[Clipart] [Open Clip Art Library] Completely unusable site

[Erik Moeller]

On 9/24/07, Jon Phillips <jon at rejon.org> wrote:

> So, then considering picture #2, this seems like a great way to know the
> content will be live, there already is svg gallery, versioned metadata
> and versioned files. However, there are some key things that wikimedia
> commons doesn't have like structured system for rating, comments, tags,
> and several other cool features of a true content management system.

On the subject of comments, I am managing the development of
LiquidThreads, a fairly neat MediaWiki discussion system:
http://wikixp.org/lqt/index.php/Main_Page

It'll go into production on a couple of non-Wikimedia sites soon, and
I'll promote it inside Wikimedia as well once I'm comfortable with it.
It already has a few very neat features, and if I can continue to
secure funding for it, we'll pursue things like chat and e-mail
integration.

Categories to me are superior to tags in some ways, and inferior in
others. They are hierarchically organized, which makes them more
powerful, and they're standardized (one name per category). But since
there's no synonymy, that makes it harder to add the right ones.
There's a bunch of tools in Commons that are trying to help users with
that, though.

Both tags and categories suck in one important way: they're not
language-independent. I'd prefer to move towards "concept-tagging";
I'm working on a fairly exciting project called OmegaWiki.org, which
is a structured wiki-driven ontology, somewhat similar to Cyc if that
rings a bell. Concepts have an internal ID, and can be identified by
any associated synonym or translation. They also have a relational
network. So my dream is to replace tags and categories with OmegaWiki
concepts.

> There is a way to associate contributions with licensing, but there
> would be much work necessary to merge over this project into wikimedia
> commons. Erik, would there be assistance for this type of merge in a
> timely manner, how could resources be allocated?

Uploading the content shouldn't be very hard, perhaps 2-3 man days of
work + the time it takes to run the bot. (I speak from experience,
having uploaded some 20K media files to Commons via bots, some with
more complex metadata than OCAL.) So within any budget we allocate for
a migration it would be a fairly small component. We would want to do
existence-checks; MediaWiki doesn't do hash duplicate checking yet,
and it might be worth implementing that before running a mass upload.

It would probably be worth preserving the OCAL author URLs for
attribution purposes.

> Could all the developers who have accounts, get accounts on Wikimedia
> Commons to help develop?

It should be easy enough to get your devs committer access to our SVN,
which contains the codebase we work with. Getting shell access to the
Wikimedia clusters is a little harder ;-) but updates are taken live
fairly quickly through a process of continuous integration.

In general, I'd want to secure some funding for work on tech issues
_before_ seriously pursuing a content migration. But if you and the
rest of the OCAL folks are principally in favor, we could start making
a budget and then trying to identify one or multiple sources of
funding.

-- 
Toward Peace, Love & Progress:
Erik

DISCLAIMER: This message does not represent an official position of
the Wikimedia Foundation or its Board of Trustees.