The issue with Commons is actually not whether Wikipedia uses the picture
or not. The issue is the validity of description. If an image depicts A and
the description says it is B, then the data on Commons are obviously
invalid, and this would be the analog of false info at Wikidata sources to
unreliable sources or unsourced.This does not happen so often, despite the
fact that many Commons images have dubious provenance, and is currently
much more common on Wikidata. This is why the English Wikipedia community
opposes to the usage of the BLP Wikidata data, but does not oppose to the
usage of Commons images.
Cheers
Yaroslav
On Wed, Sep 27, 2017 at 7:49 AM, Gerard Meijssen <gerard.meijssen(a)gmail.com>
wrote:
Hoi,
When a database is linked to, there are many reasons for linking. One is it
is "authoritative" so the data is of a high quality or it is the standard
bearer in a particular field. Another reason is because there is a clear
operational purpose. Linking to the Open Library for instance has such a
purpose; it allows us to link to free content; it provides the basics for a
mechanism so that Wikipedia readers can read books of an author or read a
particular book.
One reason often neglected is that the other database is actively
maintained and its maintainers collaborate with people at Wikidata to
mutual advantage. This is the case with the people at Open Library, with
the people at OCLC. It is most powerful because past activities have had a
measurable effect in their projects and in Wikidata. From my personal
perspective active collaboration is to be preferred over the authority of
another source.
The reason why both red, blue and black links ought to be linked to
Wikidata is because it enables comparison and evaluation. When red links
are linked to a Wikidata item they will not turn blue whan an autonym is
created. Blue links have an implicit link to a Wikidata item. It happens
all too often, particularly in lists, that a blue link is associated with
the wrong article. It is reasonable to expect that multiple instances of
the same list contains links to the same items. With an explicit link it
becomes easy for bots to compare lists in the different Wikipedias and find
these differences. It is also possible to compare with Wikidata but that is
of a secondary relevance..
With red links and blue links linked to Wikidata, the similarity of the
data on an item with the data in an article indicates a probability that
the quality in Wikidata is high. Given the huge number of statements on
items that have no reference it is the best that can be done given the
enormous amount of data in Wikidata.
Given the policies of Wikidata, there will be references to living people
that only exist to complete a list. I am adding many Dutch authors at this
time to complete the award winners of Dutch literature awards. They consist
of a label in Dutch, the fact of their humanity often a gender indication
and the fact that they won the award. This pattern is true for many awards
and, it is an accepted consequence of the Wikidata notability policy. These
are in effect red links in a Wikipedia.
Thanks,
GerardM
On 27 September 2017 at 05:08, Alessandro Marchetti <alexmar983(a)yahoo.it>
wrote:
Personally, I think that if person has an ID on
some databases, than it
can stay on wikidata. Once in a while some database can be removed if
issues are pointed out about their accuracy, but if a database is sound
and
professional, we should use it to fix an item. it
could be the same for a
databases of sites, buildings or museum items too. Creating a
wikipedia-style averaged policy on the issue is much more vague.
Especially
when local pages do not exist, the IDs is the key
parameter to start,
IMHO.
It is ok if a wikipedia has only a fraction of relevant "photographers"
or
"painters" or "athletes"...
but a database should be complete and
objective, it cannot rely on the funnel of what some wikipedia accepts
and
other don't, it would make it more biased and
unbalanced importing a
local
bias. What's the point for example if I find
an archive of Dutch
photographers with IDs to import only those that have a page on nlwiki
(or
maybe enwiki, dewiki, frwiki)? You import all the
codes, some items will
have wikipedia pages, some will not, what's the real issue on this
aspect?
Being standardized and coherent is more important
for an archive.
About the quality of the items, this comes as a second step. Some of them
will always be less cured, we can say that for a BLP a minimum
requirement
of properties is necessary for example. I can
accept that an item with
just
one ID is removed if no additional information
can be found. That is, a
BLP
item with a limited number of properties and no
platform and just one ID
can be proposed for deletion, although this should not be an automatism.
But if you care about an item, you can improve it if it risks to be
deleted. This is a functional issue, if an item does not tell me if
you're
a man or a woman, your age, your profession...
well it is basically few
things more than a ugly duplicate of a string in the url of the original
database, so what's its utility? Some more complete output in some basic
query here and there, maybe, but it should be possible to ask more. The
point is that this should be considered in the framework of a database
and
its use, a more "functional" than
"philosophical" perspective.
P.S. Not sure I have understood the blue and red link request, in some
minor wiki red link can be linked to wikidata, but why the blue one?
Il Martedì 26 Settembre 2017 19:07, Gerard Meijssen <
gerard.meijssen(a)gmail.com> ha scritto:
Hoi,
There is a lot to do about the current absence of a BLP policy at
Wikidata.
Many people, particularly those involved in
Wikipedia, insist on one and
a
policy that is a mirror image of their policy.
I am opposed to such an approach because it will be detrimental to the
best
practices in Wikidata and it will stifle the
inclusion of data.
Nevertheless there is a need for better quality particularly where it
concerns BLP.
Only being against is a bad position so I have laid out the arguments
for a
more inclusive BLP and quality approach [1]. It
does bring many of the
relevant questions together.
What this approach accomplishes is:
* better quality in both Wikipedia and Wikidata
* an opt in change in the Wikipedia environment that links blue and red
links to Wikidata items
* it allows for the Wikidata best practices
* it invites any Wikimedia collaborator to make a positive difference for
our overall BLP.
What it does not provide is an instant BLP solution for Wikidata, this is
not realistic given the huge number of items involved, people often
specific to one or no Wikipedia. It will not convince everyone and that
too
is to be expected. After all the proof of the
pudding is in the eating
and
not so much in the endless bickering.
Thanks,
GerardM
[1]
https://ultimategerardm.blogspot.nl/2017/09/wikimedia-
and-its-blp-approach.html
_______________________________________________
Wikimedia-l mailing list, guidelines at:
https://meta.wikimedia.org/
wiki/Mailing_lists/Guidelines
and
https://meta.wikimedia.org/wiki/Wikimedia-l
New messages to: Wikimedia-l(a)lists.wikimedia.org
Unsubscribe:
https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
<mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe>
_______________________________________________
Wikimedia-l mailing list, guidelines at:
https://meta.wikimedia.org/
wiki/Mailing_lists/Guidelines and
https://meta.wikimedia.org/
wiki/Wikimedia-l
New messages to: Wikimedia-l(a)lists.wikimedia.org
Unsubscribe:
https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
<mailto:wikimedia-l-request@lists.wikimedia.org?subject=unsubscribe>