NeuroLab 1.2.1 Released

Released NeuroLab 1.2.1: http://neurolab.bitbucket.org

Neurocognitive Linguistics is an approach to linguistics developed by Sydney Lamb (http://www.ruf.rice.edu/~lamb/) which uses relational networks to model what the brain actually does when it handles language. You can read more about it at the LangBrain site (http://www.ruf.rice.edu/~lngbrain/main.htm) and Glottopedia (http://www.glottopedia.de/index.php/Neurocognitive_linguistics).

Neurocognitive Linguistics Lab (“NeuroLab” for short) is a program for Windows, Mac OS X and Linux that allows you to experiment with relational networks using a convenient GUI, and record the results of your experiments in tabular form.

IronMeta 2.1 Released

IronMeta version 2.1 has been released.

Version 2.1 contains some refactoring, miscellaneous bug fixes, as well as:

  • Better error handling and reporting.
  • Added IronMeta.Matcher.CharMatcher.Input() and IronMeta.Matcher.CharMatcher.Trimmed() for more convenient string handling.
  • Added min/max repeats syntax (e.g. 'a' {1, 3}).

IronMeta is an implementation of Alessandro Warth’s OMeta metaprogramming system in C#. It provides a packrat parser generator that generates parsers for Parsing Expression Grammars that operate on arbitrary streams of objects.

NeuroLab 1.2.0 Released

Version 1.2.0 of the Neurocognitive Linguistics Laboratory is now
available at http://neurolab.bitbucket.org

Highlights of this release include:

  • The primary new feature in this release is the “Grid Item”. This
    type of network item allows you to create a small “template” network,
    along with symmetrical connections to its top, bottom, and sides.
    Then the program generates a huge grid consisting of repeated
    instances of your template, with each instance connected to its
    neighbors via the edge connections. The overall topology of the grid
    is cylindrical; the sides wrap around, but the very top and bottom
    rows can connect to other network items.
  • A simple way to input and output text to a grid is provided via the
    “Text IO Item”. This is a network item that has 256 connections in
    and out; if you connect it to a grid item that has 256 horizontal
    repeats it will feed your text (in UTF-8 format) byte by byte as
    activation to the different grid repeats, and output results from the
    grid if one of its outputs is activated.
  • There are several UI improvements in version 1.2.0, including a
    palette of network items from which you can drag items into the
    network editing area.
  • The underlying network automaton is vastly sped up in version 1.2.0
    by using seqlocks instead of mutexes to manage memory consistency.

You can reply with questions or discussion about NeuroLab on this
mailing list. Please report bugs via the issue tracker:
https://bitbucket.org/kulibali/neurocogling/issues/new

Neurocognitive Linguistics is an approach to linguistics developed by
Sydney Lamb which uses relational
networks to model what the brain actually does when it handles
language. You can read more about it at the LangBrain site and Glottopedia.

Neurocognitive Linguistics Lab (“NeuroLab” for short) is a program for
Windows, Mac OS X and Linux that allows you to experiment with
relational networks using a convenient GUI, and record the results of
your experiments in tabular form.

Neurocognitive Linguistics Lab is Copyright (C) 2010,2011 Gordon
Tisher, and available under the terms of the BSD License.

IronMeta 2.0 Released

I have released version 2.0 of the IronMeta parser generator. IronMeta is an implementation of OMeta for C#. It generates Packrat PEG parsers that can recognize streams of any C# type and output any other type.

This version contains massive efficiency improvements, and an implementation of Sergio Medeiros’s algorithm for handling left recursion, which is simpler and more general than Warth et al‘s.

This algorithm can handle grammars of the form:

    A <- A a | B
    B <- B b | A | C
    C <- C c | B | d

This grammar would drive ordinary recursive-descent parsers to drink, but IronMeta handles it fine.