From seidensticker@msn.com Mon Feb 12 13:55:39 2001 Return-Path: X-Sender: Seidensticker@msn.com X-Apparently-To: lojban@yahoogroups.com Received: (EGP: mail-7_0_3); 12 Feb 2001 21:55:30 -0000 Received: (qmail 94281 invoked from network); 12 Feb 2001 21:55:29 -0000 Received: from unknown (10.1.10.26) by l10.egroups.com with QMQP; 12 Feb 2001 21:55:29 -0000 Received: from unknown (HELO bilbo.w-link.net) (206.98.114.20) by mta1 with SMTP; 12 Feb 2001 21:55:29 -0000 Received: from bobsprimary (dhcp79-dsl-gte3.w-link.net [206.129.86.79]) by bilbo.w-link.net (8.9.3/8.9.3) with SMTP id NAA20405 for ; Mon, 12 Feb 2001 13:53:25 -0800 (PST) Message-ID: <002601c0953e$8be62280$4f5681ce@wlink.net> To: Subject: Re: Punctuation Date: Mon, 12 Feb 2001 13:55:42 -0800 Organization: Microsoft Corporation MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_0023_01C094FB.7D1368A0" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.50.4522.1200 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4522.1200 From: "seidensticker" ------=_NextPart_000_0023_01C094FB.7D1368A0 Content-Type: text/plain; charset="Windows-1252" Content-Transfer-Encoding: quoted-printable --- In lojban@y..., John Cowan wrote: > Invent Yourself wrote: >=20 > > Getting jbofi'e to recognize weirder punctuation is trivial, simply run > > the text through the most simple of search/replace functions. > >=20 > > Which would you like, though? If we use "" for lu/li'u, what about > > lo'u/lu'e and la'o gy.? >=20 > A recommended practice is to place the punctuation mark in front of the > cmavo, thus: (to, )toi, "lu, "li'u, "lo'u, "le'u, ?xu, !ba'e, etc. etc. > In that way the punc mark need not be mapped to just one sumti. I had been thinking along Tengwar lines of accent-like marks over the paren= theses that would let the reader more easily find the corresponding closing= mark to each opening mark. Or, some 2-character marks might be something = like: `( )=B4 (- -) (=AF =AF) (=B0 =B0)=20=20 ....But perhaps this is too elaborate. John's recommended practice sounds = like a good approach -- it's optional, any parser could simply ignore these= punctuation marks, it doesn't require new characters, and it helps the rea= der. IMO, this last part is not simply an aid to novices like me. Skilled= writers in natural languages, too, use their more limited tools (such as p= arentheses, dashes, commas, and careful sentence construction) to minimize = the reader's mental effort. Searching through the cmavo list for "end," it looks like there are a coupl= e of dozen open/close grouping words. Which ones get which punctuation? O= ne approach might be to use: - quote for the quote wrappers lu", "li'u, lo'u", and "le'u (or, perhaps = use the =AB/=BB marks: lu=AB, =BBli'u, etc.) - parentheses for to( and )toi - [square] and {curly} brackets for the most common remaining wrappers - ...leaving for all the rest As an experiment, I put the punctuation on the _inside_ of the cmavo instea= d of on the left. Obviously, the solo ?xu and !ba'e don't have any inside. ------=_NextPart_000_0023_01C094FB.7D1368A0 Content-Type: text/html; charset="Windows-1252" Content-Transfer-Encoding: quoted-printable

--- In lojban@y..., John Cowan <jcowan@r...> wrote:
> Invent Yourself= =20 wrote:
>
> > Getting jbofi'e to recognize weirder punctuati= on is=20 trivial, simply run
> > the text through the most simple of=20 search/replace functions.
> >
> > Which would you like,= =20 though? If we use "" for lu/li'u, what about
> > lo'u/lu'e and la'= o=20 gy.?
>
> A recommended practice is to place the punctuation ma= rk in=20 front of the
> cmavo, thus: (to, )toi, "lu, "li'u, "lo'u, "le'u= ,=20 ?xu, !ba'e, etc. etc.
> In that way the punc mark need not be mapped = to=20 just one sumti.

I had been thinking along Tengwar lines of accent-li= ke=20 marks over the parentheses that would let the reader more easily find the=20 corresponding closing mark to each opening mark. Or, some 2-character= =20 marks might be something like:
`( )=B4 (- -) (=AF =AF) = (=B0=20 =B0)

....But perhaps this is too elaborate. John's= =20 recommended practice sounds like a good approach -- it's optional, any pars= er=20 could simply ignore these punctuation marks, it doesn't require new charact= ers,=20 and it helps the reader. IMO, this last part is not simply an aid to= =20 novices like me. Skilled writers in natural languages, too, use their= more=20 limited tools (such as parentheses, dashes, commas, and careful sentence=20 construction) to minimize the reader's mental effort.

Searching thro= ugh=20 the cmavo list for "end," it looks like there are a couple of dozen open/cl= ose=20 grouping words. Which ones get which punctuation? One approach = might=20 be to use:
- quote for the quote wrappers lu", "li'u, lo'u", and = "le'u=20 (or, perhaps use the =AB/=BB marks: lu=AB, =BBli'u, etc.)
- paren= theses for=20 to( and )toi
- [square] and {curly} brackets for the most common= =20 remaining wrappers
- ...leaving </> for all the rest
As=20 an experiment, I put the punctuation on the _inside_ of the cmavo instead o= f on=20 the left. Obviously, the solo ?xu and !ba'e don't have any=20 inside.

------=_NextPart_000_0023_01C094FB.7D1368A0--