Re: Is it possible to ... ?

An excellent question ... 

See here for the WLM database requirements 

https://commons.wikimedia.org/wiki/Commons:Monuments_database

It probably goes without saying that nothing will be in that exact format,
but some things would be more easily converted than others. So, there are a
few questions to ask about any resource:

* how many "monuments" does it have and what kind of coverage of Australia?
We need to get an *overall* broad coverage of Australia, rather than a few
intensive "hot spots" for a "national" competition. But the more resources
we have to work with, the more the workload increases as each one will
probably require individual processing arrangements.

* does it contain monuments that are off the main "tourist trail"? Anything
in a popular tourist area has probably already generated lots of Commons
photos. We'd like a competition to increase our coverage rather than just
attract more photos of the Sydney Opera House and the Sydney Harbour Bridge.

* does it contain records for things that are not monuments (even by the
most generous interpretation, e.g. motels, restaurants, tourist information
offices) and, if so, how easy is it to eliminate those records? 

* does it contain the range of information expected by the WLM database? If
not, how close does it get?

* is the format easily converted to the desired format? A technical
question, because if we can't automate it, it's a time-consuming manual
process.

I don't think we have to be "perfect" but we do need a decent coverage of
the monuments not already featured in Commons. I imagine we can get away
with not filling in all the fields of the database, although I gather
lat/long is important for showing "monuments near me" in the mobile app.

The first few database fields are straightforward: Australia, English, ID
(can be generated if underlying resource doesn't have one), adm0 =
Australia, adm1 = State/Territory (usually can be determined), adm2 = local
government (less likely to be known), adm3/4 not relevant to Australia, name
(should be known but may not be canonical, meaning we could end up with
duplicate entries coming from multiple sources but I don't think that is a
show-stopper). Lat/long isn't always available in older resources, although
newer resources increasingly have this data reflecting the now widespread
availability of GPS-capable devices/apps. It may be possible to geocode
using street address as an alternative (not a solution for "outback"
monuments). I think these are not the hard ones.

I note with a raised eyebrow that the database apparently requires you to
have an image of each monument, which I thought was the purpose of the
competition (to collect such images). Maybe a link to a copyright image on
the web is acceptable here? This is likely to be harder to obtain
automatically for many datasets.

The other hard things are:
* Commons category for the monument - unless we dump everything into one
category, or some simple breakup like one category per state
* the Wikipedia article or section for the monument -- now that's not easy
to automate (back to the canonical naming problem) and many monuments
probably don't yet get a mention, let alone an article.

And do we have the volunteers to do all this? Past conversations have not
yielded a lot of volunteers, and the people who do volunteer tend to be the
"usual suspects", so it really needs "not the usual suspects" to volunteer.

Kerry




_______________________________________________
Wikimediaau-l mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/wikimediaau-l

Reply via email to