====== Creating Glossaries ======

When localizing computer software, especially into languages where IT terminology is not available, new IT terms need to be created. If a glossary of computer terms does not exist in that language, then one needs to be created as these terms stand as a 'backbone' in the translation work.

Glossaries can also be created for specific projects.  One clunky way of doing it, is described [[clunky glossary creation|here]].

===== Resources =====

  * [[http://www.it46.se/show_entry.php?id=128&lang=es|KiPot]]
  * [[http://www.christophermayo.com/articles/2004/makeglossary.html|How to make a simple glossary]]

===== How glossaries are created =====


The process of creating new terms involves joint work of
linguistic and computer science experts. The development of new IT terms
requires that the linguistic expert fully understands the meaning of the IT
term and its context before trying to match or extend the meaning of a word
in the destination language.

In the words of Alberto of the KiLinux team: "After completing our first 
Swahili IT Glossary of 700 terms we understood
very quickly two things: the first is that the glossary of 700 terms requires
to be extended constantly and second the extended glossary requires a constant
and fluent communication between our technical team and Swahili linguistic
experts. Oral communication was never an option since answers were required
promtly. Email was not suitalbe either, since we wanted all participants to
have the possibility to take part in the deveopment of all new tems, which
would lead to an enormous amount of mails sent back and fourth between the
projects participants. We needed an online communication tool that would
facilitate the communication, and therefore, [[http://www.it46.se/show_entry.php?id=128&lang=es|KiPot]] was developed."

===== Hacking a glossary from existing translation =====

One option to create your first cut glossary is to use existing translations and translatable applications.  

The first step is to create a list of potential glossary words.

<code bash>
$ pogrep --search=msgid -e '^\w+(\s+\w+){0,1}$' -i templates -o short-words
$ rename .pot .po `find short-words -name "*.pot"`
$ pocompendium glossary.pot -d short-words/
</code>

Optional: edit the POT file to remove words that shouldn't be in the glossary.  See [[toolkit:pogrep]] and
[[short_strings_first]] for more information on the regex used to extract the words.

Now the glossary POT file is eady for populating.  The next step is to take your existing translations and populate the POT file to create your
first cut language glossary.

<code bash>
$ pocompendium compedium-XX.pot -d XX/ # Create a compendium for the XX language
$ cp glossary.pot glossary.po
$ msgmerge --compendium=compendium-XX.po --update glossary-XX.po glossary.pot
$ po2csv glossary-XX.po glossary-XX.csv
</code>

Your glossary is now in CSV format in the file glossary-XX.csv