Natural Language Processing / Information Retrieval Software Repository

Search SoC (via Google):

Solaris 5.8 version

The Linux 2.4.22 version (aye) is here.

Last updated: $Id: README.html,v 1.14 2005/06/30 02:11:40 rpnlpir Exp rpnlpir $

This directory and account holds centralized software and tools for natural language processing (NLP) and information retrieval (IR) research and teaching at the School of Computing at the National University of Singapore. The account is hosted off of sf3 such that students and researchers will be able to get at these tools. Access is granted to all, however, if you'd like to provide and/or install tools, you must first email the administrators (rpnlpir@comp...).

The tools here are compiled for Solaris (5.8). Installers, please keep the list of tools up to date, by checking the guidelines. Thank you. This file will also be available from the web, so if you are checking to see whether a certain package is installed locally here, you can do a find in your browser window on this webpage.

If you're looking for other pages of this sort you might try the listing of related NLP/IR software sites. We are also considering making versions of these tools readily installable from a single CD, where licensing is not an issue. Please contact us if you are interested in the availability of this software.

This site and listing is supervised by Min-Yen Kan.

Table of contents:

Corpora

Proceedings

These proceedings are available to SoC staff and members through the URL: http://www.comp.nus.edu.sg/~rpnlpir/proceedings/. Also, check out the ACL Anthology, which has the collection of all ACL related publication.

Grammars

Lexicons

Tools

Libraries

Usage / Installation Guidelines

(Last updated : Thu Jan 9 10:10:02 GMT-8 2003

If you are planning to use tools here, and the tool doesn't work straight off, here are a couple things that you want to check before sending mail to the tool's administrator (listed in the square brackets in the tool description): Are you using the tool with the right arguments? The right operating system? Did you set the right environment variables? Check the tool's status in the description, is it listed as okay to use? Any usage notes given? If it is installed and permissions are not given for you to use it, check to see whether it is a private tool (links and entries to private tools are encouraged for networking purposes but that doesn't mean you can use it without permission). Finally, check to see whether there is a usage-soc.html file in the tool's local home that was provided by the installer or maintainer. If all these things don't help you solve you problem, THEN try contacting the administrator.

If you are installing a new tool, please observe these guidelines to ensure that the tool is properly used and can be easily found by others. There are a number of rules to follow to ensure that this shared directory does not get mangled and unmanageable.

  1. Install to the proper directory (*note about libraries, see step #2 below). At the root of this account, do a "find . -type d" to see the directory tree hierarchy. Figure out which single place best fits the description of the type of tool you'd like to install. If a proper generic directory doesn't exist, please create it following the naming style.
  2. (Libraries installation). To install shared C libraries, Java archives, or perl modules, etc., please install these to the "lib" subdirectory under the generic category. Place the library under a proper language header. For example, a part-of-speech library for C fictionally called "poslib" with version 1.2 would be installed under tools/taggers/lib/c/poslib-1.2. Use CPAN with perl 5.8.0 (see above) to install new perl modules. See the current list of modules installed for rpnlpir's copy of perl 5.8.2 for details.
  3. (documentation) After the software/corpora is installed, note any particularities in detail in a file called usage-soc.html in its home.
  4. (symlinking) If the package originally comes with a version number, include it in the actual directory. Symlink the directory to the package name without version number (e.g. poslib -> poslib-1.2) so that the most up-to-date package can be always accessed by the generic name, but such that older versions have a stable reference to access the version.
  5. (symlinking 2): If you think that this software or any generic directories created for it might also be well placed in other directories, create symlinks to the package or directory, such that users can find in by multiple pathways.
  6. (documentation 2): Edit this file (~/README.html) in the root directory of the account and create a new entry for the package you've installed. Include information on the status of the tool (if it works, or hasn't been tested) as well as pointers to its home page and a quick synopsis of quirks (if there are too many to list, you might want to point to the more comprehesive usage-soc.html you just wrote. The template for the package description is below:
  7. TOOL [Installed WHEN by UID under DIR (LANG) Maintained by UID ] DESC

  8. (running it): Finally, edit the ~/.profile file and include any aliases, environment variables and paths necessary to run the tool when a person is logged in as "rpnlpir". Create a short comment for it so people can copy the appropriate environment variable settings to have the tool run correctly.
  9. (documentation 3): if it is a restricted access resource, update the restrictedACL.txt file and change the appropriate permissions.
  10. (notify) send email to rpnlpir@comp saying what a fantastic job you've done!
  11. Whew! That's it. Thanks for installing the new package!
$Id: README.html,v 1.14 2005/06/30 02:11:40 rpnlpir Exp rpnlpir $