Getting Started

This is a list of programs you will need to use this infrastructure for language analysis and development.

Hardware and system requirements

  • Machine requirements:
    • 8 Gb RAM is needed for Hfst transducer compilation, the more the better (but 4 Gb is fine if you only use the Xerox tools)
    • fast hard-disk preferred
  • OS/System - you will need unix:
    • Mac (used by most of the Divvun and GT teams)
    • Any recent Linux distribution is fine (also used by the Divvun & GT teams)
    • Windows doesn't have a unix shell, but there are two alternatives:
      • installing Cygwin. Warning! This is very briefly tested, and you will be on your own. See more info further down
      • Alternatively, you might log on to a remote unix machine, see our Windows help page
  • Filesystem: Unicode-compatible filename encoding

System preparations - software to install

Getting the Divvun/Giellatekno infrastructure & source code for your language

There are three alternatives:

1. Everything

Check out  our svn repository, then run $GTHOME/gt/script/gtsetup.sh - this gives you everything but is quite big, more than 3.5Gb in download size, and twice as much on your hard-disk.

Not recommended when you are on slow lines or if you are low on hard-disk space!

2. Only the core and single languages (on your own computer)

Follow these instructions (under the Only the GT core and the wanted language(s)heading) to only check out the required parts for working with a single or a few languages.

3. Only the core and single languages (on a multiuser server)

To use this configuration, follow these instructions:

Start the development!

See this page for details on how to use our infrastructure to develop your morphologies, lexicons and more to create tools for yourself and your language community.

A final note

This list is written for people not working at Divvun or Giellatekno. You may also look at the longer list of what we install for our new workers and the checklist for what new users should know, and see if some of that is interesting. But this document will give you what you need to get started.