Envelope-to: rlpowell@digitalkingdom.org
Delivery-date: Thu, 06 Jul 2006 07:17:46 -0700
Date: Thu, 6 Jul 2006 16:17:03 +0200 (CEST)
From: Arnt Richard Johansen <arj@nvg.org>
To: Jbovlaste <jbovlaste@lojban.org>
Subject: [jbovlaste] Re: drunk
In-Reply-To: <737890611.20060706204301@mail.ru>
Message-ID: <Pine.LNX.4.60.0607061547500.23471@hagbart.nvg.ntnu.no>
References: <1833870320.20060706195811@mail.ru> <Pine.LNX.4.60.0607061533180.23471@hagbart.nvg.ntnu.no> <737890611.20060706204301@mail.ru>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Sender: jbovlaste-bounce@lojban.org
Errors-to: jbovlaste-bounce@lojban.org
Precedence: bulk
Reply-to: jbovlaste@lojban.org
Content-Length: 2353
Lines: 53

Not a stupid question at all, Yanis. The fact that we appear to be simply 
discarding an old dictionary and doing all the work all over again, is a 
situation that demands an explanation. I suppose we have been relying to 
much on unwritten tribal knowledge in the process, which makes things 
difficult for enthusiastic new volunteers like you.

On Thu, 6 Jul 2006, Yanis Batura wrote:

> On 06.07.2006, 20:37, Arnt Richard Johansen wrote:
>
>
>> Noralujv.txt has recorded at least three suggestions for this.
>
> Isn't NORALUJV.txt present in jbovlaste??

Yes and no. All entries have been imported into Jbovlaste, but have by 
default no votes. Hence, they don't appear as search results.

It is possible to vote specific definitions into becoming live, but in 
most cases this is not recommended. This is because the definitions have 
all been automatically generated from the place structure codes in 
noralujv.txt, and are for the most part difficult or impossible to 
understand, or simply wrong.

> Why?

There are many reasons. Most can be boiled down to the fact that 
noralujv.txt simply isn't good enough. Noralujv is a fairly comprehensive 
catalogue of early usage, but it was never intended as an official 
dictionary in itself, merely as an aid in creating one. Also, it has not 
been updated for at least five years.

Also some of the lujvo *themselves* are very obviously bad coinings, and 
should therefore not be used. Since Jbovlaste aspires to be the 
authoritative source of Correct Lojban Words(tm), we can and should 
exclude those words that are not good enough.

The final reason we can't use noralujv.txt directly, is the format of the 
place structure definitions. They consist of mere pointers to gismu 
places. When the work with Jbovlaste started, it was decided that this 
format is simply too terse. It makes lujvo definitions stand out too much 
from gismu definitions, making lujvo seem to be compositionally derived 
from the source gismu -- which is wrong. Also, oblique places often have 
a slightly different meaning than that of the corresponding place in the 
source gismu, which needs to be recorded in the definition.

-- 
Arnt Richard Johansen                                http://arj.nvg.org/
Confusion among -ate ~ -ant pairs is even more prominate, since both
are legitimant suffixes. --Adam Albright