MeatballWiki | RecentChanges | Random Page | Indices | Categories

This page may turn out to have just a temporary existence. It was created to hold a bit of material that needed to be removed from another page, but if it evolves towards some other role, that's quite acceptable. -- HansWobbe

sniped from PatternTheory

Hans, I think the main point of such work is to find data structures that allow for representations of meaning and their transformation. But, it's no problem to store any information in e. g. (1) one large text, XML or not (2) a tree of files or (3) an array of arrays or (4) a LISP list of lists or (..) and even define conversions. Apart from syntax and performance, they are all equivalent. Because of their omnipotence they are also useless as a conceptual tool. What I have found recently, following the higher level Arcle concept (http://www.communitywiki.org/en/ArcleModelling), is a nicely fitting data structure that I named "symbol meaning array tree". It has a nice small number of restrictions that suggest a certain form of representations and seem to lead to KI-solutions.

-- HelmutLeitner

Helmut, Thank you for this reference. Unfortunately, I have to set aside thinking about this further until I return from my trip in the fall. -- HansWobbe

Helmut due to a typo or a deliberately set low "hanging fruit" as to stimulate activation of otherwise lazy wikizens here REXEXP didn't point to regular Expressions, which in some literature also appear as regular patterns, so I corrected the typo and added the Google prefix in you last posting. Compared to the extremely complex Alexandrian patterns, which may currently only be "recognized" by human "recognizers" (by way of intuition) the regular expressions over finite alphabets (also called "regular events") are recognized by certain "finite state machines". So Hans' posting implicitely poses the question: What kind of grammar (in the sense of Chomsky) may be appropriate to describe Alexandrians patterns. I know that it is currently a daunting task to give such a grammar (contextfree in BNF :-) as long as we have not realized the visual perceptor automaton, that can emulate the human visual capabilities. From the theory of automata we know that for each regular pattern there exists a "recognizer" and for each "recogizer" there exists a corresponding grammar that describes the recognizing capabilities of the recognizer, i.e. in this case .. that can define the regular event, the automaton is recognizing. -- FridemarPache

Hans replies to Lion, and acknowledges Helmut's request for "more..." + ...

Yes, Lion, I am indeed continuing to develop some of the concepts I tried to articulate as "StringsTheory?" (in a teasing attempt to explain them to some of my associates who are engaged in studying the "string theory" of modern physics). In fact, we've published a bit of the work as part of the conceptual work using the term "TheArray?" to describe an EnDimensionalRaggedArray? (my APL roots are showing, again) that consists of all of the (bit)"strings" contained in all of the computers that can be accessed via the InterNet.

Helmut, I will try to clarify some of my thoughts as quickly as possible, but I cannot promise that this will happen before I have to start traveling in late August.

Fridemar, Thanks for the expansions that you've added. I found them quite provocative of some further thoughts.

-- HansWobbe


MeatballWiki | RecentChanges | Random Page | Indices | Categories
Edit text of this page | View other revisions