Talk:BPFK Section: Termsets
Archived Tiki forum discussion for BPFK Section: Termsets. Original authors, UTC times, and post IDs are retained below.
BPFK Section: Termsets (Tiki thread #3476)
Re: BPFK Section: Termsets
xorxes · 2005-01-11 18:34 UTC · Tiki post #3477 (reply to #3476)
*mi klama ge le zarci le zdani gi le zdani le zarci "I go to the market from the house and to the house from the market"
It is currently ungrammatical, to make it grammatical we have to add a nu'i somewhere. But the only reason that this was made ungrammatical is that you need more than one word of lookahead to resolve it. PEG can handle it easily. Shouldn't we make it grammatical with the PEG grammar?
mu'o mi'e xorxesBPFK Section: Termsets
rlpowell · 2005-01-11 19:26 UTC · Tiki post #3479 (reply to #3476)
> Re: BPFK Section: Termsets > > Consider this sentence: > > *mi klama ge le zarci le zdani gi le zdani le zarci > "I go to the market from the house and to the > house from the market" > > It is currently ungrammatical, to make it grammatical we have to > add a nu'i somewhere. But the only reason that this was made > ungrammatical is that you need more than one word of lookahead to > resolve it. PEG can handle it easily. Shouldn't we make it > grammatical with the PEG grammar?
Doesn't that entirely obviate nu'i ?
This is probably not a bad thing, but I just want to check.
-RobinBPFK Section: Termsets
Anonymous · 2005-01-11 19:26 UTC · Tiki post #3480 (reply to #3476)
> Shouldn't we make it grammatical with the PEG grammar?
No.
-- "They tried to pierce your heart John Cowan with a Morgul-knife that remains in the http://www.ccil.org/~cowan wound. If they had succeeded, you would http://www.reutershealth.com
become a wraith under the domination of the Dark Lord." --GandalfBPFK Section: Termsets
rlpowell · 2005-01-11 19:27 UTC · Tiki post #3484 (reply to #3476)
> [email protected] scripsit: > > > Shouldn't we make it grammatical with the PEG grammar? > > No.
Umm, *because*?
In fourty-eight *THOUSAND* lines of IRC, there are *SIX* instances ce'e, nu'i, nu'u and pe'e. That's *combined*. I can't imagine a reason to
- not* build in a work around for an aspect of the language that is
- that* hard to use.
Here they are:
[15:18] <fracture> nu'ige zo by. .a zo beiste nu'ugi zo zy. .a zo zgana toji'a zo viska toi
Parses.
[15:27] <fracture> liste nu'ugi zo zy. zo zgana toji'a zo viska toi
Does not parse.
10 Apr 2004 14:49:58 <rlpowell> simsa zo pe'e
quoted, not actually used.
06 Dec 2004 00:41:01 <rlpowell> .i pilno lo nandu zenba tadji ce'e lo certu kelci pe'e je lo nandu jdika tadji ce'e lo to'e certu kelci
Parses (took me 10 minutes to get that working before I pasted it to the channel).
16 Dec 2004 12:19:00 <rlpowell> zbasu da ce'e de pe'e je di ce'e do
Parses.
16 Dec 2004 12:19:57 <rlpowell> zbasu nu'i da de nu'u pe'e je nu'i di do
Parses.
So of the six, two people used it, 5 of them count as usage, and 4 of those parse. Of those four, three are known for a fact to only parse because I tested them offline first; I know this because it was last month.
That's a pretty crappy track record.
-Robin
-- http://www.digitalkingdom.org/~rlpowell/ *** http://www.lojban.org/ Reason #237 To Learn Lojban: "Homonyms: Their Grate!"
Proud Supporter of the Singularity Institute - http://singinst.org/BPFK Section: Termsets
rab.spir · 2005-01-11 19:27 UTC · Tiki post #3485 (reply to #3476)
> [email protected] scripsit: > > > Shouldn't we make it grammatical with the PEG grammar? > > No.
Why not? --
Rob SpeerBPFK Section: Termsets
Anonymous · 2005-01-11 21:35 UTC · Tiki post #3491 (reply to #3476)
> In fourty-eight *THOUSAND* lines of IRC, there are *SIX* instances > ce'e, nu'i, nu'u and pe'e. That's *combined*. I can't imagine a reason to > *not* build in a work around for an aspect of the language that is > *that* hard to use.
The article I wrote on nanba for wikipedia features termsets. :) (http://jbo.wikipedia.org/wiki/Nanba)
I think I even used them right!
But it did take me 10 or 15 minutes, at least, of rereading the CLL and feeding them into jbofi'e to make sure it was right.
--
Jay Kominek <[email protected]>BPFK Section: Termsets
Anonymous · 2005-01-12 02:45 UTC · Tiki post #3686 (reply to #3476)
Consider this sentence:
*mi klama ge le zarci le zdani gi le zdani le zarci "I go to the market from the house and to the house from the market"
It is currently ungrammatical, to make it grammatical we have to add a nu'i somewhere. But the only reason that this was made ungrammatical is that you need more than one word of lookahead to resolve it. PEG can handle it easily. Shouldn't we make it grammatical with the PEG grammar?
mu'o mi'e xorxesBPFK Section: Termsets
xorxes · 2005-01-12 02:46 UTC · Tiki post #3687 (reply to #3476)
> In fourty-eight *THOUSAND* lines of IRC, there are *SIX* instances > ce'e, nu'i, nu'u and pe'e. That's *combined*. I can't imagine a reason to > *not* build in a work around for an aspect of the language that is > *that* hard to use. > > Here they are: > > [15:18] <fracture> nu'ige zo by. .a zo beiste nu'ugi zo zy. .a zo zgana > toji'a > zo viska toi > > Parses.
This is the only relevant instance for the structure involved, which is forethought connection of termsets. All the others use afterthought connectives with pe'e, which would not be replaced by ge.
That instance parses, but it is not the intended use of nu'ige. In that example, you would get the exact same meaning without the nu'i: {ge zo by. .a zo beiste gi zo zy. .a zo zgana} is already grammatical and has the same meaning, everything is part of the same term.
So in the fourty-eight thousand lines of IRC there are zero instances of proper {nu'i gek term term /nu'u/ gik term term /nu'u/}.
I think the PEG rule would have to be something like this:
gek-termset <- gek terms-gik-terms
terms-gik-terms <- term (gik / terms-gik-terms) term
That way the number of terms grabbed after the gik would be the same as the number of terms between gek and gik.
mu'o mi'e xorxes
__________________________________
Do you Yahoo!?
The all-new My Yahoo! - Get yours free!
http://my.yahoo.com
BPFK Section: Termsets
Anonymous · 2005-01-13 18:04 UTC · Tiki post #3707 (reply to #3476)
> On Tue, Jan 11, 2005 at 11:20:31AM -0800, Robin Lee Powell wrote: > > In fourty-eight *THOUSAND* lines of IRC, there are *SIX* instances > > ce'e, nu'i, nu'u and pe'e. That's *combined*. I can't imagine a reason > > to *not* build in a work around for an aspect of the language that is > > *that* hard to use. > > The article I wrote on nanba for wikipedia features termsets. :) > (http://jbo.wikipedia.org/wiki/Nanba)
One more: cire cilce carce jarco ce'e reci cilce carce cu jarco ..i cire cilce carce jarco cu jarco reci cilce carce ..ije'i reci cilce carce cu se jarco cire cilce carce jarco?
mu'omi'e pier. --
le xruki le ginxre xrixruba xu xrula cu xrani?BPFK Section: Termsets
rlpowell · 2005-01-13 18:04 UTC · Tiki post #3709 (reply to #3476)
> On Tuesday 11 January 2005 16:03, [email protected] wrote: > > On Tue, Jan 11, 2005 at 11:20:31AM -0800, Robin Lee Powell > > wrote: > > > In fourty-eight *THOUSAND* lines of IRC, there are *SIX* > > > instances ce'e, nu'i, nu'u and pe'e. That's *combined*. I > > > can't imagine a reason to *not* build in a work around for an > > > aspect of the language that is *that* hard to use. > > > > The article I wrote on nanba for wikipedia features termsets. :) > > (http://jbo.wikipedia.org/wiki/Nanba) > > One more: > > cire cilce carce jarco ce'e reci cilce carce cu jarco > .i cire cilce carce jarco cu jarco reci cilce carce > .ije'i reci cilce carce cu se jarco cire cilce carce jarco?
No forethought, though.
-Robin
-- http://www.digitalkingdom.org/~rlpowell/ *** http://www.lojban.org/ Reason #237 To Learn Lojban: "Homonyms: Their Grate!"
Proud Supporter of the Singularity Institute - http://singinst.org/BPFK Section: Termsets
rlpowell · 2005-01-29 02:47 UTC · Tiki post #3810 (reply to #3476)
> So in the fourty-eight thousand lines of IRC there are zero > instances of proper {nu'i gek term term /nu'u/ gik term term > /nu'u/}. > > I think the PEG rule would have to be something like this: > > gek-termset <- gek terms-gik-terms > > terms-gik-terms <- term (gik / terms-gik-terms) term > > That way the number of terms grabbed after the gik would be the > same as the number of terms between gek and gik.
Done.
mi klama nu'i ge le zarci le briju nu'u gi le zdani le ckule
and
mi klama ge le zarci le briju gi le zdani le ckule
now both parse in basically exactly the same way.
Note that nu'i...nu'u termsets are still perfectly valid.
-Robin
-- http://www.digitalkingdom.org/~rlpowell/ *** http://www.lojban.org/ Reason #237 To Learn Lojban: "Homonyms: Their Grate!"
Proud Supporter of the Singularity Institute - http://singinst.org/
