!!!Risten2 & Termwiki

Topics:
* import
* export
* next steps

!!!import

Concepts are imported, but without subcategories. The next step is to import the expressions - all expressions in all language files.

Remember to exclude (both for the expressions and the concepts) the Z category.

When is the import done? By Friday.

When done, add a news field on the front page highlighting the new data.

Change default language to SME, other languages as translations. The translations should be administered by Giellagáldu, to make sure all languages are kept up to date.

__TODO:__
* finish import of expressions
* add subcategories

!!!export

presently there are three code bases (at least):
* java class by Antti (producing xml directly)
** Antti is on a two-week vacation, no java before he is back
* scala program by Tomi (using the media wiki php export script as the first step)
* Børre's import script if "reversed"
* an export php script producing a csv with the concept pages, made by Antti
** leaves out categories and their ID's - not usable

Tomi's code:

* termcenter almost ready - printed with a few systematic errors
* language files not yet printed

Decision: we continue with Tomi's code.

The different categories are handled by the PHP script (exporting one category at a time).

Questions:
* how can we automatise the PHP export (now it requires a manual form + download)

__TODO:__
* fix remaining termcenter issues
* use the numeric namespace id as the term collection id, + "termwiki":
** {{termwiki-1234}}
* add support for multilingual definitions etc
* add support for related concepts
** replace the string-based concept ID with the numerical concept ID used in the
   termcenter file: {{Boazodoallu:guolmmámiessi}} => {{194}}
*** use the numeric id as an xpointer ref to the related concept
*** in eXist, when displaying related concepts, look up that concept and expand
    XIncludes, returning a list of language-tagged expressions - then one can
    display the expression for that concept that is closest to the language
    preferences of the user
* print language files
** combine senses for each language entry: if entry A is used for concepts 1 and
   2, language entry A has two senses, one pointing to concept 1, and another
   pointing to concept 2.
* automatise downloading
* all converted files automatically checked in in svn
* create metadata files (stored in svn) to hold metadata needed for Risten2 but
  not found in the Termwiki (for each category)

!!!Next steps

* import smaller collections from SMA and other languages
* merge Mekanikk-termer with the existing term collections
* delete (or move to an attick branch) the old terminology collections in our
  svn, and replace them with the term collections exported from the termwiki
* import other term collections, such as the legal terms (provided that the
  group behind the terms is willing to use the termwiki as their term editor)