[Mediawiki-l] WiktionaryZ/Wikidata updates; new tarball

[Erik Moeller]

The WiktionaryZ project has made rapid progress in the last few weeks,
thanks in no part to our growing team of developers, now including
myself, Peter-Jan Roes, Karsten Uil, and Rod Smith. There's a growing
community at http://www.wiktionaryz.org/ , and our intrepid editors
have already added more than 20,000 words to the original GEMET
database. While there are still many glaring deficiencies and lacking
key features, the WZ core model of terminology works.

Key functionality added to the code in the last few weeks includes:
- Basic versioning. All tables now have versioning information, based
on a transaction model where each record has an associated transaction
record. It is possible to get a simple aggregated history view on a
per-page level; more advanced history views are to come.
- Object identifiers. Key tables now have object IDs associated with
their records. These object IDs are themselves linked to a
database-independent universally unique identifier (UUID), which
allows us to expose data through APIs without using IDs that may not
be available in a target application. The object model also provides
us with a flexible method of grouping or annotating entitites.
- Sticky UI. Rod A. Smith has hacked a nice feature where your view of
the "tree hierarchy" in WiktionaryZ will remain persistent if cookies
are enabled. This allows editors to see only the information relevant
to them.
- New import scripts. We have some exciting new scientific data
sources that will be imported into WZ soon (target date is mid/end
October).
- Small things: hacks for search, adding languages, finding words that
need translation. Splitting exact meanings and imprecise ones.
Improved changelogging.

The versioning is arguably the biggest step here and puts us one step
closer to becoming a true wiki. Next in the roadmap: rollback and a
better class model for ontologies.

Now, let me be absolutely clear: WZ is still pre-alpha code, and it's
messy. We're still extending an old MediaWiki (not really a big deal
since there are relatively few points of attachment, but we'll still
need to fix it), and we have very little documentation. That said, if
you follow the instructions below, you should be able to get your own
edition up and running.

1) Download http://epov.org/temp/wz-sep14.tgz - it's a big file
because it contains the entire database _and_ uploads of
wiktionaryz.org.

2) Decompress the file to a directory in your Apache DocumentRoot.

3) Create and import the database from the .sql file in the dump (if
you don't know how to do this, you should probably stop here).

4) Correct LocalSettings.php $wgDBname to point to the correct
database and authentication, and fix the paths (/www/wz) to point to
the appropriate subdirectory of your DocumentRoot.

5) Run the wiki. Create a user through the "Create an account"
interface. Give the user the "wikidata" permission by inserting a row
into the user_groups table with the ug_user being the user ID of the
user you just created, and the ug_group being 'wikidata' (you should
probably give it 'sysop' and 'bureaucrat' rights as well).

You should now be able to edit the relational data in your WZ
installation, and can start playing with the code, particularly in the
extensions/Wikidata/WiktionaryZ subdirectory. You may want to get
updates to that code from the Wikimedia subversion repository.

I am currently unavailable but will be back on Sep 20 if you have
questions or problems. We are happy to accept new developers and will
prod Brion for you to give you Subversion access if necessary. ;-) We
also always have fun and interesting small and large projects to work
on.

-- 
Peace & Love,
Erik