Lojban and the Turing machine

[David Brookshire Conner <nellardo@concentric.net>]

PILCH Hartmut writes:
 > From: PILCH Hartmut <phm@a2e.de>

 > Lojban texts are parsable, and it should at least be possible to formulate
 > computer programs in some lojban-compatible syntax.

Sure - syntax is no big deal.

 > I wonder if Lojban texts themselves could be considered as programs, just
 > as postscript documents can be.

Yes, but like you do with Postscript, you need a machine to run them,
even an abstract one.

 > Lojban parsing is very similar to SGML parsing, but SGML, unlike
 > Postscript, isn't a full-fledged programming language
 > 
 > What is needed to make Lojban "turing-complete"?

A (virtual) machine that interprets enough of the semantics of lojban
to get it over the Turing threshold. Something that understood lojban
math would put it over the hump.

Interestingly, interpreting variable assignment in the expected way is 
enough - two integers is enough memory to provide Turing equivalence.

Without such a machine for the semantics, i.e., a syntax-only lojban,
lojban is not turing equivalent, as it is represented by a
context-free grammar (the reference grammar).

 > I am just looking for more arguments to further confuse the debate about
 > software patentability, see also
 > 
 > 	http://swpat.ffii.org/lojban/

I'm missing something - how does Turing equivalence confuse the issue
of software patentability? Are you suggesting that Turing equivalence
be a metric for determining whether a language can be patented? This
would suggest that Bjarne Stroustrop could have patented C++, which
seems a little silly.

Copyrighting a reference grammar, sure (but writing something that
conforms to that grammar would not violate the copyright).  You can
copyright a dictionary, but people could still use the words defined
in it freely ("small passages" and such), else Merriam-Webster would
have a doozy of a lawsuit against all sorts of people. You could even
copyright a word list, but using individual words from it would still
be legit. At least that's how I understand copyrights to work - being
an artist and a writer, I pay some attention to this. John Cowan may
have other insights based on the events of the Loglan/lojban split.

Trademarking a name used in business (i.e., trade), sure. This is part
of the reason Microsoft can trademark "Visual Basic", even though
Basic is a widely available language - the trademark covers the
product, technically.

But patents are broader beasts. I can make a copy of something
copywritten for personal use. I cannot (legally) make *any*
implementation of something patented, even if it never sees the light
of day. Being a software engineer, I pay some attention to this, as
well (and have technically violated the RSA patent like oh ten years
ago when I took an algorithms class where RSA was an assignment).

The hairy question of software patents is when does software change
from "speech" (covered by copyright law) to "implementation" (and thus
covered by patent law)?


Brook

---------
RAM disk is not an installation procedure.

---------
Fancy. Myth. Magic.
http://www.concentric.net/~nellardo/