.TH "smujajgau" 1L "April 2000"
.SH NAME
smujajgau \- Dictionary builder for use with jbofihe and cmafihe
.SH SYNOPSIS
.PP
.B smujajgau
[-v]
.PP
.B smujajgau
.I dictionary-name
[
.I file1
.I file2
.BR "" ...
.I filen
]
.SH DESCRIPTION
.I smujajgau
is a program that reads a set of English definitions for Lojban words,
and formats them into a presorted dictionary for use by the jbofihe
and cmafihe programs.  The dictionary is arranged for rapid access.
.SH OPTIONS
.TP
.B -v
Show the program version and exit.
.TP
.B dictionary-name
This is the name of the formatted dictionary to be generated /
modified.  If the file exists, the definitions in the other file will
be added, replacing existing entries where they clash.
.TP
.B file1 .. filen
The source files to be added.  Lines beginning with # are treated as
comments and discarded.  Other lines should have one of the forms
.PP
lojban:english
.PP
lojban:english:comment

.SH SEE ALSO
.PP
.TP
.B jbofihe
Checks Lojban grammar checker and provides English translations of the
individual words.
.TP
.B cmafihe
Provides English translations of Lojban words in the input, without
checking the grammar.  Useful for getting a rough translation of
grammatically invalid text.
.SH FILES
.TP
.B /usr/local/lib/jbofihe/smujmaji.dat
This file is the default location where
.I jbofihe
and
.I cmafihe
expect to find the dictionary.  It should therefore be the default first argument to smujajgau (unless the software was installed to an alternative location.)
.SH SPECIAL TRANSLATION FORMATS
Dictionary entries for brivla (gismu & lujvo) are expected to provide
entries for each place of the word.  The English translation should
indicate the type of the word and the translation.  The types are
shown in the following table.  In addition, the gloss for a
translation X is shown depending on the context where it will arise in
the translation.  These defaults may be overridden.

.TS
center tab(@);
cb | cb || cb | cb | cb | cb
c | c || c | c | c | c.
Letter@Type@Noun@Verb@Qualifier@Tag
=
A @ Act @ X-er(s) @ X-ing @ X-ing @ X-er(s)
D @ Discrete @ X(s) @ being X @ X @ X
S @ Substance @ X @ being X @ X @ X
P @ Property @ X thing(s) @ being X @ X @ X thing(s)
R @ Rev. prop @ thing(s) X @ being X @ X @ things(s) X
I @ Idiomatic @ thing(s) X-ing @ X-ing @ X-ing @ thing(s) X-ing
E @ Event @ X(s) @ being X @ X @ X
.TE

To specify the dictionary entries, the lojban should take the form 'brivlaN',
where brivla is the word and N is the place number.  One of
the following may be suffixed to provide an override of the defaults
in the table : n v a or t (for noun, verb, adjective, tag
respectively.)

As an example, 'nanmu' might have entries
.PP
nanmu1:D;man
.PP
nanmu1n:D;man/men

.PP
whereas 'nandu' might have the definition
.PP
nandu1:P;difficulty
.PP
nandu2:I;have* difficulty
.PP
nandu3:S;conditions for difficulty
.PP
nandu3t:under conditions

a '*' is used in the places where the affixed -s, -er and -ing should
be applied (instead of putting them at the end of the English
translation, which is the default.)

The 'places.dat' file included in the distribution shows many
examples.

Where a translation with an 'n' suffix exists, this is used in place
of some other default forms in the table.  For example, this allows
special plural forms to be used in other places (e.g. tags.)

The dictionary also supports some 'pattern' translations.  This allows
defaults to be automatically generated for forms ending in '-gau',
'-zma, '-rai' etc, where there is a standard lujvo place structure
defined in terms of the prefix.

The 'Lojban' form for such patterns should be defined in the
dictionary as '*Mprefix+N' or '*M+suffixN' for prefix forms (e.g. nu+)
and suffix forms (e.g. +zmadu) respectively.  M is the 'precedence' (5
highest, 0 lowest), defining the order in which prefix v suffix
matches will be attempted.  N is the place number as usual.  The
letters n, v, a or t may be suffixed to define a particular form if
required, as for normal definitions.

The 'English' form is either a standard definition or a place
redirection. In standard definitions, the symbol % is used to mean the
translation of the rest of the lujvo form.  Place redirections take
the form @N, and mean that the lojban pattern form should be
translated as place N of the rest of the lujvo.

An example makes this clearer : zmadu.

*2+zmadu1:R;more %1q
.br
*2+zmadu2:R;less %1q
.br
*2+zmadu2t:than
.br
*2+zmadu3:@2
.br
*2+zmadu4:@3
.br
*2+zmadu5:@4
.br
*2+zmadu6:@5

The components are defined in terms of the full gismu forms, rather
than rafsi (hence zmadu rather than mau).  This is necessary because
the form of a rafsi can change when components are added to or
subtracted from a lujvo form.

When the 'English' form is given as '-', it indicates that the next
components inwards should be concatenated to form a new 'Lojban' form
for lookup.  This facility is only used for one thing so far - to
handle the rafsi 'zil' followed by the rafsi for a cmavo of selma'o PA
to puncture a place from the following form.  An example (to delete
the 1st place of a word, e.g. zilpavykla) :

*4zi'o+1:-
.br
*4zi'o+2:-
.br
*4zi'o+3:-
.br
*4zi'o+4:-
.br
*4zi'o+5:-
.br
*4zi'o+pa+1:@2
.br
*4zi'o+pa+2:@3
.br
*4zi'o+pa+3:@4
.br
*4zi'o+pa+4:@5
.br
*4zi'o+pa+5:@6

The pattern forms are all defined in a file 'patterns' in the
distribution.

For cmavo, there are special forms used for certain selma'o,
particular tenses.  These allow dependence on the context in which the
cmavo is used.  Logical connectives are also defined in a special way.
The file 'extradict' in the distribution provides examples.

(More documentation is required!)

.SH BUGS
ju'oru'e so'imei (Surely there are many).
.SH REFERENCES
.TP
.B http://go.to/jbofihe
Home page for the jbofihe project
.TP
.B http://www.rrbcurnow.freeuk.com/lojban/
My Lojban page.
.TP
.B http://www.lojban.org/
Home page of the Lojban community
.SH AUTHOR
Richard Curnow
.B <rpc@myself.com>