Re: Further Lojban->Prolog: relative clauses (Repost)

[cowan@snark.thyrsus.com (John Cowan)]

> I'm about
> to implement the lo/le distinction (well, next week: done enough for this),
> and would welcome any opinions.

I think that you should send "le" sumti through undisturbed:

	le gerku cu xekri -> xekri(le_gerku).

and then insert a subsidiary fact

	gerku(le_gerku).

If the latter turns out to be false in the actual circumstances, it can be
retracted without loss of modularity.  This is what I did in the example
published in JL, which had facts like "sofa(sofa1)", where sofa1 = "le sfofa".

Where do those unspeakable (:-)) variables like FIODG and FIPZL come from?

-- 
John Cowan	cowan@snark.thyrsus.com		...!uunet!lock60!snark!cowan
			e'osai ko sarji la lojban.