Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.compilers > #385

Public statement about Meta-S and The Grammar Forge

Show all headers | View raw

Hello.

[ I ask John's indulgence with this bit of administratrivia. I am
posting this in comp.compilers because this is where I have posted
about Meta-S over the years, and this is where the other parser
generator authors also seem to congregate. This forum is also highly
indexed, searchable, and is moderated. Thus, I figure this is where
various authorities who discuss parser generators come to be informed
"from the source". - qtj]


I have noticed various references and public speculation about the
Meta-S parsing system and The Grammar Forge. I have also received
private queries about the technical and legal status of the system.

* Rumors of my death are greatly exaggerated.
* The Meta-S parsing system is alive and well and under daily development.
* For a time in early the early 2000's, I was partnered with SandStone
(VisualParse++) on a resale agreement. That arrangement dissolved on
very good terms some years ago.
* For a time, I was partnered with Thothic Technology Partners, LLC. I
reacquired the IP on good terms and am still partnered with TTP.
* I own and continue to develop the system, and am partnered with
Black Dog Research, LLC. All business development and licensing is
through BDR -- I am the technical R&D lead, doing business as Q
Agamemnon Technologies.
* No other party is authorized to be distributing the system, where
here "no other party" means quite definitively: "NO OTHER PARTY."

Although I mention some technical aspects of the parser publicly, much
of what Meta-S is currently capable of is under strict non-disclosure
for IP protection reasons. The last definitive public disclosures are
contained in the book Adapting to Babel. To the best of my knowledge,
the last definitive in-depth third-party citations can be found in:

Dick Grune and Ceriel J.H. Jacobs, Parsing Techniques--A Practical
Guide, (2nd ed.), Springer Verlag, (In Press) 2006.

Some technical nuts and bolts disclosures:

* The system is an adaptive parser. I had a large part in this
Wikipedia article, which pretty much sums up what that implies:

http://en.wikipedia.org/wiki/Adaptive_grammar

* The system uses the adaptive(k) algorithm. (Essentially an adaptive
predictive infinite look ahead system.) Since the exact constraints of
this algorithm (outlined in Adapting to Babel) are specific to Meta-S,
I chose not to rehash some variant of LL or whathaveyou, and asked a
third party to coin a term.

* The formal underpinnings of Meta-S are known as the Meta-S Calculus.
The quick summary of this is "adaptive set theory."

* As discussed in Grune et al (ibid.), although Meta-S is capable of
accepting Type 0 languages, in practice it does so in O(n^m) time
where here m < 3. Grammars have been written to accept Collatz
sequences, Fibonacci, Catalan numbers, RNA pseudoknots, English, C++,
C#, et cetera ad nauseam, and never in worse than O(n^3), and often in
linear time.

* The system has a visual development environment (The Grammar Forge)
that runs on Windows systems (7, Vista, XP), but the parsing engine is
more or less portable to any system with a C++ compiler that handles
standard C++. It has been compiled into 64-bit and runs just fine. It
also can handle UNICODE.

* Although adaptive, the system can be used to generate C++ hosted
grammars that are compiled statically as well.

* As already mentioned, there are features that cannot be discussed
for IP protection reasons.

Any specific questions are best directed either to:

http://blackdogresearch.com/

or to me at thothic.quinn@gmail.com.


Cheers,
Quinn Tyler Jackson

Back to comp.compilers | Previous | Next | Find similar

Thread

Public statement about Meta-S and The Grammar Forge Quinn Tyler Jackson <quinn_jackson2004@yahoo.ca> - 2011-12-06 07:55 -0800

csiph-web