR1 STR2) (string->number STR) (string->integer STR) (string-length STR) (string-append) (string-append STR1 ...) (substring STR NUM1 NUM2) (cons ANY1 ANY2) (car CONS) (cdr CONS) (list) (list ANY1 ...) 3. Library functions (and values) PI E empty UPPERCASE-LETTER-CHARACTERS LOWERCASE-LETTER-CHARACTERS LOWER-TO-UPPER-LETTER-MAPPING UPPER-TO-LOWER-LETETR-MAPPING LETTER-CHARACTERS NUMERIC-CHARACTERS WHITESPACE-CHARACTERS (assert BOOL) (assert-equals ANY1 ANY2) (empty? ANY) (null? ANY) empty? and null? are aliases of nil? (pair? ANY) pair? is an alias of cons? (list? ANY) (lambda? ANY) lambda? is an alias of procedure (eq? ANY1 ANY2) (eqv? ANY1 ANY2) eq? and eqv? are aliases of equal? (number? ANY) number? is an alias of numeric? (== NUM1 NUM2) == is an alias of = (!= NUM1 NUM2) != is an alias of /= (zero? NUM) (positive? NUM) (negative? NUM) (abs NUM) (sign NUM) (max NUM1 NUM2) (min NUM1 NUM2) (truncate NUM) (quotient NUM1 NUM2) (remainder NUM1 NUM2) (floored-quotient NUM1 NUM2) (modulo NUM1 NUM2) (gcd NUM1 NUM2) (lcm NUM1 NUM2) (sqrt NUM) (even? NUM) (odd? NUM) (exp NUM) (succ NUM) (pred NUM) (char? ANY) (char<? CHAR1 CHAR2) (char-ci<? CHAR1 CHAR2) (char<=? CHAR1 CHAR2) (char-ci<=? CHAR1 CHAR2) (char=? CHAR1 CHAR2) (char-ci=? CHAR1 CHAR2) (char/=? CHAR1 CHAR2) (char-ci/=? CHAR1 CHAR2) (char>? CHAR1 CHAR2) (char-ci>? CHAR1 CHAR2) (char>=? CHAR1 CHAR2) (char-ci>=? CHAR1 CHAR2) (char->integer CHAR) (integer->char NUM) integer->char is an alias of integer->string (id ANY) (const ANY) (flip LAMBDA) (dot LAMBDA1 LAMBDA2) (subst LAMBDA1 LAMBDA2) (fix LAMBDA) (pair ANY1 ANY2) pair is an alias of cons (first CONS) first is an alias of car (rest CONS) rest is an alias of cdr (fold-left LAMBDA ANY LIST) (foldl LAMBDA ANY LIST) foldl is an alias of fold-left (fold-right LAMBDA ANY LIST) (foldr LAMBDA ANY LIST) foldr is an alias of fold-right (unfold-left LAMBDA ANY) (unfoldl LAMBDA ANY) (unfold LAMBDA ANY) unfoldl and unfold are aliases of unfold-left (map LAMBDA LIST) (reverse LIST) (join LIST) (bind LAMBDA LIST) (filter LAMBDA LIST) (all? LAMBDA LIST) (for-all? LIST LAMBDA) (any? LAMBDA LIST) (there-exists? LIST LAMBDA) (length LIST) (zip LIST1 LIST2) (member ANY LIST) (memq ANY LIST) (memv ANY LIST) memq and memv are aliases of member (element-of ANY LIST) (append LIST1 LIST2) (ith LIST NUM) (list-ref LIST NUM) list-ref is an alias of ith (last LIST) (init LIST) (take NUM LIST) (list-head LIST NUM) (drop NUM LIST) (list-tail LIST NUM) (sort-with LAMBDA LIST) (merge-with LAMBDA LIST1 LIST2) (merge-sort-with LAMBDA LIST) (quick-sort LIST LAMBDA) (merge-sort LIST LAMBDA) (sort LIST LAMBDA) sort is an alias of merge-sort (assoc ANY LIST) (assq ANY LIST) (assv ANY LIST) assq and assv are aliases of assoc (range-step NUM1 NUM2 NUM3) (range NUM1 NUM2) (make-list NUM1 ANY) (list->string LIST) (string->list STR) (string-null? STR) (string-ref STR NUM) (string-head STR NUM) (string-tail STR NUM) (char-alphabetic? CHAR) (char-numeric? CHAR) (char-whitespace? CHAR) (char-upper-case? CHAR) (char-lower-case? CHAR) (char-upcase CHAR) (char-downcase CHAR) (integrate LAMBDA LIST LIST) (integrate-rk LAMBDA LIST LIST) (map\has? MAP ANY) (map\elt-already-exists ANY) (map\elt-not-found ANY) (map\get MAP ANY) (map\update-without-checking MAP ANY ANY) (map\update MAP ANY ANY) (map\delete-without-checking MAP ANY) (map\delete MAP ANY) (map\insert-without-checking MAP ANY ANY) (map\insert MAP ANY ANY) (map\insert-or-update MAP ANY ANY) (map\keys MAP) (map\values MAP)