Received-SPF: pass (google.com: best guess record for domain of kali9putra@yahoo.com designates 66.94.237.209 as permitted sender) client-ip=66.94.237.209;
References: <CAOQ8VDV21KdHLncsr2yhqnneNpE-rbV8gcjdkXhjnPJyFrJP6A@mail.gmail.com> <518FC9B5.1090805@lojban.org> <CAOQ8VDX+4HDxrkUCF+YzBVeviExzeZFGvTODJwcc0ssCpU1+fg@mail.gmail.com> <CA+gZtgJOV_u8fWOryhuqu_r+PpoCuKYUxg+=YG0-H2oEMr2fVw@mail.gmail.com> <519276C9.9090604@lojban.org>
Message-ID: <1368555355.51683.YahooMailNeo@web184403.mail.bf1.yahoo.com>
Date: Tue, 14 May 2013 11:15:55 -0700 (PDT)
From: John E Clifford <kali9putra@yahoo.com>
Reply-To: lojban@googlegroups.com
Subject: Re: [lojban] Higher-precedence Grouping with {bo} Is Impossible for
 TAG Sentence Connectives
To: "lojban@googlegroups.com" <lojban@googlegroups.com>
In-Reply-To: <519276C9.9090604@lojban.org>
MIME-Version: 1.0
Precedence: list
Mailing-list: list lojban@googlegroups.com; contact lojban+owners@googlegroups.com
Sender: lojban@googlegroups.com
Content-Type: multipart/alternative; boundary="-6906265-1898919435-1368555355=:51683"
X-Spam_score: 0.0
X-Spam_score_int: 0
X-Spam_bar: /

---6906265-1898919435-1368555355=:51683
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

Or, in short, while Lojban could use full FOPL (augmented, to be sure) to c=
over every possible situation, the results are too ugly to even consider as=
 a language to speak or write.=A0 Hence, the project is to find the way to =
take away the maximum of the preciding features of Logic without losing the=
 precision.=A0 To some extent, this is trial and error, but, given that no =
significant changes have been made in the overall underlying logic, Lojban =
has done a rather good job of it.=A0 At some point we will get to the aesth=
etics of whether what we have is really a language, but the vector of progr=
ess has generally been that way.


________________________________
 From: "Bob LeChevalier, President and Founder - LLG" <lojbab@lojban.org>
To: lojban@googlegroups.com=20
Sent: Tuesday, May 14, 2013 12:39 PM
Subject: Re: [lojban] Higher-precedence Grouping with {bo} Is Impossible fo=
r TAG Sentence Connectives
=20

Ian Johnson wrote:
> {bo} is hackish, but its meaning is at least determined by the grammar.
> Grammatically the worst thing about it is failure to elide certain
> terminators. For example, if {I TAG BO} were instead {I TAG BOhOI} or
> something, {.i fi'o broda fe'u bo} could be {.i fi'o broda bo'oi} with
> the {fe'u} elided. This isn't that bad, however, all things said; I
> wouldn't say it's worth having its own cmavo. Incidentally, why is it
> that the PEG can't make {.i fi'o broda bo mi brode} parse correctly,
> whereas it can fix things like JOI connecting LE-sumti without {ku}?
>=20
> Evidently {ke} is actually even worse, since you can't coerce {GIhA KE
> SELBRI} to not make a gihe-kek. Why were these designed to be so
> awkwardly general purpose?

I'm not sure I understand your question.=A0 one point to note is that TLI L=
oglan grammar was originally designed with no formal parser analysis at all=
, and this history has driven the subsequent Lojban effort.

a) JCB discovered the works of Victor Yngve sometime in the 60s or early 70=
s, and therefore got the idea of codifying the grammar in a set of rules.=
=A0 He also conceived of several goals for this codified grammar which I do=
n't remember very well.=A0 But he was unable to achieve these goals, whatev=
er they were, though they were associated with encoding what JCB understood=
 as the "human grammar"
b} Around 1976-78, the effort changed to using a YACC LALR-1 grammar as a s=
tandard for codifying the grammar because several people knew how to use YA=
CC.=A0 Much of the grammar was encoded, but it seemed to be impossible to g=
et the "machine grammar" to parse things quite the same as the "human gramm=
ar" did.
c) The problem was solved around 1980, I believe by Jeff Prothero, then a s=
tudent at the University of Washington, to use elidable terminators to brac=
ket constructs, which elisions YACC would supply using its error processing=
.=A0 It took until around 1982-1983 to actually achieve a complete YACC gra=
mmar for the language, using error correction.

When we started redeveloping Lojban, the intent was to retain the Loglan gr=
ammar in its entirety, changing only the words.=A0 Thus we were bound by th=
e design limitations of the original language.=A0 JCB attempted to play cop=
yright games with the formal grammar (as he had with the words of the langu=
age), but he was on impossible legal ground given that so much of the work =
had been done by Prothero and others, along with known legal issues in copy=
righting a computer algorithm.

But we had reinvented the cmavo lexicon, and we wanted to include grammar c=
omponents for tense and MEX that JCB had never managed.=A0 Thus, initially =
with the help of Prothero and a guy named Jeff Taylor and others who knew Y=
ACC, I attempted to reimplement the YACC grammar from scratch, but not real=
ly trying to reinvent any wheels.=A0 In 1991, Cowan took over what I had do=
ne, and cleaned it up considerably, eventually achieving the baseline gramm=
ar listed in CLL (which is still the official grammar).=A0 But the grammar =
was still a YACC grammar, with all its limitations.

Attempts to create a PEG grammar remain unofficial, and frankly I've never =
looked at the PEG grammar and probably wouldn't understand it if I did.=A0 =
YACC was hard enough for me, and having learned the YACC grammar for Lojban=
, I never managed to fluently use the supposedly simpler E-BNF grammar (eve=
n though I had learned a couple of computer languages using BNF).

Thus the long answer to your question, as I understand it is that the gramm=
ar was always intended to be as general purpose as possible. Elidability of=
 terminators wasn't a high priority in general, though certain ones were de=
sirable; there was nothing more obnoxious that trying to figure out what wa=
s and was not terminated when you expressed a string like kukukeiku.=A0 (JC=
B's language used gu instead of ku, and thus it sounded a lot like baby tal=
k.=A0 Lojban with full terminators, is simply kuku.)

Some of the non-general purpose constructs arose because they couldn't get =
YACC to work with fully general constructs, or they required too much use o=
f obnoxious terminators.=A0 Hence the plethora of different families of log=
ical connectives, each linking a different type of construct.=A0 Those deci=
sions generally dated from the JCB era, though we added some new things tha=
t were connectable (such as relative clauses), and hence some new families,=
 most of which eventually went away (leaving for example zi'e which no long=
er is the basis for a family of logical connectives).=A0 We also abandoned =
the effort to impose a formal grammar on PA and UI compounds, so that there=
 are strings of each of those cmavo that are technically grammatical but ma=
ke no sense:
pi'epaime'ipipi'e.=A0 But for the most part, the fundamental language gramm=
ar remains that of JCB's pre-formal language, with elidable terminator cons=
tructs added where they could enable useful and yet syntactically unambigio=
us constructs.

lojbab
-- Bob LeChevalier=A0 =A0 lojbab@lojban.org=A0 =A0 www.lojban.org
President and Founder, The Logical Language Group, Inc.

-- You received this message because you are subscribed to the Google Group=
s "lojban" group.
To unsubscribe from this group and stop receiving emails from it, send an e=
mail to lojban+unsubscribe@googlegroups.com.
To post to this group, send email to lojban@googlegroups.com.
Visit this group at http://groups.google.com/group/lojban?hl=3Den.
For more options, visit https://groups.google.com/groups/opt_out.

--=20
You received this message because you are subscribed to the Google Groups "=
lojban" group.
To unsubscribe from this group and stop receiving emails from it, send an e=
mail to lojban+unsubscribe@googlegroups.com.
To post to this group, send email to lojban@googlegroups.com.
Visit this group at http://groups.google.com/group/lojban?hl=3Den.
For more options, visit https://groups.google.com/groups/opt_out.


---6906265-1898919435-1368555355=:51683
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<html><body><div style=3D"color:#000; background-color:#fff; font-family:ti=
mes new roman, new york, times, serif;font-size:12pt">Or, in short, while L=
ojban could use full FOPL (augmented, to be sure) to cover every possible s=
ituation, the results are too ugly to even consider as a language to speak =
or write.&nbsp; Hence, the project is to find the way to take away the maxi=
mum of the preciding features of Logic without losing the precision.&nbsp; =
To some extent, this is trial and error, but, given that no significant cha=
nges have been made in the overall underlying logic, Lojban has done a rath=
er good job of it.&nbsp; At some point we will get to the aesthetics of whe=
ther what we have is really a language, but the vector of progress has gene=
rally been that way.<br><div><span><br></span></div><div><br></div>  <div s=
tyle=3D"font-family: times new roman, new york, times, serif; font-size: 12=
pt;"> <div style=3D"font-family: times new roman, new york, times, serif;
 font-size: 12pt;"> <div dir=3D"ltr"> <hr size=3D"1">  <font face=3D"Arial"=
 size=3D"2"> <b><span style=3D"font-weight:bold;">From:</span></b> "Bob LeC=
hevalier, President and Founder - LLG" &lt;lojbab@lojban.org&gt;<br> <b><sp=
an style=3D"font-weight: bold;">To:</span></b> lojban@googlegroups.com <br>=
 <b><span style=3D"font-weight: bold;">Sent:</span></b> Tuesday, May 14, 20=
13 12:39 PM<br> <b><span style=3D"font-weight: bold;">Subject:</span></b> R=
e: [lojban] Higher-precedence Grouping with {bo} Is Impossible for TAG Sent=
ence Connectives<br> </font> </div> <div class=3D"y_msg_container"><br>
Ian Johnson wrote:<br>&gt; {bo} is hackish, but its meaning is at least det=
ermined by the grammar.<br>&gt; Grammatically the worst thing about it is f=
ailure to elide certain<br>&gt; terminators. For example, if {I TAG BO} wer=
e instead {I TAG BOhOI} or<br>&gt; something, {.i fi'o broda fe'u bo} could=
 be {.i fi'o broda bo'oi} with<br>&gt; the {fe'u} elided. This isn't that b=
ad, however, all things said; I<br>&gt; wouldn't say it's worth having its =
own cmavo. Incidentally, why is it<br>&gt; that the PEG can't make {.i fi'o=
 broda bo mi brode} parse correctly,<br>&gt; whereas it can fix things like=
 JOI connecting LE-sumti without {ku}?<br>&gt; <br>&gt; Evidently {ke} is a=
ctually even worse, since you can't coerce {GIhA KE<br>&gt; SELBRI} to not =
make a gihe-kek. Why were these designed to be so<br>&gt; awkwardly general=
 purpose?<br><br>I'm not sure I understand your question.&nbsp; one point t=
o note is that TLI Loglan grammar was originally designed with no
 formal parser analysis at all, and this history has driven the subsequent =
Lojban effort.<br><br>a) JCB discovered the works of Victor Yngve sometime =
in the 60s or early 70s, and therefore got the idea of codifying the gramma=
r in a set of rules.&nbsp; He also conceived of several goals for this codi=
fied grammar which I don't remember very well.&nbsp; But he was unable to a=
chieve these goals, whatever they were, though they were associated with en=
coding what JCB understood as the "human grammar"<br>b} Around 1976-78, the=
 effort changed to using a YACC LALR-1 grammar as a standard for codifying =
the grammar because several people knew how to use YACC.&nbsp; Much of the =
grammar was encoded, but it seemed to be impossible to get the "machine gra=
mmar" to parse things quite the same as the "human grammar" did.<br>c) The =
problem was solved around 1980, I believe by Jeff Prothero, then a student =
at the University of Washington, to use elidable terminators to
 bracket constructs, which elisions YACC would supply using its error proce=
ssing.&nbsp; It took until around 1982-1983 to actually achieve a complete =
YACC grammar for the language, using error correction.<br><br>When we start=
ed redeveloping Lojban, the intent was to retain the Loglan grammar in its =
entirety, changing only the words.&nbsp; Thus we were bound by the design l=
imitations of the original language.&nbsp; JCB attempted to play copyright =
games with the formal grammar (as he had with the words of the language), b=
ut he was on impossible legal ground given that so much of the work had bee=
n done by Prothero and others, along with known legal issues in copyrightin=
g a computer algorithm.<br><br>But we had reinvented the cmavo lexicon, and=
 we wanted to include grammar components for tense and MEX that JCB had nev=
er managed.&nbsp; Thus, initially with the help of Prothero and a guy named=
 Jeff Taylor and others who knew YACC, I attempted to reimplement
 the YACC grammar from scratch, but not really trying to reinvent any wheel=
s.&nbsp; In 1991, Cowan took over what I had done, and cleaned it up consid=
erably, eventually achieving the baseline grammar listed in CLL (which is s=
till the official grammar).&nbsp; But the grammar was still a YACC grammar,=
 with all its limitations.<br><br>Attempts to create a PEG grammar remain u=
nofficial, and frankly I've never looked at the PEG grammar and probably wo=
uldn't understand it if I did.&nbsp; YACC was hard enough for me, and havin=
g learned the YACC grammar for Lojban, I never managed to fluently use the =
supposedly simpler E-BNF grammar (even though I had learned a couple of com=
puter languages using BNF).<br><br>Thus the long answer to your question, a=
s I understand it is that the grammar was always intended to be as general =
purpose as possible. Elidability of terminators wasn't a high priority in g=
eneral, though certain ones were desirable; there was nothing more
 obnoxious that trying to figure out what was and was not terminated when y=
ou expressed a string like kukukeiku.&nbsp; (JCB's language used gu instead=
 of ku, and thus it sounded a lot like baby talk.&nbsp; Lojban with full te=
rminators, is simply kuku.)<br><br>Some of the non-general purpose construc=
ts arose because they couldn't get YACC to work with fully general construc=
ts, or they required too much use of obnoxious terminators.&nbsp; Hence the=
 plethora of different families of logical connectives, each linking a diff=
erent type of construct.&nbsp; Those decisions generally dated from the JCB=
 era, though we added some new things that were connectable (such as relati=
ve clauses), and hence some new families, most of which eventually went awa=
y (leaving for example zi'e which no longer is the basis for a family of lo=
gical connectives).&nbsp; We also abandoned the effort to impose a formal g=
rammar on PA and UI compounds, so that there are strings of each of
 those cmavo that are technically grammatical but make no sense:<br>pi'epai=
me'ipipi'e.&nbsp; But for the most part, the fundamental language grammar r=
emains that of JCB's pre-formal language, with elidable terminator construc=
ts added where they could enable useful and yet syntactically unambigious c=
onstructs.<br><br>lojbab<br>-- Bob LeChevalier&nbsp; &nbsp; <a ymailto=3D"m=
ailto:lojbab@lojban.org" href=3D"mailto:lojbab@lojban.org">lojbab@lojban.or=
g</a>&nbsp; &nbsp; www.lojban.org<br>President and Founder, The Logical Lan=
guage Group, Inc.<br><br>-- You received this message because you are subsc=
ribed to the Google Groups "lojban" group.<br>To unsubscribe from this grou=
p and stop receiving emails from it, send an email to lojban+<a ymailto=3D"=
mailto:unsubscribe@googlegroups.com" href=3D"mailto:unsubscribe@googlegroup=
s.com">unsubscribe@googlegroups.com</a>.<br>To post to this group, send ema=
il to <a ymailto=3D"mailto:lojban@googlegroups.com"
 href=3D"mailto:lojban@googlegroups.com">lojban@googlegroups.com</a>.<br>Vi=
sit this group at <a href=3D"http://groups.google.com/group/lojban?hl=3Den"=
 target=3D"_blank">http://groups.google.com/group/lojban?hl=3Den</a>.<br>Fo=
r more options, visit <a href=3D"https://groups.google.com/groups/opt_out" =
target=3D"_blank">https://groups.google.com/groups/opt_out</a>.<br><br><br>=
<br><br></div> </div> </div>  </div></body></html>

<p></p>

-- <br />
You received this message because you are subscribed to the Google Groups &=
quot;lojban&quot; group.<br />
To unsubscribe from this group and stop receiving emails from it, send an e=
mail to lojban+unsubscribe@googlegroups.com.<br />
To post to this group, send email to lojban@googlegroups.com.<br />
Visit this group at <a href=3D"http://groups.google.com/group/lojban?hl=3De=
n">http://groups.google.com/group/lojban?hl=3Den</a>.<br />
For more options, visit <a href=3D"https://groups.google.com/groups/opt_out=
">https://groups.google.com/groups/opt_out</a>.<br />
&nbsp;<br />
&nbsp;<br />

---6906265-1898919435-1368555355=:51683--