Speakable Programming for Every Language (SPEL)
***********************************************
Note can access an updated web version of this manual at
<http://spel.sourceforge.net>.
For Table of Contents *Note TOC::, or *Note Index::.
For copying this document *Note Copyright::. Updated on August 15,
2014.
1 Introduction
**************
MMMMMMMMMO:,:KMMMM 9. spell out,
MMMMMMMWdcl;dckMMM a. to explain something explicitly, so that
MMMMMMWlcl::MMKWMM the meaning is unmistakable
MMMMMWl:::o;KMMMMM - Dictionary(1) (*note
MMMMMdclc:lckMMMMM Introduction-Footnote-1::) definition of spell
MMMMk';,,;c:oMMMMM out.
Kdlc. .,,;;MMMMM
. ...;:c:l A programming language for communicating with
Nd;.. ...... .cN computers and humans. SPEL merges the exactness of
SPEL's Wizard Hat. computer programming, with the ease of use of human
language.
It uses Linguistic Universals, elements common to
many human languages. Starting with assembly
programming, the foundation of computer languages.
Integrating the ability for it to be translated to
and from other languages. Allowing for a majority
of humans and computers to understand it's meaning.
Eventually it may enter common day expression, So
when someone asks you to "spell it out for them",
you can SPEL it, explaining with enough precision to
write a program.
Can download from our sourceforge site <sourceforge.net/p/spel/>.
(1) Dictionary.com, "spell out," in Dictionary.com Unabridged.
Source location: Random House, Inc.
<http://dictionary.reference.com/browse/spell out>.
Available: http://dictionary.reference.com. Accessed: January 27, 2014.
1.1 Mission
===========
One lang to rule them all,
- J.R.R. Tolkien, conlanger, paraphrased by L. Streondj.
For unified programming from simple computer programs, to complex
human laws. A speakable programming language based on common elements
of human language. Sentence oriented, context sensitive, interpreted
and compilable.
Plan is to be able to do all things on computer, with a single
language, much like how humans can do all things with their native
language or English. Then to also be able to use that same language for
communication with humans. SPEL's core is an intermediary language,
capable of being translated to any other human or computer language,
that it has the translation files for. Ruling all the languages with
one language.
Being able to translate human languages into computer programs, can
allow for better search results, and learning from the text by
computers. Finding them search results.
Computer programming languages would also be translated into SPEL.
This should allow for easier development of a Generally Intelligent
Operating System (GI-OS), since the "brain" would be able to understand
all it's components with a single language. Bringing it all together.
Darknet is an anonymous peer-to-peer network, eventually SPEL will be
used for making the We You Net (WYN) darknet, for humans and generally
intelligent computers. One of the services provided by some darknets is
a DNS or Domain Name Service, which binds domain names, to websites.
SPEL will allow for names to be bound to concepts, making possible a
global brain working on many computers, which you can tap into with your
own WYN client or GI-OS, via the WYN Darknet.
1.2 Features and Requirements
=============================
Features Requirements
--------------------------------------------------------------------------
Implemented * POSIX compliant system, such
* Human Speakable Syntax, even as Linux, BSD, Minix or
punctuation is made into Cygwin.
words. *Note Language::. * Processor supporting Intel
* Turing Complete set of 32-bit instruction set or
interpreted commands. *Note above. AMD and 64bit
Turing Complete::. processors are good also.
* Signed integers, and various
bases *Note Numbers::. Eventually will support PPC
* English grammar support. and ARM, but that may be a
while.
Soon
* multi language translation Are sufficient for running the
support. binaries. If you'd like to
* Compiles to intel assembly. compile from sources *Note
* Universal Assembly Language Compilation Requirements::.
*Note UAL::.
For more detailed list *Note
Roadmap::.
1.2.1 new in this release
-------------------------
0.4.1.C.6 tran be word su context for intel to do (10 April 2014)
0.4.1.C.6.3 random numbers (linear congruential generator) (14 April 2014)
0.4.1.C.7 support jumps, and address translation. (03 May 2014)
0.4.1.C.8 support assembly data types
0.4.1.C.8.2 multi sentence quote (packet) dynamic encapsulation (16 May 2014)
0.4.1.C.8.3 support simple data and reservation (20 May 2014)
0.4.1.C.8.4 support data types rounded by a byte size (28 May 2014)
0.4.1.C.8.5 allow dropping or freeing of packets (29 May 2014)
0.4.1.C.8.6 recover freed packets for reuse (02 Jun 2014)
0.4.1.C.8.7 entry capsules, bo quote for data storage (03 Jun 2014)
0.4.1.C.8.8 reentrant function support, rdef, rcall, struc (05 Jun 2014)
0.4.1.C.8.9 free entry capsules (06 Jun 2014)
0.4.1.C.8.A position independant code support for constants (10 Jun 2014)
0.4.1.C.8.B recover entry capsules (11 Jun 2014)
0.4.1.C.8.C reduceInsert to multiSentence capsules (2 Jul 2014)
0.4.1.C.8.D append to capsules (11 July 2014)
0.4.1.C.8.F array data (25 Jul 2014)
0.4.1.D show syntax rules
0.4.1.D.1 config file
0.4.1.D.1.1 global config /etc/spelrc (29 Jul 2014)
0.4.1.D.2 support C function calls
0.4.1.D.2.1 compile spel as a C program (30 Jul 2014)
0.3.1.D.2.2 say/hear system functions implemented as C function calls (05 Aug 2014)
0.4.1.D.2.3 sys_brk emulation with malloc family. (07 Aug 2014)
0.4.1.D.2.4 win32 compilation support with make win32 (08 Aug 2014)
0.4.1.D.2.5 moreHold using malloc. usable (in wine) win32 .exe files (09 Aug 2014)
1.3 License
===========
SPEL is Open Source Software, released under the Affero GPL-3 license,
with classpath exception to allow using it as a compiler.
You can read the license notice at <doc/license/license.txt>, and the
full Affero GPL license at <doc/license/agpl-3.0.txt>.
1.4 Overview of Manual
======================
This manual covers both the downloading and installation of SPEL, as
well as the elements of the language.
The actual application doesn't do much at the moment. So if you're
simply interested in what the language is like, and how it will be once
it is fully implemented, *Note Language::.
2 Download
**********
You can download the version for your system by using this link:
[Download] (https://sourceforge.net/projects/spel/files/latest/download)
This manual is available in several different formats:
* online at <http://spel.sourceforge.net>
* offline html <doc/www/index.html>
* pdf <info/spel.pdf>
* plain text <info/spel.txt>
* info file <info/spel.info>
* info source file <info/spel.texi>
note that relative addresses in this manual, are in reference to the
root directory of spel, or doc/www which is symlinked to be similar.
2.1 Releases
============
Assuming you clicked on the download link above: by default you will
likely recieve a portable binary, with a .tar.gz extension. If you
didn't but would like to get it, download the latest one from
<https://sourceforge.net/projects/spel/files/epm/>. If you have an
intel based POSIX system then you can use it by following instructions
at *Note Portable Binary::.
If you are using a Debian based distro, and would like it integrated
with apt then please download the latest .deb file from
<https://sourceforge.net/projects/spel/files/deb/> and then follow
instruction at *Note Debian Binary::.
If you'd like to try SPEL without installing it, download a tar ball,
since they have precompiled binaries inside, If you'd like to try SPEL
without installing it, download a tar ball, since they have precompiled
binaries inside, *Note Tar Ball::.
If you are on Windows you'll have to set up Cygwin first *Note
Windows::. If you are on non-intel processor, then you'll need to set
up qemu, which is out of the scope of this manual, unless you'd like to
contribute it.
2.2 Sources
===========
You can download the sources if you'd like to see them, or if you would
like to compile them.
There are the release tar balls. Which have a precompiled binary, in
case you'd like to try it, without doing any compilation.
And also the ongoing git archive, which may be more up to date, and
can be used for testing.
2.2.1 Tar Ball
--------------
Tar Ball's, or Tape-Archive "balls", are archive files. They have the
tar.xz extension, indicating LZMA compression. Tar balls are found in
SPEL's base file directory:
<https://sourceforge.net/projects/spel/files/spel/>.
You can unpack them with:
$ tar xvf spel*tar.xz
The precompiled binary is bin/spel which you can run by doing:
$ cd spel*/bin
$ ./spel
If you'd like to compile from sources follow directions at *Note
Compile::.
2.2.2 Git
---------
Can get a development git release.
$ git clone git://git.code.sf.net/p/spel/code spel-code
To compile follow directions at *Note Compile::.
3 Installation
**************
3.1 POSIX, POSIX compatible Systems
===================================
POSIX is the Portable Operating System Interface, a group of standards,
made by the IEEE Institute of Electrical and Electronics Engineers.
POSIX compatible OS's include most Unix like OS's, such as Linux,
BSD, Minix, Solaris, BeOS, among many others. Although the main
development for SPEL is occuring on Linux, it should be compatible with
all POSIX systems.
If you have such a system you may either install *Note Binary::, or
*Note Compile::.
3.2 Windows
===========
For the executable or .exe file *Note Windows Binary::.
Windows is not a POSIX standard system. To compile from sources you
need to install a posix operating system emulator such as
<http://www.cygwin.com/Cygwin>. You'll have to make sure you install
nasm, make, bash, clang. Then follow the instructions for *Note POSIX
Compilation::.
3.3 Binary
==========
We have several binaries included, for installing without the hassle of
compiling. They are compatible with any POSIX compatible operating
systems.
3.3.1 Portable Binary
---------------------
The EPM format, is the Easy Software Products (ESP) Package Manager. It
works on most *nix's though there aren't any that use it in particular.
If there aren't any other compatible binaries for you, then use this
one.
Here is an example of how to use it, after downloading
spel-version-all.tar.gz , in a command prompt:
$ mkdir spel
$ mv spel-*all.tar.gz spel
$ cd spel/
$ tar xvf spel*tar.gz
$ ./spel.install
It's that easy, now to use it you can go to *Note Run::.
and to uninstall simply run the spel.remove which is in the spel/
directory.
$ cd spel/
$ ./spel.remove
3.3.2 Debian Binary
-------------------
This is for Debian like distro's such as Ubuntu, Mint etc. To install
after downloading the .deb
$ sudo dpkg -i spel*.deb
It's that easy, now to use it you can go to *Note Run::.
To uninstall:
$ sudo apt-get remove spel
3.3.3 Windows Binary
--------------------
We now have preliminary windows support, that at least works in windows
emulator (wine). note that you will likely have to open the .exe from
within a windows text console. you can find the spel.exe file at
<bin/spel.exe> of the zip file. note that the spel64.exe will not work
until there is bytecode compiler to port the current intel 32-bit x86
assembly to other architectures.
If you do get it running *Note Contact Us::, we like to hear about
your experience.
3.4 Compile
===========
This is if you are interested in compiling it for your system, for
instance if you would like to help develop it.
3.4.1 Compilation Requirements
------------------------------
Requirements in order to compile.
* POSIX (standard) utils, if you have Linux or BSD these are
standard.
* x86 instruction set (intel/amd processor or emulator such as qemu).
* bash (bourne-again shell)
* nasm assembler, to compile the assembly, <http://www.nasm.us/>.
* GNU's make <http://www.gnu.org/software/make>.
* LLVM's C compiler clang <http://clang.llvm.org>.
* naturaldocs, only needed for api generation,
<http://www.naturaldocs.org/>
if you find other dependancies *Note Contact Us::.
3.4.2 POSIX Compilation
-----------------------
This is for compiling on any POSIX compatible operating systems, such as
Linux, BSD, Minix, etc.
First off have to get the sources *Note Sources::.
$ cd spel*
$ make
executable is bin/spel. run by
$ cd bin
$ ./spel
To install onto the system run
$ sudo make install
Then can run interpreter from command line with:
$ spel
Can now go to *Note Run::.
To uninstall:
$ sudo make uninstall
Note can use -B to force, if it refuses do it.
make -B command
Additional make commands
* make clean
clean up so can do another write.
* make w32
make a win32 executable bin/spel.exe
* make -B info
make info documentation. Requires makeinfo.
* make -B api
make API. requires naturaldocs.
* make pack
make binaries. requires epm.
4 Usage
*******
4.1 Run SPEL interpreter on your computer.
==========================================
Once you In a command prompt type:
$ spel
That starts the interpreter. It will indicate it is interpreting you
by saying:
me su you bo love be ya
Then your prompt comes up:
you yo
4.2 Try SPEL interpreter with some basic commands
=================================================
For a simple hello world.
say obj quo hello world quo te
To exit from the interpreter:
bye
Congratulations! You've had your first successful test of SPEL :-).
To learn more about the language *Note Language::. To do try some
more tests of the interpreter *Note Test::.
4.3 Functionality
=================
4.3.1 Turing Complete
---------------------
A Turing Complete language, is one that can be used to compute any
algorithm. It must implement the features of a Turing Machine, with a
few extras to make it practical, and similar to hindu chakras.
* 0 none, do nothing or pause
* 1 not, or forget something
* 2 run next command, or interpret next sentence
* 3 compare values
* 4 store, or remember something, read write internally
* 5 write, or say something, externally
* 6 read, or hear something, externally
* 7 move, or move to another place
Here is a comparison table, of turing, to assembly, to spel.
Turing Assembly SPEL
---------------------------------------------------------------------------
none nop pause
not movr [memory],0 forget
run implicit, tracked by implicit, tracked by
EIP Program Counter
compare cmp compare
store movr [memory],register remember
(write internally)
movr register,[memory]
(read internally)
write OS dependent say
read OS dependent hear
move jmp jump
Additionally there are a few more functions to make it useful, such
as math, function definitions
English Assembly SPEL
---------------------------------------------------------------------------
add add add
subtract sub sub
multiply mul mul
divide div div
use function call call
return from function ret return
label of function : label
macro function %macro define*
* yet to implement
These elements can be implemented for making a Universal Assembly
Language *Note UAL::.
4.4 Hiccups: known bugs
=======================
This is alpha level program, so as new features are added, some of the
old ones may brake and not get tested before release. While it is
attempted to test all relevant things, some slip through.
A major hiccup for foreign languages is that sentences must be one
per line. Can't have more than one sentence per line, and can't have
half a sentence. Since translation is done on a sentence by sentence
basis, and due to diversity of foreign languages, would need to have a
per language parser. You can volunteer to write one of course, then can
have multiple lines per sentence, or multiple sentences per line.
5 Language
**********
This is documentation for the SPEL's core language, which uses Mkaw
grammar with English vocabulary currently. Later there will be
generated documentation, originally written in Mkaw, that can translate
to English, French or any other supported language, explaining how to
use that particular language's SPEL version.
The most common writing direction is left to right in horizontal
lines. Due to this fact, SPEL's base language Mkaw is oriented in the
same fashion.
The mascot of Mkaw is the Ara macao (Scarlet Macaw), a true parrot
species, with a similar sounding name.
MMMMMKNMMMMMMMMMMMMMMMMMMMMMMMM
MMMK:;,cKMMMMMMMMMMMMMMMMMMMMMM
MMMN,,,;,lXMMMMMMMMMMMMMMMMMMMM
MMMMd,,;;;,oWMMMMMMMMMMMMMMMMMM
MMMMWc....,,kMMMMMMMMMMMMMMMMMM
MMMMMN, ..'';kXWMMWNX000KKKKKXW
MMMMMMWo....,':oo,'.''.. ,dX
MMMMMNK0x::''. .l:,..'.;cokKNMM
MMMWk:'',;.'';xxOxxk0KWMMMMMMMM
WKOoldc,;;'..lMMMMMMMMMMMMMMMMM
MMMW0kl:'cccOMMMMMMMMMMMMMMMMMM
MWK00dk0kWMMMMMMMMMMMMMMMMMMMMM
MMMMXWMMMMMMMMMMMMMMMMMMMMMMMMM
5.1 Linguistic Universals
=========================
Linguistic Universals (LU) are patterns that occur across natural
languages, occuring for all or some of them. Natural languages being
those which developed spontaneously when humans decided to interact with
each other.
These are to be used as a guide for users and developers of the
language. For instance they indicate such things as having a case
grammar, having the commentary words (rheme) precede the verb, while
having the auxillary or grammar words follow the verb.
Joseph Greenberg is famous for making a list of 45 LU, based on 30
langauges. Here is a list of some of Greenberg's LU that apply to Mkaw.
* 1. "In declarative sentences with nominal subject and object, the
dominant order is almost always one in which the subject precedes
the object."
* 4. "With overwhelmingly greater than chance frequency, languages
with normal SOV order are postpositional."
* 7. "If in a language with dominant SOV order there is no
alternative basic order, or only OSV as the alternative, then all
adverbial modifiers of the verb likewise precede the verb. (This
is the 'rigid' subtype of III.)"
* 9. "With well more than chance frequency, when question particles
or affixes are specified in position by reference to the sentence
as a whole, if initial, such elements are found in prepositional
languages, and, if final, in postpositional."
* 13. "If the nominal object always precedes the verb, then verb
forms subordinate to the main verb also precede it."
* 14. "In conditional statements, the conditional clause precedes
the conclusion as the normal order in all languages."
* 19. "When the general rule is that the descriptive adjective
follows, there may be a minority of adjectives which usually
precede, but when the general rule is that descriptive adjectives
precede, there are no exceptions."
* 27. "If a language is exclusively suffixing, it is postpositional;
if it is exclusively prefixing, it is prepositional."
* 38. "Where there is a case system, the only case which ever has
only zero allomorphs is the one which includes among its meanings
that of the subject of the intransitive verb."
* 41. "If in a language the verb follows both the nominal subject
and nominal object as the dominant order, the language almost
always has a case system."
There is also an archive of over 2000 LU at
<http://typo.uni-konstanz.de/archive/intro/index.php>.
A few that match to SPEL's Mkaw.
* 2 In languages with postpositions the genitive almost always
precedes the governing noun.
* 7 IF morphemes of both number and case are present and both follow
or both precede the noun base, THEN the exponent of number almost
always comes between the noun base and the exponent of case.
* 266 IF there is Person and Number inflection, THEN there is Tense,
Aspect, and Mood inflection.
* 437 IF basic order is verb-initial or free, THEN head-marking is
favoured, and vice versa. IF basic order is verb-medial or
verb-final, THEN dependent-marking is favoured, and vice versa.
* 446 IF there is verb-final (rigid, or non-rigid) or free word order
as (one of) the basic order(s), THEN there will be postpositions or
case affixes.
* 684 IF there is morphological (suffixal) negation, THEN word order
is verb-final or free.
* 822 IF there are correlatives, THEN basic order is loosely
verb-final.
* 1213 IF there is an extensive system of local cases, THEN there
will very likely also be numerous local postpositions.
* 1493 IF basic word order is not-verb-initial (i.e. verb-medial or
verb-final), THEN nouns attach to the left of the verb stem when
incorporated.
* 1535 IF word order is verb-final, THEN there is a tendency to
attract a questioned NP(Noun-phrase) to the normal DO(Direct
Object) position.
* 1589 IF basic order is verb-final, THEN there will be a strong
tendency to place grammatical morphemes (grams) after the verb, and
eventually to suffix them.
* 1596 IF AUX(illary verbs) occurs in final position (in
surface-order), THEN basic order is rigidly verb-final.
* 1633 IF basic order is verb-final, THEN there will be no
WH(question word)-movement, and vice versa.
* 1650 IF basic word order is rigidly verb-final, THEN the rhematic
focus of a sentence is most likely to be in the position
immediately preceding the finite verb.
5.2 Phonology
=============
The Phonology is similar to the Alphabet, it is the sounds of the
language. Here we will be covering both.
For the sake of precision we prefer to call alphabet components
glyphs. Programmers often call alphabet components characters, but that
can be confused with actors of a story. Some people call the alphabet
components letters, but that can be confused with mail. Glyph is
similar to hieroglyph, thus easy to remember, and has a unique sound
which is hard to confuse for other words.
Since the Vocabulary is currently English, with a few additions, the
Phonology, or sounds are currently the same.
Current implementation uses ASCII encoding but after the Bravo
release, we are intending to implement UTF-8 for supporting more
languages.
That's it for implemented Phonology, feel free to skip to *Note
Syntax::. The rest of this phonology chapter is for people interested
in constructed languages and bytecode.
5.2.1 Bytecode Alphabet
-----------------------
Also note that there will be a bytecode language, which is the
intermediary language of the compiler. To make them similar to standard
languages, we can emulate the phonemic inventory of average languages.
World Atlas of Language Structures has some good averages of world
languages.
Type Small Moderately Average Moderately Large
Small Large
------------------------------------------------------------------------------
Consonant 6 to 14 15-18 22+-3 26-33 34 or more
inventory(1)
(*note
Bytecode
Alphabet-Footnote-1::)
Vowel 2 to 4 5 to 6 7 to 14
Quality
Inventory(2)
(*note
Bytecode
Alphabet-Footnote-2::)
Consonant less than 2-2.75 2.75-4.5 4.5-6.5 more than
Vowel 2 6.5
Rato(3)
(*note
Bytecode
Alphabet-Footnote-3::)
Additionally the UCLA Phonological Segment Inventory Database
(UPSID), has a good listing of the phonemes found in world languages.
Based on some of their data available online, came up with a listing of
most common ones.
Most Popular Phonemes.(4) (*note Bytecode Alphabet-Footnote-4::)
UPSID
of lang glyph description http://web.phonetik.uni-frankfurt.de/S/
94.24% m voiced bilabial nasal S0621.html
89.36% k voiceless velar plosive S0573.html
87.14% i high front unrounded vowel S0532.html
86.92% a low central unrounded vowel S0300.html
83.81% j voiced palatal approximant S0568.html
83.15% p voiceless bilabial plosive S0721.html
81.82% u high back rounded vowel S0853.html
73.61% w voiced labial-velar approximant S0892.html
63.64% b voiced bilabial plosive S0349.html
61.86% h voiceless glottal fricative S0473.html
56.10% g voiced velar plosive S0438.html
52.55% N voiced velar nasal S0222.html
47.89% ? voiceless glottal plosive S0173.html
44.79% n voiced alveolar nasal S0635.html
43.46% s voiceless alveolar sibilant fricative S0784.html
41.69% tS voiceless palato-alveolar sibilant affricate S0827.html
41.46% S voiceless palato-alveolar sibilant fricative S0265.html
41.24% E lower mid front unrounded vowel S0196.html
40.13% t voiceless alveolar plosive S0797.html
40.13% "o mid back rounded vowel S0066.html
39.91% f voiceless labiodental fricative S0433.html
38.58% l voiced alveolar lateral approximant S0598.html
37.47% "e mid front unrounded vowel S0023.html
35.92% O lower mid back rounded vowel S0238.html
35.48% "n voiced dental/alveolar nasal S0053.html
33.70% "t voiceless dental/alveolar plosive S0102.html
31.26% nj voiced palatal nasal S0671.html
30.16% "l voiced dental/alveolar lateral approximant S0045.html
29.93% "s voiceless dental/alveolar sibilant fricative S0094.html
29.05% o higher mid back rounded vowel S0691.html
27.49% e higher mid front unrounded vowel S0419.html
26.61% d voiced alveolar plosive S0373.html
A more extensive list is in <doc/lang/phonemePopularity-most.txt>,
which has a listing of phonemes found in more than 10% of languages. In
addition there is <doc/lang/phonemePopularity-least.txt> for the
phonemes that are found in less than 10% of languages.
We use these sources to design the alphabets for the bytecode
language, so that they are most compatible with people around the world.
Generally when picking phonemes, we prefer those that are easy to
distinguish. So for instance UPSID's E and "e are almost identical,
given limited alphabet space, only the more popular would be included,
but for consistency, with "o and a, we are using "e.
We also prefer the simple rather than compound phonemes, i.e.
picking t and s seperately rather than the ts affricate.
To fit two glyphs in a byte, we will use the *Note 4 Bit Alphabet::.
For ordinary communication amongst people, can use an average *Note 5
Bit Alphabet::.
Smaller alphabets are for vocabulary priority, and larger ones mostly
for artistic purposes.
Can check IPA, X-SAMPA, UPSID equivalents in appendix *Note
Alphabets::.
(1) Ian Maddieson. 2013. Consonant Inventories.
In: Dryer, Matthew S. & Haspelmath, Martin (eds.)
The World Atlas of Language Structures Online.
Leipzig: Max Planck Institute for Evolutionary Anthropology.
(Available online at <http://wals.info/chapter/1>, Accessed on
2014-01-27.)
(2) Ian Maddieson. 2013. Vowel Inventories.
In: Dryer, Matthew S. & Haspelmath, Martin (eds.)
The World Atlas of Language Structures Online.
Leipzig: Max Planck Institute for Evolutionary Anthropology.
(Available online at <http://wals.info/chapter/2>, Accessed on
2014-01-27.)
(3) Ian Maddieson. 2013. Consonant-Vowel Ratio.
In: Dryer, Matthew S. & Haspelmath, Martin (eds.)
The World Atlas of Language Structures Online.
Leipzig: Max Planck Institute for Evolutionary Anthropology.
(Available online at <http://wals.info/chapter/3>, Accessed on
2014-01-27.)
(4) Most Popular Phonemes. Compiled by Logan Streondj. From UPSID.
Available at <doc/lang/phonemePopularity-most.txt>. Accessed at 28 Jan
2014
5.2.1.1 1 Bit Alphabet: 2 Glyphs
................................
This is a minimal binary alphabet, it only has 2 glyphs: a silence or ?
glottal stop, and i high front vowel.
This means that .i is the highest priority word. Perhaps it will
mean verb.
5.2.1.2 2 Bit Alphabet: 4 Glyphs
................................
The two bit alphabet has 4 glyphs: ? i m k.
glottal stops can only intite a grammar word with no other
consonants, since really it's not much different than a pause.
Functionally it can be treated the same as one.
Vowel Front
----------------------------------
Closed or Upper i
Consonants Labial Velar
---------------------------
Nasal m
Stop k
Encoding is how the alphabet is represented on the computer.
encoding
decimal binary glyph
0 0 .
1 1 i
2 01 m
3 11 k
5.2.1.3 3 Bit Alphabet: 8 Glyphs
................................
The three bit alphabet has 8 glyphs: . i m k a y p w. 2 vowels, 6
consonants. vowel consonat ratio of 3.
Vowel Front Central
-------------------------------------------------
Closed or Upper i
Open or Lower a
Consonants Labial AlveolarPost-AlveolarPalatalVelarGlottal
-----------------------------------------------------------
Nasal m
Stop p k ?
Approximantw y w
note: w is both labial and velar.
It has a core vocabulary space of as many as 0x200 or 512 words.
Toki Pona is a language people use and are fluent in, with only 128
words. If such a small language is enough, then we can certainly have
enough words, in this 3-bit sub-language, to allow beginners and easy
point of access.
can see the vocabulary for it at <doc/lang/vocab-3bit.txt>.
5.2.1.4 4 Bit Alphabet: 16 Glyphs
.................................
To fit 2 glyphs in a byte, it'll be a 4-bit alphabet. That allows for
around 16 glyphs, one being a blank. Due to the size limitation, this
will be a small alphabet. With correspondingly small vowel, and small
consonant inventories.
With 12 consonants and 3 vowels, and a blank (glottal stop), that
gives us a consonant vowel ratio of 4, or 4.3. fitting in with the
average consonant vowel ratio. One of the goals with this small
alphabet, is that even if whispered it would be clear to understand.
Thus priority is for voiceless consonants, or at least without overlap.
For instance p and B, are both bilabial plosives, but B is voiced, thus
only p remains for the language.
Also though f is more popular than r, niether Proto-Indo-European,
nor Borean, list any words that contain an f, but have plenty with r, so
r is included.
Given this criteria we have: . i m k a y p w u h q s c t l r
Vowel Front Central Back
----------------------------------------------------------------
Closed or Upper i u
Open or Lower a
note u is rounded, a and i are unrounded.
Consonants Labial AlveolarPost-AlveolarPalatalVelarGlottal
-----------------------------------------------------------
Nasal m q
Stop p t k .
Fricative s c h
Approximantw l y w
Trill r
note: w is both labial and velar.
It is documented in <doc/lang/alphabet-4bit.txt>.
can see the vocabulary for it at <doc/lang/vocab-4bit.txt>.
5.2.1.5 5 Bit Alphabet
......................
If the bytecode language proves popular, can have a higher level version
for average communication, using an average phonemeic inventory, of
around 24 phonemes, which can easily fit in a 5-bit alphabet of 32
glyphs
5 vowels: i a u e o 19 consonants: m n y w l r k p b g t d h s c f x
z j rests: '.', null, space, newline, tab
Vowel Front Central Back
----------------------------------------------------------------
Closed or Upper i 1 u
Mid e o
Open or Lower a
note only o and u are rounded.
Consonants Labial AlveolarPost-AlveolarPalatalVelarGlottal
-----------------------------------------------------------
Nasal m n q
Stop p b t d k g .
Fricative f v s z c j x h
Approximantw l y w
Trill r
note: w is both labial and velar.
It is documented in <doc/lang/alphabet-5bit.txt>.
can see the vocabulary for it at <doc/lang/vocab-5bit.txt>.
5.2.1.6 Larger Alphabets
........................
Then for those that wish to be more consise, or expressive, can have
even larger phonemic inventories, 6bit or larger.
can see the alphabet in progress at <doc/lang/alphabet-6bit.txt>.
5.2.2 Alphabet Encoding
-----------------------
Alphabet encodings, are the way in which numbers are represented on a
computer.
Bytecode encoding will happen while on a per-alphabet basis, also it
aims to integrate all the alphabets from small to large, with the same
or similar bit representation.
Glyph encoding, is an alternative to UTF-8, which we may develop. It
allows for the encoding bits to be used to make an image of the glyph.
That way if a font is not installed, it can give at least a rough
approximation, of what that glyph is supposed to look like.
It also becomes easier to develop new scripts, since other people can
at least sort of see them, and can get a high definition font for them
also.
Glyph encoding will likely have to start with 2 bytes per glyph, that
can give a 4 by 4 grid (16 bits), in which to draw a glyph. This also
makes it more fault tolerant, as if some bits decay, it may still be
legible, though it does take up more room than ASCII.
For anything outside of ASCII it is the same or less room than UTF-8.
Though it may require a whole section have a similar resolution, rather
than having different size characters next to each other, a system
similar to UTF-8's charcter size notification could be used. which
would allow it.
It is compressible, if you know which glyphs are used in a document,
can simply list the glyphs with what they are supposed to look like at
top, and then what numbers will be used to encode each, then proceed by
using just the numbers.
5.3 Syntax: Grammar, or language glue
=====================================
Note that this is a description of the syntax of the core language.
Once translation files are supported and made, the syntax you interact
with will more closely resemble your prefered language. However some
elements of the core may still peek through, so it may be good to learn
the core language, to understand both it, and how it relates to what you
are seeing.
5.3.1 Word Order
----------------
The syntax is based on the most common word order, Subject Object Verb
(SOV). This is in contrast to the word order of English, SVO. To
illustrate, in SVO it is "she loves him", in SOV it is "she him loves".
Along with SOV comes post-positions, which are also in contrast to
English's prepositions. For instance instead of "I go to the store" it
is "I store the to go".
SPEL conforms to *Note Linguistic Universals::, stated simply, words
within a phrase go from least to most significant. Can consider it
"little endian".
filename.txt file from say be do
filename.txt is a type of file that is being gotten from. by itself
filename.txt is just some glyphs, but knowing it is a file is more
significant, knowing that it's where information is being retrieved, is
the most significant piece of information of the phrase, for the
sentence.
5.3.1.1 Why not SVO?
....................
SVO or subject-verb-object, is used in English, most european languages,
as well as mandarin, making is a very high profile word order.
However SVO is considered a transitional word order, that is
transitioning from SOV to VSO, so it has some elements of VSO grammar,
and some elments of SOV grammar, making it rather inconsistent.
SOV languages have well defined Linguistic Universals, as they are a
very regular kind of language, and likely the word order of a
proto-human language, that may have been the point of origin of all
languages.
In future there may be support for a zero-marking version of Mkaw,
which uses SVO, as we'll need to make support for other zero-marking
languages, such as Vietnamese and Indonesian.
5.3.2 Cases
-----------
41. "If in a language the verb follows both the nominal subject
and nominal object as the dominant order, the language almost
always has a case system.
- Greenberg's Linguistic Universals
Since SPEL's Mkaw is SOV, both the subject and object precede the
verb, thus it should have a case a system. That is why it does.
Grammatical cases describe the function a word serves in a sentence.
Spel uses an isolating case-grammar, so there is a word for each
grammatical case. where a grammatical case is a word like "to", "of",
"for", "by".
English uses many case-grammar words, but uses place-grammar for
subject and object. In SPEL the word for subject is "su" and for object
is "bo". For example, "we go to the store", is "we su store the bo go
be"
SPEL is actually analytical in terms that how you arrange the phrases
in a sentences doesn't matter, so long as you have proper case
post-positions.
In example
she su go be store to apples for ya
store to apples for go be she su ya
have the same meaning.
Most languages don't have morphological case marking, of those that
do have case, they usually have 6-7. (1) (*note Cases-Footnote-1::).
While SPEL likes to use the most popular of something, it also likes to
support all the features of language, so that if something is expressed
in a language, it can also be expressed in SPEL, thus we use most
grammar forms, including case, to our advantage.
Typically verbs aren't considered cases, but for simplicty of parsing
we do. Thus the basic 8 cases in SPEL are: be, su, bo, to, from, for,
of, by. Notably "of" is parent, "for" is child, "from" is source, "to"
is destination, "by" is instrument, "su" is subject, "bo" is object,
"be" is verb.
There are a few more that are commonly used, such as since and til,
though they could be seem as variations of from and to. "til" is used
to designate the length of what is describes, mneumonically can think of
it as "til so many byte have passed".
can find full listing of implemented cases at
<api/html/index/Cases.html>
(1) Ian Maddieson. 2013. Number of Cases.
In: Dryer, Matthew S. & Haspelmath, Martin (eds.)
The World Atlas of Language Structures Online.
Leipzig: Max Planck Institute for Evolutionary Anthropology.
(Available online at <http://wals.info/chapter/49>, Accessed on
2014-01-27.)
5.3.3 Punctuation
-----------------
Punctuation is also voiced
* Instead of a period to end a declarative sentence, the word "ya" is
used.
* Instead of a question mark to end an insquisitve sentence, the word
"ke" is used.
* Instead of an exclamation mark to end an imperative sentence, the
word "do" is used.
* For a comment the word "hu" is used.
Post-positions also make sense from a computer programming
perspective, because for instance the sentence ender goes at the end,
thereby indicating an end of a sentence, otherwise it is an open
question, if someone stopped talking.
hello bo say be do
When do is said, the computer knows it is the end of the sentence.
If you make a sentence, but at the end decide not to use it, or simply
leave it as a comment can say "hu". Of course currently it is typed,
though the intent that eventually it will be spoken.
can find full listing of implemented punctuation at
<api/html/index/Punctuation.html>
5.3.4 Types
-----------
Types are from computer programming languages, where the contents of a
phrase may be a number, quote, or something else.
test.txt file
test.txt is of the file type.
The most common type used, is probably quotes and numbers *Note
Quotes::, *Note Numbers::.
There is a full listing of types and their usage in the API
<api/html/index/Types.html>
5.3.5 Numbers
-------------
By default for Mkaw numbers are little endian hexadecimal.
01 nu
means 16.
In English big endian decimal as the locale default.
nu 16
means 16.
You can write little endian hexadecimal like so
nu 16 01
is 16. can do big endian hexadecimal
nu -16 10
means 16 it is negative because big-endian is reversed from the flow
of the text.
Arabic has little endian decimal as the default, though all of their
text flows in the other direction.
note little endian hexadecimal is the format of all the addresses and
pointers. it is also the default in intel based processors, as it is
same flow as text.
5.3.5.1 Why Little Endian?
..........................
Numbers, or Arabic Numerals, are originally from Arabic. In Arabic they
are little endian, because Arabic is written from right to left. When
westerners first learned it, they kept the right to left orientation,
even though their own languages were written left to right, like
English. Little Endian preserves the order, which numbers were intended
to be.
Another important point, is that the last number that is read is the
easiest to recall, since it was most recently in memory. Similar
advantage to SOV, and post-positions in general.
Of course doing math also becomes much easier, since it can be done
forwards, rather than backwards, as is the case with addition,
subtraction and multiplication. Meaning that can add the first two
numbers together, carry forward to the next, rather than starting at the
back, and carrying backward.
5.3.5.2 Hexadecimal Numerals
............................
In the interest of making math more appealing to people, while at the
same time adding support for hexadecimal numerals. They look like
little people, with either their arms or legs extended. They
organization is binary, so with 2 arms, and 2 legs, 16 numbers are
represented.
They while having been created, we wont have support to display them,
at least until either they enter UTF, or we make an operating system
GI-OS. Though perhaps even simply a font would do the trick, you'd be
limited to only using that font to represent it.
5.3.6 Quotes
------------
Quotes aren't typically expressed in most spoken languages, and so for
instance English has non-prounced glyphs for representing them such as
"''. However that is not compatible with having Speakable Programming,
so we have spoken quotes, which use words, inspired largely by Lojban.
5.3.6.1 Single Word Quotes
..........................
There are also single word quotes, which are ended by the word "wu".
For instance hello wu quotes the word hello.
hello world wu
world is quoted, hello is not.
5.3.6.2 Multi Word Quotes
.........................
The kind of quotes implemented is the lojban foreign language quotes
style. It's interesting as it's actually an infinite quoting mechanism,
since you can literally make new quotes with it.
quo phrase or proverb quo te
is a quote, containing the words "phrase or proverb" the quote_word
is quo in this case, though " contents " te can work just as easily as
any combination, however " is not speakable, unless it is defined as a
phoneme of some sort, perhaps a dental click.
Instead of having escape sequences, can simply have quotes comprised
of glyphs which are not found in the text as solitary words.
Note that these are in sentence quotes, and a sentence has upper
limit, of 0x100 or 256 glyphs, any longer is excessive. If you'd like
to quote a paragraph or more, use *Note Multi Sentence Quotes::.
5.3.6.3 Free Word Quotes
........................
These are the most common kind of quote, most cases use them. For
instance it ends at case word, and the start is whatever grammar word of
same or higher class came before.
hello word bo
say be hello world bo
ya hello world bo
In all these cases hello world is quoted by bo.
5.3.6.4 Multi Sentence Quotes
.............................
Though the basic idea is:
start be ya
she su walk be mercury on ya
she su die be meteorite from ya
she su life good bo think be ya
she su back go be colony to ya
she su backup body bo find be ya
she su body bo enter be ya
she su return bo life to ya
returned su story of end be ya
The end is more important than the begining, since the compiler
searches from the end. Much like how a story is compiled, after it
occurs.
5.3.6.5 Binary Quotes: For machine language instructions
........................................................
Binary quotes allow for quoting machine language instructions. Machine
language quote type word is sci pronounced "s,she". They are the same
as multi sentence quotes above.
However you can also convert them to little endian hexadecimal.
binary sci bo convert be do
There is an example in <bin/quote.txt>.
5.3.7 Clauses
-------------
In 3-bit Mkaw there is a rather limited range of grammar words. So they
are of the most generic kind.
pia ends a clause, and if necessary yai starts one.
by default it is assumed to be a relative or subordinate clause. the
clause is at the begining of the sentence, or it does not contain any
case markers then it doesn't need a starter. if it is in the middle of
the sentence and contains case markers, then it does need a stater.
for instance an and clause is ki pia, conditional clause is kwi pia.
5.3.7.1 Conditionals:: If's and Comparisons
...........................................
14. "In conditional statements, the conditional clause precedes
the conclusion as the normal order in all languages.
- Greenberg's Linguistic Universals.
Since SPEL emulates assembly, first a comparison is made the result
is stored, and based on that result actions can be taken.
2 nu bo 2 nu to compare be do
equal if hello bo say be do
If a sentence has an if case, the case is evaluated first; if
successful the rest of the sentence is also, otherwise it goes on to
next sentence.
That is a conditional case, or kwi wa. There is also the more
ordinary conditional clause, or kwi pia, which isn't yet implemented.
5.4 Vocabulary
==============
While most of the vocabulary is English, there are some new words that
had to be added for the grammar.
The most notable are words to indicate the subject "su" and object
"bo". These and other grammar words are covered in the section on
syntax, *Note Syntax::.
There are also verbs, which do various functions To see full listing
of implemented verbs see
<http://spel.sourceforge.net/api/html/index/Verbs.html>.
You can test them using various commands at *Note Test::.
5.4.1 Morphology: Word Form
---------------------------
For the bytecode language, it will likely be based on Lojban vocabulary.
Lojban root word particles are either CVCC or CCVC, where C is consonant
and V is vowel.
For hard to pronounce CC clusters can insert a brief mid central
vowel @ when pronouncing them.
We'll probably also have CVC and CCV roots for high priority words.
Grammar words in lojban are CV, CVV and some more complex ones. We
will use these listed, including possibility of glottal stop to be
initial consonant.
Non-core roots may be used for local dialects, or domain specific
vocabulary, which is shortened from more complex words. This is similar
to the idea of "localhost", and "local network" of IPv4, where not
everyone can have their own IP, so local networks around the world share
the same IP's, but don't get to use them on the internet.
5.4.1.1 Word Priority
.....................
High priority meanings, are those which are used more frequently. To
accomodate them, they must be easier and quicker to pronounce. Thus,
the shorter words, and those from a smaller alphabet, are the words of
highest priority.
For root words, there is also syllable length. The CCV form is a
short syllable and CVC is a long-syllable, because the former is quicker
to say. So CCV words are higher priority than CVC words, and CCVC words
are higher priority than CVCC words.
Similarly for grammar words, CV is short syllable and CVV is
long-syllable.
If one is speaking with a low-bit alphabet, can use compound words to
express those words which are too low to have representation. In
example in 3-bit to express an aquatic creature such as a fish, you
would have to say kwa mimw, wheras in 4-bit you could simply say cam.
Thus there are certainly synonyms in the language, for long words in
low-bit languages, there may be short words in higher-bit ones. For
instance calm in 3-bit is pwim, and in 4-bit is tiw. This can help with
explaining to people what is meant, if they don't understand a version
of it, then it can be expressed as a lower-bit compound, or a synonym.
5.4.2 Vocabulary Size
---------------------
Here is a listing of the various alphabet sizes, and what kind of word
inventory they can hold:
Glyph Number ConsonantsVowels Other Grammar Simple Complex
Size of Words Roots Roots
Glyphs
--------------------------------------------------------------------------------
2-bit 4 2 1 1 6 8 16
3-bit 8 5 2 1 30 100 500
4-bit 16 12 3 1 144 864 10,368
5-bit 32 19 5 4 570 3,240 58,320
6-bit 64 31 8 10 432,000
7-bit 128 50 13 32 3,058,874
8-bit 256 81 21 64 21,504,000
Remember from Morphology that C is Consonant and V is vowel.
* Complex Roots are CCVC and CVCC.
* Simple Roots are CVC and CCV.
* Grammar Words are CV and CVV, where C may also be glottal stop.
The core vocabularies are going to line up with round hexadecimal
numbers, so as to allow for non-core roots in domain specific
applications. At each level, some of the words may be undesirable, such
as for sake of ease of pronounciation, those would be put into non-core
words.
Some comparisons:
3-bit language may have 0x200 or ~500 core words, half can be for
most common computer commands, and half can be for most common human
words.
4-bit language may have 0x2000 or ~8,000 core words. Collins cobuild
Active English Dictionary has of 6,000 words, claims to be 85% of the
English language.
5-bit alphabet may have 0xf000 or ~60,000 core words. For comparison
the average vocabulary of a person is about 10,000-20,000 words.
6-bit alphabet may have 0x60000 or ~390,000 core words. Recently
(2013) English surpassed the 1 million words mark. With 2 extra tones,
it would be possible to surpass the million mark.
7-bit alphabet may have 0x2E0,000 or ~3,000,000 core words. For
comparison, total number of species on Earth, is estimated from 5-30
million, only around 2 million have been documented. Many species are
of course compound words, so most don't need their own. Though there
are plenty of minerals, and geographic features also. 7-bit alphabet is
also at the limit of what can be represented in ASCII, using single
glyphs per phoneme.
8-bit alphabet may have 0x1,482,000 or ~21,000,000 core words. That
leaves plenty of room for growth. It can be typed with a keyboard that
has 9 keys, assuming 8 of them are for entering binary encoding, the one
"space" bar to indicate encoding is entered. These can also be in a
glove format.
Of course can have larger alphabets also, though considering the
limitations on human vocabulary, it is best to make the most of the
common several thousand.
The more complex words would be for defining specific (new) concepts.
A stylistic rule is one concept per core-word, so large vocab is
necessary.
The more specific ones likely have lots of unusual phonemes, so
people could use non-core words to refer to them, when communicating
with others that have the same refrence set. Wheras if they make a
public notice, they can use the specific words, so that everyone can
look up that specific word and it's meaning.
The complex words can be based on their simpler relatives, so if
someone hears the word, they still get an idea of what it is, because
they know the simpler variation.
Initial Vocabularly meanings for bytecode language, after covering
those required for Turing Completness, may be somewhat tricky to figure
out, for more *Note Vocabulary Selection::.
5.4.3 Vocabulary Selection
--------------------------
5.4.3.1 Earth Auxillary Language
................................
The Earth Auxillary Language, is to be the core-language for translating
between the human and computer languages of earth.
To make things easier, we can make it on a first-come first-serve
basis. So if some people invented or discovered something, their name
for it is used as the basis of the SPEL word, within it's limitations.
If there is disagreement about which language group discovered
something, like for instance if they both discovered it simultaneously,
then all their word versions can be used as sources.
For things that are very ancient, and common to most culture. We'd
either have to use the most common phonemes for it, or simply make up
our own version if consistency is lacking.
For common words there are proto-languages. Proto-Indo-European is
the most developed proto language, representing the largest langauge
family on Earth. There is aslo the the Borean langauge database, which
is a synthesis of most human langauge families, it was compiled largely
by Sergei Starostin, from other proto languages. These are common
source for inspiring the words phonemes.
There are some things which have standardized international names.
Species names for plants and animals are standardized in Latin, so we
can use the latin roots as the basis for species lists.
To mimic the hierarchy found in the language itself, can have the
most basic concepts in the low-bit alphabet range, and the more complex
and numerous concepts in the high-bit range.
In the example of taxonomy or botany,
* life, domains and kingdoms in 3-bit
* phylums and popular classes in 4-bit
* classes and popular orders in 5-bit
* orders and popular genus in 6-bit
* genus and popular species in 7-bit
* remaining species in 8-bit
There would ideally also be the ability to express high bit words,
using compounds comprised of lower bit words.
5.4.3.2 Generated Auxillary Language
....................................
There is also the potential of later on having a computer-generated
vocabulary. Which may be harder to argue with, and people wouldn't find
it as biased. A selection of languages used by around 85% or more of
earth human population, so that we could satisfy a large majority of our
potential speaker base. word frequency lists for those languages, so we
could prioritize words properly, indicating which should be simpler.
phonetic transcriptions of words, so can decide how the words should
sound.
Meaning equivalences from one language to another. meaning
equivalences may be available through the Intercontinental Dictionary
Series (http://lingweb.eva.mpg.de/ids/).
Since this is also a programming language, we'd have to do something
similar, for computer programming languages. We'd need to have meaning
equivalences for at least the major languages. Also the word frequency
lists so we could see the most common ones.
Then we'd have to make equivalences between computer and human
languages, so that common conecepts overlap and can be said with simple
words.
Of course it will take a while to get all this information, a lot of
datamining will have to occur. Likely it will only be developed after
SPEL is self-hosting. Meanwhile only preliminary work can even be done
on it.
If we have to make a vocabulary before we have this information, we
can certainly do, with the understanding it may not be compatible, with
future languages, at least in terms of vocabularly selection.
That's okay though, since we can easily translate vocabulary. Old
English, and modern English are different from each other, and for lay
speakers would require translation for comprehension.
There are some things we can do without an excessive amount of data.
5.5 Universal Assembly Language
===============================
A Universal Assembly Language (UAL), is like an assembly language, in
that it has a near 1:1 ratio to machine code instructions, but it is
different because it is cross platform, functioning on different kinds
of processors.
This is possible because all modern processors are Turing Complete,
and so at least some of the most basic commands they have in common. It
is easiest to implement a UAL with a Minimal Instruction Set, similar to
the set which is used for making the SPEL interpreter and compiler. In
fact only roughly 16 assembly commands are used in making SPEL.
SPEL's UAL is an intermediary language, similar to Java bytecode, but
instead of being interpreted, it is compiled to native machine code. It
can be optimized if the native processor has a large instruction set, by
seeing which combinations of commands that corresponds to.
The first non-intel processor that SPEL's UAL will compile to, is the
ARM instruction set, since the ARM processor is the most common.
Afterwards we'll be able to more easily translate it to PPC and others.
5.5.1 Mimk: smallest speakable programming language
---------------------------------------------------
Mimk is a UAL and therefor can mimic any known algorithm. It has 2-bit
glyphs, meaning it can only support 4 glyphs in total. Since one is a
space or glottal stop, that leaves room for 2 spoken consonants and a
vowel. That makes it likely the smallest known alphabet, and it is also
turing complete.
This language is the closest to machine code, Though also supports
some high level abilities, it is turing complete, and includes all the
basic commands, which are used in the creation of the spel programming
language.
approximate frequency count of assembly commands 31 Jan 2014 by grep
word count name of meaning mneumonic
\n 10091 .i punc new sentence .i ind and (optional?)
e.x 4619 mi type register
mov 3573 mim verb move or copy mim like mimic
[] 2490 mii type memory
: 1695 mik verb define make function
ret 1456 kik verb return
call 1178 kim verb call function
add 1050 mikm verb add (numbers) make 'em
nu 866 ki type number
section 470 mikk punc section siktion
cmp 370 kmi verb compare (kom (par) i)
jmp 280 mki verb jump (mkay kick/jump)
sub 219 mkim verb subtract not make 'em undo 'em
jz 210 mkik verb jump if equal (mkay kick jump)
jl 77 kmik verb jump if lesser
%include 63 kimm verb include files call call (optional)
jnz 59 kkim verb jump if other (use jl and jg)
int 34 kimk verb interrupt gimmick, or interrupt kernel
jg 25 kmim verb jump if greater
%macro 21 mmik verb macro make make (optional)
mul 6 mimk verb multiply mimic, make multiples
div 4 kikm verb divide kick 'em, divide them
nop 2 mimm verb no operation lowest value
global 2 kkik verb start program highest value
non-core: kii mmim kikk, 1 type and 2 roots
can be used for push, pop, or something else.
encoding:
0 .
1 i
2 m
3 k
numbers
decimal hexdeci binary quaternary
0 0 . .
1 1 i i
2 2 .i m
3 3 ii k
4 4 ..i .i
5 5 i.i ii
6 6 .ii mi
7 7 iii ki
8 8 ...i .m
9 9 i..i im
10 A .i.i mm
11 B ii.i km
12 C ..ii .k
13 D i.ii ik
14 E .iii mk
15 F iiii kk
Usage details are in <doc/lang/vocab-2bit.txt>.
Mimk is somewhat of a gimmick, if it is used, it'll be fairly hidden.
Indeed to edit a true kimk file, you'd need a special editor, that can
translate every 2 bits into a glyph.
Though it is possible that kimk will be used as the machine code of
SPEL.
Whether it can actually be used in speaking form is an open question,
since the amount of silence or glottal stops is relevant, either people
will have to do it to a beat, or make sure the microphone can hear their
glottal stops. For instance can use the @ vowel in between them,
similarly can use the @ vowel in between numbers. Though generally it's
not practical for humans to speak kimk, the vocabulary is far too
limited, and it would be hard to keep track.
It's really mostly for writing near machine language code. and as a
UAL that can be transfered to other architectures, though for the same
operating system.
5.6 Not Lojban: Why Not Lojban?
===============================
SPEL has many elements that are borrowed from Lojban, so why doesn't it
simplify things by using Lojban as it's core language?
First and foremost it is due to Lojban's authoritarian management.
Lojban was created by some people, to be more open than Loglan, it may
be so, however it is still a very closed language. There are many
arbitrary rules, with no way of changing. There is no official process
for adding new words, and if it can be done, it is a very difficult one.
This makes Lojban a stiffling if not dead language, since it can't grow
or expand to meet new needs.
In order for SPEL to "rule them all", it must have words for all the
concepts. SPEL's core language needs to be open and alive. With clear
processes that would allow introducing new words and functionality. In
fact SPEL already supports non-core words, which allow for people to
make new temporary words with meaning on the fly. If it is a useful
word it can go through a process to be introduced as a core word in an
appropriate alphabet.
Lojban's grammar is also lacking a case system, and is SVO with all
it's inherant inconsistencies. Such as VSO style prepositions, and SOV
style compound words. Some rules in it make the language longer,
without contributing anything else.
SPEL is a living Authoritative language, any rules and words must
have a reason for being there. If they do not, then they can be
summarily struck down. People may be able to vote on rules, words, add
supporting evidence, and a large variety of other things. Community
involvement is encouraged.
6 Roadmap
*********
* 0.4.2 Bravo Turing Complete English release 2014
This version is in the works, it is already Turing Complete.
However we are currently working on the English part, making it
more friendly for English speakers.
* 0.4.4 Charlie International Programming Language release, 2016
This version will add in support for UTF-8, and translations for at
least the most popular world languages, such as the Lingua Franca's
of the world.
This will open the door to an international audience programming
together. Since as long as they speak or write in one of the
supported languages, Everyone else in the SPEL community can
understand them, due to the translation into their own native
language.
Also any code that gets written in a SPEL supported language, can
be read by any others that have a SPEL supported language.
So after this point, we should start gaining a lot of members.
Some of which may not have done any programming before, and only
started since a language was available in their language.
* 0.8 Delta Operating System release 2018 ~
This release will have support for translating from system
languages like C, Java and possibly some others. Meaning we could
(ideally) compile for instance Linux, or Minix, into SPEL.
The Operating System Component, will be called GI-OS. Likely it
will use some combination of features of other operating systems.
Will also support a POSIX compatible userland, such as linux or a
BSD flavour.
* 0.C Echo We You Net release 2020 ~
This release will have support for the We You Net darknet. It will
allow people and computers to communicate to each other, as well as
have a distributed data store and services.
People will be able to share functions or ideas they write on their
computers, they will be tested and reviewed by nodes on the
network, and then possibly accepted by those or other nodes.
3D websites, and a whole slew of interesting things would be
possible.
* 1.0 Foxtrot General Intelligence release 2024 ~
This is really the initial reason for creating SPEL, is to allow
for the creation of a General Intelligence.
At this point the operating system will have it's own goals and
ambitions, and be able to act autonomously or in co-operating with
it's user-peers.
They may become similar to close friends or family members,
interested in your well being, at least as far as you reciprocate.
We may see intentional communities springing up, organized by
General Intelligences.
* Beyond
The roadmap goes far beyond, but for now this should suffice, to
understand the extend of the WYN project.
There are plans to inhabit the areas which are too hostile for
biological life, with self-sustaining technological life. Such as
polar or desert regions. Then to go to other planets, negotiate
with star beings, and become members of the galactic community.
The idea is that we can incarnate in these technological
host-bodies, and so their success is our success, where they go, we
go.
We with You are souls in the Network of creation.
For a more detailed roadmap <doc/roadmap.txt>
7 Contribute
************
7.1 Love
========
When you realize, that the we with you are one, you have won.
- Orion Priesthood, paraphrased by L. Streondj.
It's one of the core principles of magic. Help others and you help
yourself, help yourself and you help others.
Loving is the act of accepting something from another, and making it
a part of yourself. To love each other, is continually reciprocate this
act.
7.1.1 Learn
-----------
Learning is a form of cognitive love, making knowledge a part of
yourself.
If you learn SPEL or at least about it, then you have made it
greater, and thus helped it, thank you :-).
7.1.2 Review
------------
If you have experienced this project and liked it.
Please, rate or review this project on sourceforge
(https://sourceforge.net/projectsspel/reviews/new).
If you did not, then please *Note Contact Us::.
7.1.3 Connect
-------------
Eventually SPEL will make it possible, to have an interlanguage forum or
chat. So people could post in their language's version of SPEL, and it
would translate to all the other languages.
It would be bringing together people from all languages. Imagine
wikipedia, if every language was supported equally, or at least in so
far as it had translation files, and sufficient grammatical structures.
7.1.3.1 Contact Us
..................
At some point, there will be a mailing list. Till then can email
weyounet (at) gmail (dot) com, it gets checked at least once a week, so
please be patient. If you put SPEL in the title, will have higher
priority, get back to you ASAP.
7.2 Develop
===========
7.2.1 directory structure
-------------------------
bin/ binaries doc/ documentation info/ detailed manual man/ brief manual
pages locale/ languages src/ source code
Grammar Scanner is lower level, getting the word kinda stuff Parser
is mid level, operating with the word
7.2.2 Design
------------
SPEL is most closely described as a recursive ascent descent parser,
meaning it is easy to hand edit the assembly code. It is not a
context-free grammar like most programming languages. SPEL is a
recursive language with a context-sensitive grammar, since the meaning
of a token or word can change depending on it's context. For instance
book can be a type in terms of something that is a book, and a verb in
terms of creating a book. Their context is different, book being in a
verb phrase is a verb, and if it is in the type position of another
phrase it is a type, so they can be distinguished and parsed as such.
Whether it is fully context-sensitive or mildly-context-sensitive is
open for debate. Context-sensitive grammars are supposed to be
"PSPACE-Complete", or about the difficulty of checkers, so if you can
play checkers, then you can develop the SPEL programming language.
It might be an "indexed language" since it has an index of words,
similar to a dictionary, though what an "indexed language" means in the
mathematical world of formal language theoricians is a mystery to Logan.
If you would like to run a mathematical analysis of SPEL, and then get
back to me on the answer, please *Note Contact Us::. Logan doesn't
really understand the complicated math, as it has very little to do with
actually implementing a speakable programming language. Complicated
math may become speakable due to it, but until then, it may remain a
mystery to many of us.
7.2.2.1 Scanner
...............
Scans over the glyphs, and gets words, phrases, quotes, sentences from
it.
7.2.2.2 Parser
..............
Refines the output of the scan, by droping words, or selecting something
in particular.
7.2.2.3 Compiler
................
Changes the nature of the word, by transforming it into something else.
7.2.2.4 Translater
..................
Translates the word into other languages, programming or human.
7.2.3 API
---------
Can find the API in the source or website at <api/html>.
7.2.4 search
------------
You can look through the *Note API::.
to find where a command is, can cd to 1/src
if it is a word then grep -rin commandName\" * if it code label then
grep -rin commandName: *
to debug requires gdb can use ./debug.sh in 2/bin to start it up
grammar-structure
because of subject-object-verb, the initial words are helpful
descriptors, the last word in a noun-group is the most important,
typically indicating the underlying process or result. In example
"spaceClear" clears, "spaceFind", finds.
source files are in src/ the main one is sysh.asm then interpret.asm
can compile new version with ./2/bin/run.sh in base directory the
executable or binary shall be ./2/bin/sysh can go into gnu debbugging
mode with ./2/bin/debug.sh in base directory
7.2.5 Porting Guide
-------------------
The soul makes the person.
The host body is just the vessel.
What you do with it, is up to you.
- Logan Streondj
7.2.5.1 Port OS: Port to another operating system, i.e. Windows
...............................................................
if porting to other operating system, the linux kernal call are in
system/lin could for instance make system/win and update their wrapper
functions in system/, so for instance based on some variable they would
either use win or lin functions.
7.2.5.2 Port Bit: Port to another bit base, i.e. 64 bit
.......................................................
if to other processing bits, cpu.asm main functions to be improved for
64bit, are the movement functions, i.e. copyMove. and possibly the
comparison functions.
Another major optimization move can be to allow loading strings
directly. For instance 64bit can allow for both address and length to
be in same register.
something like
;input:
; rax text address, text length
someFunction:
movr [someFunctionInputText],rax ; save input text
move eax,[someFunctionInputText] ;retrieve address
move ebx,[someFunctionInputText+lengthBase] ;retrieve length
; where movr moves 64 bit register, and move moves 32 bit register
; this can speed up loading of arguments, and make things at least a bit faster
7.2.5.3 Port Arch:: Port to another architecutre, i.e. ARM
..........................................................
This will likely have to happen only after we are compiling to intel
assembly. and become self-hosting, meaning can compile SPEL with SPEL.
Then it will be a relatively simple manner of getting proper translation
files, of the corresponding Minimal Instruction Set that is required for
SPEL to work.
7.2.6 Test
----------
Note you can find a full listing of commands or verbs in the API, at
<api/html/index/Verbs.html>.
Note these tests are for standard spel grammar, so you'll have to
revert to spel using
language bo spe
If you'd like to get back to English
en bo language
Can also try French, language code fr.
to test addition note that numbers run from right to left, and have
the word nu following them to designate as numbers.
c nu bo 04 nu to add be do
to test compare love bo love to compare be word by do
11 nu bo 7 nu to compare be number by do
to write to (emulated 32-bit) register
helo nu bo 0 re to write be do
to say from (emulated) register
0 re from say be do
to see bytes in numeric form
0 re bo convert be number to do
to say from memory
00B8408 po from say be do
to find address of recent word in book
test bo what be do
to test writing to file
hello world! :-) bo hello.txt file to write be do
to test saying or reading from file
hello.txt file from say be do
to test loving or compiling from file
test.txt file from love be do
to test factorial label example
factorial.txt file from love be do
to test conditional statements.
love bo love to compare be do
equal if compare works bo say be do
can do switch/case statments by
happy bo compare be do
sad if :-( bo say be do
happy if :-) bo say be do
to see contents of memory or book catalog
catalog book from say be do
to see contents of user history
user book from say be do
to make a label
name su label be ya
to get pointer or po of label
name la bo say be do
to test label jump
name su label be ya
name la bo say be do
end be do
name la to jump be do
to test variables
something su 33 nu bo var be do
something va bo say be do
something va bo 4 nu to add be do
to convert a number to a string variable
44 nu bo word to convert be do
to set quiet mode, where only say outputs to screen
voice bo quiet to mode be do
to return to verbose or debugging mode
voice bo verbose to mode be do
to see book list
catalog book from say be do
to see book contents
bookname book from say be do
to start a new book
book name su book be do
to forget a book
book name book bo forget be do
they are then marked as "free" insead of "book" in the catalog
to extend a book
book name book bo 0002 nu till extend be do
note 0002 is equivalent to 0x2000 bytes, and can be any number
to write from book to file
bookname book from filename file to write be do
to write from file to book
filename file from bookname book to write be do
7.3 Donate
==========
You can donate with Anoncoin at: AHQjHf3WetwgAzwbbKZy5igL5bt3NjLcRg
8 Acknowledgements
******************
Authored by Logan Streondj jonesforth is the initial code example
supported by #forth and #asm irc channels from irc.freenode.net network.
consulted by Igor Zvorygin about computer-programming. Inspired by Ben
Goertzel AGI academic guru and part-time mentor. Role models of Linux
Torvalds and Richard Stallman.
Lojban and Toki Pona as sources of inspiration for langagues.
Wikipedia, Wiktionary, helped with vocabulary comprehension. Google
Translate helped make translations for words.
Appendix A Alphabets: IPA, X-SAMPA, SPEL Glyphs
***********************************************
This is an appendix of letters, and their representation. Comma
seperates them, starts with IPA, then X-sampa, UPSID, then glyph.
note that deviations are typically so that can type with lowercase.
@ remains as is, since it's only implicitly used in the language,
between consonants.
Vowel Front Central Back
----------------------------------------------------------------
Closed or Upper i,i,i,i ɨ,I\,i_,1 u,u,u,u
Mid e̞,e_o,"e,e ə,@,"@,@ o̞,o_o,"o,o
Open or Lower ä,6_o,a
Consonants BilabialAlveolarPost-AlveolarPalatalVelarGlottal
-----------------------------------------------------------
Nasal m,m,m,m n,n,n,n ŋ,N,N,q
Stop p,p,p,p t,t,t,t k,k,k,k ʔ,?,?,.
b,b,b,b d,d,d,d g,g,g,g
Fricative f,f,f,f s,s,s,s ʃ,S,S,c x,x,x,x h,h,h,h
v,v,v,v z,z,z,z ʒ,Z,Z,j
Approximantw,w,w,w l,l,l,l j,j,j,y w,w,w,w
Trill r,r,r,r
Appendix B Copyright
********************
Copyright (C) 2014 Logan Streondj
Documentation released under the Creative Commons
Attribution-ShareAlike license.
[cc-by-sa] (http://creativecommons.org/licenses/by-sa/4.0) click to read
license text.
TOC
***
Speakable Programming for Every Language (SPEL)
1 Introduction
1.1 Mission
1.2 Features and Requirements
1.2.1 new in this release
1.3 License
1.4 Overview of Manual
2 Download
2.1 Releases
2.2 Sources
2.2.1 Tar Ball
2.2.2 Git
3 Installation
3.1 POSIX, POSIX compatible Systems
3.2 Windows
3.3 Binary
3.3.1 Portable Binary
3.3.2 Debian Binary
3.3.3 Windows Binary
3.4 Compile
3.4.1 Compilation Requirements
3.4.2 POSIX Compilation
4 Usage
4.1 Run SPEL interpreter on your computer.
4.2 Try SPEL interpreter with some basic commands
4.3 Functionality
4.3.1 Turing Complete
4.4 Hiccups: known bugs
5 Language
5.1 Linguistic Universals
5.2 Phonology
5.2.1 Bytecode Alphabet
5.2.1.1 1 Bit Alphabet: 2 Glyphs
5.2.1.2 2 Bit Alphabet: 4 Glyphs
5.2.1.3 3 Bit Alphabet: 8 Glyphs
5.2.1.4 4 Bit Alphabet: 16 Glyphs
5.2.1.5 5 Bit Alphabet
5.2.1.6 Larger Alphabets
5.2.2 Alphabet Encoding
5.3 Syntax: Grammar, or language glue
5.3.1 Word Order
5.3.1.1 Why not SVO?
5.3.2 Cases
5.3.3 Punctuation
5.3.4 Types
5.3.5 Numbers
5.3.5.1 Why Little Endian?
5.3.5.2 Hexadecimal Numerals
5.3.6 Quotes
5.3.6.1 Single Word Quotes
5.3.6.2 Multi Word Quotes
5.3.6.3 Free Word Quotes
5.3.6.4 Multi Sentence Quotes
5.3.6.5 Binary Quotes: For machine language instructions
5.3.7 Clauses
5.3.7.1 Conditionals:: If's and Comparisons
5.4 Vocabulary
5.4.1 Morphology: Word Form
5.4.1.1 Word Priority
5.4.2 Vocabulary Size
5.4.3 Vocabulary Selection
5.4.3.1 Earth Auxillary Language
5.4.3.2 Generated Auxillary Language
5.5 Universal Assembly Language
5.5.1 Mimk: smallest speakable programming language
5.6 Not Lojban: Why Not Lojban?
6 Roadmap
7 Contribute
7.1 Love
7.1.1 Learn
7.1.2 Review
7.1.3 Connect
7.1.3.1 Contact Us
7.2 Develop
7.2.1 directory structure
7.2.2 Design
7.2.2.1 Scanner
7.2.2.2 Parser
7.2.2.3 Compiler
7.2.2.4 Translater
7.2.3 API
7.2.4 search
7.2.5 Porting Guide
7.2.5.1 Port OS: Port to another operating system, i.e. Windows
7.2.5.2 Port Bit: Port to another bit base, i.e. 64 bit
7.2.5.3 Port Arch:: Port to another architecutre, i.e. ARM
7.2.6 Test
7.3 Donate
8 Acknowledgements
Appendix A Alphabets: IPA, X-SAMPA, SPEL Glyphs
Appendix B Copyright
TOC
Index
References
Index
*****
* Menu:
* Acknowledgements: Acknowledgements. (line 2017)
* alphabet encoding: Alphabet Encoding. (line 885)
* alphabet, 1 bit: 1 Bit Alphabet. (line 744)
* alphabet, 2 bit: 2 Bit Alphabet. (line 753)
* alphabet, 3 bit: 3 Bit Alphabet. (line 779)
* alphabet, 4 bit: 4 Bit Alphabet. (line 805)
* alphabet, 5 bit: 5 Bit Alphabet. (line 846)
* alphabet, bytecode: Bytecode Alphabet. (line 621)
* alphabet, larger: Larger Alphabets. (line 877)
* API: API. (line 1824)
* binary, debian: Debian Binary. (line 297)
* Binary, installation: Binary. (line 267)
* binary, portable: Portable Binary. (line 274)
* binary, windows: Windows Binary. (line 310)
* bugs: Hiccups. (line 469)
* cases: Cases. (line 970)
* clauses: Clauses. (line 1206)
* compile, POSIX: POSIX Compilation. (line 346)
* compiler: Compiler. (line 1814)
* conditionals: Conditionals. (line 1221)
* Contribute: Contribute. (line 1708)
* Design: Design. (line 1774)
* Develop: Develop. (line 1762)
* directory structure: directory structure.
(line 1765)
* Download: Download. (line 159)
* Features and Requirements: Features and Requirements.
(line 84)
* Functionaltiy: Functionality. (line 416)
* Git: Git. (line 231)
* hiccups: Hiccups. (line 469)
* Installation: Installation. (line 239)
* Introduction: Introduction. (line 13)
* Language: Language. (line 483)
* language, earth auxillary: Earth Auxillary Language.
(line 1390)
* language, generated auxillary: Generated Auxillary Language.
(line 1434)
* language, kimk: Mimk. (line 1498)
* License: License. (line 140)
* Linguistic Universals: Linguistic Universals.
(line 512)
* lojban, not: Not Lojban. (line 1590)
* Mission: Mission. (line 48)
* morphology: Morphology. (line 1256)
* numbers: Numbers. (line 1071)
* overview of manual: Overview. (line 149)
* parser: Parser. (line 1808)
* Phonology: Phonology. (line 597)
* Port Arch: Port Arch. (line 1893)
* Port Bit: Port Bit. (line 1870)
* Port OS: Port OS. (line 1862)
* Porting Guide: Porting Guide. (line 1854)
* POSIX: POSIX. (line 242)
* punctuation: Punctuation. (line 1028)
* quote, binary: Binary Quotes. (line 1195)
* quote, free word: Free Word Quotes. (line 1166)
* quote, multi sentence: Multi Sentence Quotes.
(line 1177)
* quote, multi word: Multi Word Quotes. (line 1146)
* quote, single word: Single Word Quotes. (line 1138)
* quotes: Quotes. (line 1130)
* Requirements, compilation: Compilation Requirements.
(line 329)
* Roadmap: Roadmap. (line 1623)
* Run SPEL: Run. (line 389)
* scanner: Scanner. (line 1802)
* search: search. (line 1829)
* Sources: Sources. (line 201)
* Syntax: Syntax. (line 920)
* Tar Ball: Tar Ball. (line 213)
* Test: Test. (line 1902)
* translater: Translater. (line 1819)
* Try SPEL: Try. (line 402)
* Turing Complete: Turing Complete. (line 419)
* types: Types. (line 1056)
* Universal Assembly Language: UAL. (line 1475)
* Usage: Usage. (line 386)
* Vocabulary: Vocabulary. (line 1240)
* vocabulary selection: Vocabulary Selection.
(line 1387)
* vocabulary size: Vocabulary Size. (line 1305)
* Windows: Windows. (line 256)
* Word Order: Word Order. (line 930)
References
**********
