[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
jbofi'e examples [was Re: [lojban] useful tools]
On Thu, Aug 31, 2000 at 12:44:38PM -0400, Bob LeChevalier (lojbab) wrote:
>
> That sounds like the old version of Nora's parser/glosser, which took
> Cowan's "official" parser's parenthesized output (the parentheses are
> produced by the parser and show the levels of nesting for grammar
> structures; braces and brackets are alternated with parentheses) and lines
> the English up with the Lojban words in that parenthesized output. It of
> course MUST be used with a fixed font or nothing will line up, but since it
> is a DOS program, fixed font should be the norm.
>
> The new version, which should be the one on lojban.org, does not leave the
> parentheses in, and the only symbols are braces around the case tags that
> label the place structures. You might want to take a look if you can stand
> a sickly DOS program.
>
> What I don't know is how this compares with jbofi'e.
>
In rough terms :
- jbofi'e was developed under Unix and is very unlikely to be portable
to DOS (memory size issues).
- the parser and glosser functions are all integrated into the one tool.
- the parser was build using the bnf.300 file as a reference - I set out
to avoid using John Cowan's yacc grammar entirely. I thought this
would ultimately provide a route for validating the equivalence of the
bnf grammar and the yacc grammar, by checking whether the same texts
parse or fail to parse on the two tools.
- some of the output formats are similar (though not identical to) those
produced by the 'official' software.
Here are a few example outputs :
Input text (my attempt at translating the maxim 'there's a time and a
place for everything'):
mapti jai ca je vi me ro da
jbofihe on its own gives
(0[mapti {jai <ca je vi> <me (1ro da)1>}])0
Which is fairly similar to the original parser format (parentheses are
numbered so you can match things in very deeply nested constructions.)
jbofihe -x gives a gloss as well :
[<<^mapti /compatible/ /[type-of]/ jai /being the/ ca /time of/ je /and/ vi /location of/ me /amongst those that are/ ^ro /every/ /(of)/ da /X/^^>>]
and finally jbofihe -x -b puts the gloss in 'block' mode.
[ << | mapti jai ca je vi
[ << | compatible [type-of] being the time of and location of
[ << \
1 2 3
me / ro da \ | >> ]
amongst those that are | every (of) X | | >> ]
| | / >> ]
4 4 3 2 1
In the -x cases, the differently shaped brackets are used to envelop
particular constructions (e.g. sumti, sentence etc).
I've just noticed that last example doesn't show the case tagging
feature, so here's one that does
le zarci cu se klama mi la london.
[ ( le zarci ) cu << se klama
[ ( the trading place(s) ) is/does << being destination(s)
[ ( klama2 (destination(s)) ) <<
1 2 2 3
>> ( mi ) ( la london. ) ]
>> ( I, me ) ( [NAME] ) ]
>> ( klama1 (go-er(s)) ) ( klama3 (origin(s)) ) ]
3 4 4 5 5 1
--
----------------------------------------------------------------------
Richard P. Curnow rpc@myself.com
Weston-super-Mare
United Kingdom http://go.to/richard.curnow/