Re: [lojban] Error in bnf.300

[Bob LeChevalier <lojbab@lojban.org>]

At 10:44 AM 3/21/04 -0800, Robin Lee Powell wrote:
> There's a contradiction between grammar.300 and bnf.300 and, regardless
> of baselining issues, bnf.300 is *clearly* wrong:
>
>     text-1<2> = [(I [jek | joik] [[stag] BO] #) ... | NIhO ... #] 
> [paragraphs]
>
> The problem is that there's supposed to be a "text-1" betweev "BO]" and "#)".

I don't believe so, though I don't pretend to be able to read EBNF as well 
as the YACC grammar.

I could be wrong, but I believe that the EBNF is NOT designed to show the 
structure within a single rule parses - it just shows what is a legal 
string.  EBNF rules are also not supposed to be self-referencing so it 
would be invalid to have a text-1 included in the rule for text-1; that is 
what the "..." construct is supposed to communicate.

The first three rules of text_B_2 are in the in order to allow you to stutter:

.i   .i   .i   mi klama

This was generalized to allow
.ije  .ije  .ije   mi klama

The "mi klama" gets included when you finally get to the text_c_3 YACC 
rule/paragraphs EBNF rule.

It is possible that the EBNF is incorrect in how it represents the 4th rule 
of text_B_2, but my mind is twisted in the abbreviated 
recursion/alternation, so I cannot be sure.  If it is incorrect, I think 
the fix would be to change "| NIHO" to "] [NIHO]", because the EBNF as 
stated appears to forbid ".i ni'o mi klama"

> Here's the grammar.300 part (w/o comments); you mostly want to look at 
> text_B_2.
>
> %start text_0
>
> %%
>
> text_0                  :  text_A_1
>                         |  indicators_411  text_A_1
>                         |  free_modifier_32  text_A_1
>                         |  cmene_404  text_A_1
>                         |  indicators_411  free_modifier_32  text_A_1
>                         |  NAI_581  text_0
>                         ;
>
> text_A_1                :  JOIK_JEK_422  text_B_2
>                         /* incomplete JOIK_JEK  without preceding I */
>                         /* compare note on paragraph_10 */
>                         |  text_B_2
>                         ;
>
> text_B_2                :  I_819  text_B_2
>                         |  I_JEK_820  text_B_2
>                         |  I_BO_811  text_B_2
>                         |  para_mark_410  text_C_3
>                         |  text_C_3
>                         ;
>
> text_C_3                :  paragraphs_4
>                         |  /* empty */
>                         ;
>
>
> The test case is "mi broda .i je no da zo'u broda", which jbofi'e chokes on
> because of this problem.

I believe that it should be rejecting it.  (See my other post).

lojbab

--
lojbab                                             lojbab@lojban.org
Bob LeChevalier, Founder, The Logical Language Group
(Opinions are my own; I do not speak for the organization.)
Artificial language Loglan/Lojban:                 http://www.lojban.org