Received-SPF: pass (google.com: domain of almikes@aol.com designates 205.188.109.204 as permitted sender) client-ip=205.188.109.204;
Date: Fri, 10 Jan 2014 18:58:40 +0100
From: Wuzzy <almikes@aol.com>
To: lojban@googlegroups.com
Subject: [lojban] Quoting with ZOI can be ambigious
Message-ID: <20140110185840.22249c17@aol.com>
Mime-Version: 1.0
Reply-To: lojban@googlegroups.com
Precedence: list
Mailing-list: list lojban@googlegroups.com; contact lojban+owners@googlegroups.com
Sender: lojban@googlegroups.com
Content-Type: text/plain; charset=windows-1252
Content-Transfer-Encoding: quoted-printable
X-Spam_score: 0.0
X-Spam_score_int: 0
X-Spam_bar: /

TL;DR: Quoting with ZOI can be ambigious. I point out the flaws and show
how the Reference Grammar does not help here. Then I propose a change=97or
more precisely: a clarification=97to the grammar which hopefully fixes
the flaws.

Full text:
I noticed that you run into trouble when your ZOI quote has trailing
or leading whitespace, or trailing or leading periods. Also you run
into trouble if you want to decode a ZOI quote because you can=92t tell
if the text truly has trailing/leading whitespace/periods or not.

Just for extra clarity: In this e-mail I will make use of the so-called
logical quotation style so you know exactly what I=92m talking about and
because the British quotation style is just damn awful.
Logical quotation style is basically this: All characters which are
between the quotation marks are quoted. No magic! ;-)
<https://en.wikipedia.org/wiki/Quotation_mark#Punctuation>

Let=92s go!
Just look at this:
What does =93zoi gy..gy=94 quote?
I could think of these answers:
(1) =93=94 (empty string (string =3D sequence of characters))
(2) =93..=94 (two periods)
(3) =93.=94 (one period)

All three answers seem plausible to me. Remember: according to the
Reference Grammar, all periods are optional:
=93Technically, the period is an optional reminder to the reader of a
mandatory pause that is dictated by the rules of the language;
because these rules are unambiguous, a missing period can be inferred
from otherwise correct text. Periods are included only as an aid to
the reader.=94=97Reference Grammar, chapter 3, section 3.

(1) would be true if both periods are considered to be part of the
delimeter, therefore, they are part of the Lojban text and not of the
quoted text.
(2) would be true if the =93pause periods=94 of Lojban have
been ommitted. Therefore, the periods must be part of the quoted text.
(3) is a weird case which could be true if one of the =93pause periods=94
has been ommited, but not the other one. Therefore, there is one period
left which is therefore part of the quoted text.

All three possibilities are plausible to me under the rules. But
under these rules, I can=92t tell which is the correct one.

There is another problem, with whitespace:
=93zoi gy gy=94 =3D =93=94 or =93 =94?

Here are more examples (look closely at spaces and periods):
=93zoi gy. .gy=94 =3D =93 =94, =93=94, =93. =94, =93.=94, =93. =94?
=93zoi gy.Peter Smith.gy=94 =3D =93Peter Smith=94, =93.Peter Smith=94, =93P=
eter Smith.=94,
                          =93.Peter Smith.=94?
=93zoi gy. Peter Smith .gy=94 =3D =93Peter Smith=94, =93 Peter Smith=94,
                            =93Peter Smith =94, =93 Peter Smith =94?
                            With or without periods?
=93zoi gy.pdf.gy=94 =3D =93pdf=94, =93.pdf=94, =93.pdf.=94, =93pdf.=94?
=93zoi gy..pdf.gy=94 =3D =93pdf=94, =93.pdf=94, =93.pdf.=94, =93pdf.=94, =
=93..pdf=94 or even
                   =93..pdf.=94?

Now here comes the real problem:
The Reference Grammar does NOT have an answer,
so any of these answers I suggested could be true.

Here is what the grammar chapter of the Reference Grammar says:
=93a. If the Lojban word =91zoi=92 (selma'o ZOI) is identified, take the
following Lojban word (which should be end delimited with a pause for
separation from the following non-Lojban text) as an opening delimiter.
Treat all text following that delimiter, until that delimiter recurs
after a pause, as grammatically a single token (labelled =91anything_699=92
in this grammar). There is no need for processing within this text
except as necessary to find the closing delimiter.=94
This text does not explicitly state wheather to write down the periods.
I only know all periods are optional, leading to the situation
I pointed out before (pun intended).
The section on ZO and ZOI is of no help either.

The point I want to make here is that the Reference Grammar does NOT say
which possiblity to choose. So please don=92t argue about what you think
would be the correct choice here. There is no correct choice, the
Reference Grammer does not define these borderline cases at all. Which
means: It is impossible to reliably quote any string with whitespace or
a period at the beginning or end.

Since the Reference Grammar is disappointing me here, there must be a
clear definition for these cases.

I therefore suggest the following rule changes to be applied to the
Reference Grammar:
- Written Lojban:
  - periods are MANDATORY at the following places: Right after the
    opening delimeter word and right before the closing delimeter word.
    The mandatory period is considered part of the delimeter.
  - The period of the opening and closing delimeter may not be merged
    into one period.
  And just to be sure:
  - The quote starts after the period of the opening delimeter.
  - The quote ends before the period of the closing delimeter.
  - Everything, even whitespace between those two periods is considered
    quoted.
- Spoken Lojban:
  - The sound of the delimeter word may not appear in the quoted text,
    no matter if there=92s a glottal stop before or after it.

Here are some valid and invalid examples under these proposed rules:
=93zoi gy..gy=94			=3D =93=94. The opening delmeter is =93gy.=94,
                                the closing delimeter is =93.gy=94.
=93zoi gy.gy=94			is an error! 2nd delimeter missing
=93zoi gygy=94			is an error! 1st delimeter missing
=93zoi gy. .gy=94			=3D =93 =94
=93zoi gy..pdf.gy=94		=3D =93.pdf=94
=93zoi gy.pdf.gy=94			=3D =93pdf=94
=93zoi gy.Peter Smith.gy.=94	=3D =93Peter Smith=94
=93zoi gy.Peter Smith.gy=94		=3D =93Peter Smith=94 (note the period
				AFTER the closing delimeter is still
				optional)
=93zoi gy.Wuzzy eats Pizza..gy.=94	=3D =93Wuzzy eats Pizza.=94
=93zoi gy. Peter Smith .gy.=94 	=3D =93 Peter Smith =94
=93zoi lo..lo.lo=94			=3D =93=94 followed by an article
=93zoi gy...gy=94			=3D =93.=94
=93zoi gy. . .gy=94			=3D =93 . =94
=93zoi gy.....gy=94			=3D =93...=94

Here are some borderline cases:
=93zoi gy.gy.gy=94 violates the spoken Lojban rule iff the 2nd
               =93gy=94 has the sound of Lojban=92s =93gy=94.
=93zoi gy.gy.gy=94 is valid and is =93gy=94 iff the 2nd =93gy=94 has NOT th=
e sound
               of Lojban=92s =93gy=94.
=93zoi gy.gyrations.gy=94: same as previous ones.

Note the last three cases are borderline cases and should be used with
care. The pronounciation of the quoted text must be known for this to
work. While these may be valid, at best such borderline cases are
avoided by choosing another delimeter. Because you aren=92t going to
tortue the recipient, aren=92t you? ;-)

That=92s it.
I have made some research and it seems nobody has pointed out these
problems before. Anyhow, this looks like serious trouble and
certainly must be fixed.
It is embarrassing that I am able to quote more precisely with English
than with Lojban. ;-) ;-) ;-)

Please say wheather you agree with me in general and wheather you think
the proposed change is okay. If my proposal or reasoning has flaws in
is, please point them out.

Greetings!
--=20
Wuzzy
XMPP: Wuzzy2@jabber.ccc.de
E-Mail: wuzzy2@mail.ru

--=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.
For more options, visit https://groups.google.com/groups/opt_out.