path: root/README

This directory contains the distribution of Scheme Library slib2c0.
Slib conforms to Revised^4 Report on the Algorithmic Language Scheme
and the IEEE P1178 specification.  Slib supports Unix and similar
systems, VMS, and MS-DOS.

The maintainer can be reached at jaffer@ai.mit.edu.

			       MANIFEST

  `README' is this file.  It contains a MANIFEST, INSTALLATION
	INSTRUCTIONS, and proposed coding standards.
  `FAQ' Frequently Asked Questions and answers.
  `ChangeLog' documents changes to slib.
  `slib.texi' has documentation on library packages in TexInfo format.

  `Template.scm' Example configuration file.  Copy and customize to
	reflect your system.
  `chez.init' is a configuration file for Chez Scheme.
  `elk.init' is a configuration file for ELK 2.1
  `gambit.init' is a configuration file for Gambit Scheme.
  `macscheme.init' is a configuration file for MacScheme.
  `mitscheme.init' is a configuration file for MIT Scheme.
  `mitcomp.pat' is a patch file which adds definitions to SLIB files
	for the MitScheme compiler.
  `scheme2c.init' is a configuration file for DEC's scheme->c.
  `scheme48.init' is a configuration file for Scheme48.
  `scsh.init' is a configuration file for Scheme-Shell
  `scm.init' is a configuration file for SCM.
  `t3.init' is a configuration file for T3.1 in Scheme mode.
  `vscm.init' is a configuration file for VSCM.
  `mklibcat.scm' builds the *catalog* cache.
  `require.scm' has code which allows system independent access to
	the library files.

  `format.scm' has Common-Lisp style format.
  `formatst.scm' has code to test format.scm
  `pp.scm' has pretty-print.
  `ppfile.scm' has pprint-file and pprint-filter-file.
  `obj2str.scm' has object->string.
  `strcase.scm' has functions for manipulating the case of strings.
  `genwrite.scm' has a generic-write which is used by pp.scm,
	pp2str.scm and obj2str.scm
  `printf.scm' has printf, fprintf, and sprintf compatible with C.
  `scanf.scm' has scanf, fscanf, and sscanf compatible by C.
  `lineio' has line oriented input/output functions.
  `qp.scm' has printer safe for circular structures.
  `break.scm' has break and continue.
  `trace.scm' has trace and untrace for tracing function execution.
  `debug.scm' has handy higher level debugging aids.
  `strport.scm' has routines for string-ports.
  `strsrch.scm' search for chars or substrings in strings and ports.

  `alist.scm' has functions accessing and modifying association lists.
  `hash.scm' defines hash, hashq, and hashv.
  `hashtab.scm' has hash tables.
  `sierpinski.scm' 2-dimensional coordinate hash.
  `soundex.scm' English name hash.
  `logical.scm' emulates 2's complement logical operations.
  `random.scm' has random number generator compatible with Common Lisp.
  `randinex.scm' has inexact real number distributions.
  `primes.scm' has primes and probably-prime?.
  `factor.scm' has factor.
  `root.scm' has Newton's and Laguerre's methods for finding roots.
  `cring.scm' extend + and * to custom commutative rings.
  `selfset.scm' sets single letter identifiers to their symbols.
  `determ.scm' compute determinant of list of lists.
  `charplot.scm' has procedure for plotting on character screens.
  `plottest.scm' has code to test charplot.scm.
  `tek40.scm' has routines for Tektronix 4000 series graphics.
  `tek41.scm' has routines for Tektronix 4100 series graphics.
  `getopt.scm' has posix-like getopt for parsing command line arguments. 
  `psxtime.scm' has Posix time conversion routines.
  `cltime.scm' has Common-Lisp time conversion routines.
  `timezone.scm' has the default time-zone, UTC.
  `tzfile.scm' reads sysV style (binary) timezone file.
  `comparse.scm' has shell-like command parsing.

  `rdms.scm' has code to construct a relational database from a base
	table implementation.
  `alistab.scm' has association list base tables.
  `dbutil.scm' has utilities for creating and manipulating relational
	databases.
  `dbrowse.scm' browses relational databases.
  `paramlst.scm' has procedures for passing parameters by name.
  `report.scm' prints database reports.
  `batch.scm' Group and execute commands on various operating systems.
  `makcrc.scm' Create Scheme procedure to calculate POSIX.2 checksums
	or other CRCs.

  `record.scm' a MITScheme user-definable datatypes package
  `promise.scm' has code from R4RS for supporting DELAY and FORCE.

  `repl.scm' has a read-eval-print-loop.
  `defmacex.scm' has defmacro:expand*.
  `mbe.scm' has "Macro by Example" define-syntax.
  `scmacro.scm' is a syntactic closure R4RS macro package.
	r4rsyn.scm, synclo.scm, synrul.scm have syntax definitions
	and support.
  `scmactst.scm' is code for testing SYNTACTIC CLOSURE macros.
  `scainit.scm' is a syntax-case R4RS macro package.
	scaglob.scm scamacr.scm scaoutp.scm scaexpp.scm have
	syntax definitions and support.  `syncase.sh' is a shell
	script for producing the SLIB version from the original.
  `macwork.scm' is a "Macros that work" package.
	mwexpand.scm mwdenote.scm mwsynrul.scm have support.
  `macrotst.scm' is code from R4RS for testing macros.

  `values.scm' is multiple values.
  `queue.scm' has queues and stacks.

  `yasos.scm' is object oriented programming (using R4RS macros).
  `collect.scm' is collection operators (like CL sequences).
  `priorque.scm' has code and documentation for priority queues.
  `wttree.scm' has weight-balanced trees.
  `wttest.scm' tests weight-balanced trees.
  `process.scm' has multi-processing primitives.
  `array.scm' has multi-dimensional arrays and sub-arrays.
  `arraymap.scm' has array-map!, array-for-each, and array-indexes.

  `sort.scm' has sorted?, sort, sort!, merge, and merge!.
  `tsort.scm' has topological-sort.
  `comlist.scm' has many common list and mapping procedures.
  `tree.scm' has functions dealing with trees.
  `chap.scm' has functions which compare and create strings in
	"chapter order".

  `sc4opt.scm' has optional rev4 procedures.
  `sc4sc3.scm' has procedures to make a rev3 implementation run rev4
	code. 
  `sc2.scm' has rev2 procedures eliminated in subsequent versions.
  `mularg.scm' redefines - and / to take more than 2 arguments.
  `mulapply.scm' redefines apply to take more than 2 arguments.
  `ratize.scm' has function rationalize from Revised^4 spec.
  `trnscrpt.scm' has transcript-on and transcript-off from Revised^4 spec.
  `withfile.scm' has with-input-from-file and with-output-to-file from R4RS.
  `dynwind.scm' has proposed dynamic-wind from R5RS.
  `dwindtst.scm' has routines for characterizing dynamic-wind.
  `dynamic.scm' has proposed DYNAMIC data type.
  `fluidlet.scm' has fluid-let syntax.
  `struct.scm' has defmacros which implement RECORDS from the book:
	"Essentials of Programming Languages".
  `structure.scm' has syntax-case macros for the same.
  `structst.scm' has test code for struct.scm.
  `byte.scm' has arrays of small integers.

		      INSTALLATION INSTRUCTIONS

  Check the manifest in `README' to find a configuration file for your
Scheme implementation.  Initialization files for most IEEE P1178
compliant Scheme Implementations are included with this distribution.

  If the Scheme implementation supports `getenv', then the value of the
shell environment variable SCHEME_LIBRARY_PATH will be used for
`(library-vicinity)' if it is defined.  Currently, Chez, Elk,
MITScheme, scheme->c, VSCM, and SCM support `getenv'.  Scheme48
supports `getenv' but does not use it for determining
`library-vicinity'.  (That is done from the Makefile.)

  You should check the definitions of `software-type',
`scheme-implementation-version', `implementation-vicinity', and
`library-vicinity' in the initialization file.  There are comments in
the file for how to configure it.

  Once this is done you can modify the startup file for your Scheme
implementation to `load' this initialization file.  SLIB is then
installed.

  Multiple implementations of Scheme can all use the same SLIB
directory.  Simply configure each implementation's initialization file
as outlined above.

  The SCM implementation does not require any initialization file as
SLIB support is already built in to SCM.  See the documentation with
SCM for installation instructions.

  SLIB includes methods to create heap images for the VSCM and Scheme48
implementations.  The instructions for creating a VSCM image are in
comments in `vscm.init'.  To make a Scheme48 image for an installation
under `<prefix>', `cd' to the SLIB directory and type `make
prefix=<prefix> slib48'.  To install the image, type `make
prefix=<prefix> install48'.  This will also create a shell script with
the name `slib48' which will invoke the saved image.

			 PORTING INSTRUCTIONS

  If there is no initialization file for your Scheme implementation, you
will have to create one.  Your Scheme implementation must be largely
compliant with `IEEE Std 1178-1990' or `Revised^4 Report on the
Algorithmic Language Scheme' to support SLIB.

  `Template.scm' is an example configuration file.  The comments inside
will direct you on how to customize it to reflect your system.  Give
your new initialization file the implementation's name with `.init'
appended.  For instance, if you were porting `foo-scheme' then the
initialization file might be called `foo.init'.

  Your customized version should then be loaded as part of your scheme
implementation's initialization.  It will load `require.scm' from the
library; this will allow the use of `provide', `provided?', and
`require' along with the "vicinity" functions.  The rest of the
library will then be accessible in a system independent fashion.

  Please mail new working configuration files to `jaffer@ai.mit.edu' so
that they can be included in the SLIB distribution.

			   CODING STANDARDS

  All library packages are written in IEEE P1178 Scheme and assume that
a configuration file and `require.scm' package have already been
loaded.  Other versions of Scheme can be supported in library packages
as well by using, for example, `(provided? 'rev3-report)' or `(require
'rev3-report)'.

  `require.scm' defines `*catalog*', an association list of module
names and filenames.  When a new package is added to the library, an
entry should be added to `require.scm'.  Local packages can also be
added to `*catalog*' and even shadow entries already in the table.

  The module name and `:' should prefix each symbol defined in the
package.  Definitions for external use should then be exported by having
`(define foo module-name:foo)'.

  Submitted packages should not duplicate routines which are already in
SLIB files.  Use `require' to force those features to be supported in
your package.  Care should be taken that there are no circularities in
the `require's and `load's between the library packages.

  Documentation should be provided in Emacs Texinfo format if possible,
But documentation must be provided.