The university of Tromsø > Giellatekno
 

Anonymous CVS

Note
The source code and the documentation is distributed under the GPL-license. You will find the license text in xtdoc/LISENS.txt and gt/LISENS.txt
Note
Preparation and Check-out you will only have to do once

Preparation

Anyone can download the source code via anonymous CVS, but will not be able to contribute with modifications.

Open a terminal and use the following command to login to the cvs server:

cvs -d :pserver:anonymous@victorio.uit.no:/usr/local/cvs/repository login

You will be greeted with this message:

Logging in to :pserver:anonymous@victorio.uit.no:2401/usr/local/cvs/repository
CVS password:

Just press enter and continue. You will get a harmless error message similar to the following:

cvs login: warning: failed to open /Users/boerre/.cvspass for reading:
No such file or directory

You can ignore it.

Checkout

You are now ready to copy the files of our project to your machine, or check out the files. Issue the following command to get a copy of our source and documentation:

cvs -d:pserver:anonymous@victorio.uit.no:/usr/local/cvs/repository co xtdoc gt words

If all goes well you will be greeted with lots of lines of this sort:

U gt/smj/src/noun-smj-lex.txt
U gt/smj/src/pp-smj-lex.txt
U gt/smj/src/propernoun-smj-lex.txt
U gt/smj/src/punct-smj-lex.txt
U gt/smj/src/smj-dis.rle
U gt/smj/src/smj-lex.txt
U gt/smj/src/smj-num.txt
U gt/smj/src/spellrelax.regex
U gt/smj/src/tok.txt
U gt/smj/src/twol-smj.txt

You will also have two new directories, xtdoc and gt inside your working directory

Updating your local copy

If you want to update the documents (get new versions) at a later time, issue the following command:

cvs -d:pserver:anonymous@victorio.uit.no:/usr/local/cvs/repository up -dP xtdoc gt

in the same directory that you did the checkout, or just

cvs up -dP

within one of the subdirectories.

Building, compiling and running

When everything is checked out and residing on your computer, you probably want to compile and run our linguistic analysers. For this you need to install the required tools. They are:

  • Xerox morphological transducer tools, for morphological analysis. This is proprietary software, and not available for download, but it is included with the book Finite State Morphology by Kenneth R. Beesley and Lauri Karttunen. The book has its own home page. Unfortunately we do not have an open source alternative for compiling our morphological analysers, basically because there are none compatible with our source, and almost none able to handle Sámi in a decent way. One possible open-source alternative would be SFST, but it would require major reworking of our lexicons, and a lot of other work as well.
  • vislcg, for disambiguation. This is open source, but has no documentation accompanying the software (there are references to external/printed documentation). You need to download the source code and compile it yourself, using a GCC 2.9x compiler (or update the source code to compile with newer versions of GCC:-) ). Basic user documentation can be found on our pages.

Other tools (perl scripts, shell scripts, etc.) should have been included in the CVS check-out, but might require some environment setups (PATHs etc). See our technical documentation and the subsection in the corresponding menu.

Reading the documentation off-line

The full documentation is available online on divvun.no, but if you want to read the documentation off-line, you are best off installing and running Forrest. You will find it at our server.

We have made a document on how to install and use forrest, follow the instructions there on how to use forrest on your private machine

Sending feedback and patches

If you have improvements to our code, be it linguistic corrections, C-code bug fixes, documentation enhancements or whatever, we would welcome your contribution. Please make a cvs diff, gzip it, and send it to us.

Warning
This e-mail address may change in the future!

Interactive feedback and discussions

The whole team is available on AIM, and we frequently keep discussions through our favourite AIM client(s). Our AIM addresses can be found on our contact page.

We also use SubEthaEdit (SEE) a lot to do collaborative development, but SEE is Mac only, and commercial (version 2.2 is available for free for non-commercial purposes). A rather new cross-platform alternative is Gobby (please see the installation notes, especially Mac users - the Mac version only installs through DarwinPorts). Try it, or SEE - realtime collaborative development is both very fun and productive, and creates a whole new feeling of working together, even when sitting in different countries! Much of what we have done would have taken 10 x the time, or not been possible at all without SEE.

If you have questions or problems, either e-mail us, or start a chat. Further investigation can be done collaboratively and interactively using SEE or Gobby. If you are using a Mac with a recent OS version (10.4.x, aka Tiger), you can even invite us to (group) voice or video chats using iChat. Non-group voice and video chats should also work with earlier versions of MacOS, as well as with a recent version of the AIM client for Windows (we haven't tried voice or video chats with Windows users).

Last modified: $Date: 2007/03/15 14:02:15 $, by $Author: boerre $

by Børre Gaup, Sjur N. Moshagen