Re: [lojban] slinku'i test

[Richard Curnow <richard@rrbcurnow.freeuk.com>]

On Mon, Jul 31, 2000 at 10:58:10PM -0400, Pierre Abbat wrote:
> On Mon, 31 Jul 2000, John Cowan wrote:
> >Pierre Abbat wrote:
> >
> >> Why is the slinku'i test called that, and what's a slinku'i?
> >
> >"Slinku'i" is a hypothetical fu'ivla that fails the slinku'i test,
> >because "pa slinku'i" would be taken as the lujvo "pas+lin+ku'i".
> >Neither the fu'ivla nor the lujvo has any other meaning.
> >The name goes back to Loglan days.
> 
> I ran jbofi'e on it and got robe-chain-caretaker. Chain-robe-caretaker

It's fair to say that handling of fu'ivla is not a strong area of
jbofi'e.  As far as I can remember, I tried to have a reasonable go at
breaking a type-3 fu'ivla up into the 'classifier' and the 'proper noun'
bits.  It just tries to split type 4 words as though they are lujvo,
giving up when it gets left with a nonsensical suffix.

jbofi'e parses "paslinku'i" as a single lujvo as Pierre reports.
However, it currently parses "pa slinku'i" as 2 words, the second
failing to completely decode because it's not a valid lujvo.  John's
comment indicates this should be grouped as 1 word - maybe the
tokenising algorithm needs another revisit...

Richard

-- 
----------------------------------------------------------------------
Richard P. Curnow                                       rpc@myself.com
Weston-super-Mare
United Kingdom                            http://go.to/richard.curnow/