[Wikimedia-l] Quality issues

classic Classic list List threaded Threaded
166 messages Options
1234 ... 9
Reply | Threaded
Open this post in threaded view
|

[Wikimedia-l] Quality issues

Gerard Meijssen-3
Hoi,
At Wikidata we often find issues with data imported from a Wikipedia. Lists
have been produced with these issues on the Wikipedia involved and arguably
they do present issues with the quality of Wikipedia or Wikidata for that
matter. So far hardly anything resulted from such outreach.

When Wikipedia is a black box, not communicating about with the outside
world, at some stage the situation becomes toxic. At this moment there are
already those at Wikidata that argue not to bother about Wikipedia quality
because in their view, Wikipedians do not care about its own quality.

Arguably known issues with quality are the easiest to solve.

There are many ways to approach this subject. It is indeed a quality issue
both for Wikidata and Wikipedia. It can be seen as a research issue; how to
deal with quality and how do such mechanisms function if at all.

I blogged about it..
Thanks,
     GerardM

http://ultimategerardm.blogspot.nl/2015/11/what-kind-of-box-is-wikipedia.html
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] Quality issues

Jane Darnell
Gerard,
I think this was always the case. Most Wikidatans are as at home on
Wikipedia as they are on Commons. The issue you describe also happened to
Commons - both communities feel the other is less focussed on quality. Many
Commonists spend hours on high quality images and these are rarely picked
up by Wikipedia unless a Commonist notices and does so in their own
language. There is no requirement for Wikipedians to get to know any other
project and this is normal wiki behavior. We don't want anyone to feel
pressured to do anything they feel uncomfortable doing. It's already
difficult to get Wikipedians to do small tasks like add catagories to their
articles. The list of things necessary to create an acceptable article on
Wikipedia just seems to get longer and longer, while the associated work
for illustrations of that article or for data of that article is not even
mentioned in current AfC policies on Wikipedia. I have thought about this,
but I still think we need to break down the list of things necessary to
make new short articles on Wikipedia, not extend the list. So in summary, I
think that what you describe is normal predictable behavior for a
"Wikipedia support" project such as Commons and Wikidata. This will change
as more and more external users find out that Commons and Wikidata are
valuable resources in and of themselves. This is already the case for many
GLAMs which have found collaborations with Commons to be valuable
experiences. I have high hopes this will become the case for Wikidata as
well.
Jane

On Fri, Nov 20, 2015 at 8:18 AM, Gerard Meijssen <[hidden email]>
wrote:

> Hoi,
> At Wikidata we often find issues with data imported from a Wikipedia. Lists
> have been produced with these issues on the Wikipedia involved and arguably
> they do present issues with the quality of Wikipedia or Wikidata for that
> matter. So far hardly anything resulted from such outreach.
>
> When Wikipedia is a black box, not communicating about with the outside
> world, at some stage the situation becomes toxic. At this moment there are
> already those at Wikidata that argue not to bother about Wikipedia quality
> because in their view, Wikipedians do not care about its own quality.
>
> Arguably known issues with quality are the easiest to solve.
>
> There are many ways to approach this subject. It is indeed a quality issue
> both for Wikidata and Wikipedia. It can be seen as a research issue; how to
> deal with quality and how do such mechanisms function if at all.
>
> I blogged about it..
> Thanks,
>      GerardM
>
>
> http://ultimategerardm.blogspot.nl/2015/11/what-kind-of-box-is-wikipedia.html
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] Quality issues

Peter Southwood
In reply to this post by Gerard Meijssen-3
Gerard,
Who were you expecting would respond from the Wikipedias?
Cheers,
Peter

-----Original Message-----
From: Wikimedia-l [mailto:[hidden email]] On Behalf Of Gerard Meijssen
Sent: Friday, 20 November 2015 9:18 AM
To: Wikimedia Mailing List; Research into Wikimedia content and communities; WikiData-l
Subject: [Wikimedia-l] Quality issues

Hoi,
At Wikidata we often find issues with data imported from a Wikipedia. Lists have been produced with these issues on the Wikipedia involved and arguably they do present issues with the quality of Wikipedia or Wikidata for that matter. So far hardly anything resulted from such outreach.

When Wikipedia is a black box, not communicating about with the outside world, at some stage the situation becomes toxic. At this moment there are already those at Wikidata that argue not to bother about Wikipedia quality because in their view, Wikipedians do not care about its own quality.

Arguably known issues with quality are the easiest to solve.

There are many ways to approach this subject. It is indeed a quality issue both for Wikidata and Wikipedia. It can be seen as a research issue; how to deal with quality and how do such mechanisms function if at all.

I blogged about it..
Thanks,
     GerardM

http://ultimategerardm.blogspot.nl/2015/11/what-kind-of-box-is-wikipedia.html
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>

-----
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2016.0.7227 / Virus Database: 4460/11032 - Release Date: 11/20/15


_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] [Wiki-research-l] Quality issues

WereSpielChequers-2
In reply to this post by Gerard Meijssen-3
My experience is that pretty much all Wikimedians care about quality,
though some have different, even diametrically opposed views as to what
quality means and which things are cosmetic or crucial.

My experience of the sadly dormant death anomaly project
<https://meta.wikimedia.org/wiki/Death_anomalies_table> was that people
react positively to being told "here is a list of anomalies on your
language wikipedia" especially if those anomalies are relatively serious.
My experience of edits on many different languages is that wikipedians
appreciate someone who improves articles, even if you don't speak their
language. Dismissing any of our thousand wikis as a "black box" is I think
less helpful.

One of the great opportunities of Wikidata is to do the sort of data driven
anomaly finding that we pioneered with the death anomalies report. But we
always need to remember that there are cultural difference between wikis,
and not just in such things as the age at which we assume people are dead.
Diplomacy is a useful skill in cross wiki work.

~~~~

On 20 November 2015 at 07:18, Gerard Meijssen <[hidden email]>
wrote:

> Hoi,
> At Wikidata we often find issues with data imported from a Wikipedia.
> Lists have been produced with these issues on the Wikipedia involved and
> arguably they do present issues with the quality of Wikipedia or Wikidata
> for that matter. So far hardly anything resulted from such outreach.
>
> When Wikipedia is a black box, not communicating about with the outside
> world, at some stage the situation becomes toxic. At this moment there are
> already those at Wikidata that argue not to bother about Wikipedia quality
> because in their view, Wikipedians do not care about its own quality.
>
> Arguably known issues with quality are the easiest to solve.
>
> There are many ways to approach this subject. It is indeed a quality issue
> both for Wikidata and Wikipedia. It can be seen as a research issue; how to
> deal with quality and how do such mechanisms function if at all.
>
> I blogged about it..
> Thanks,
>      GerardM
>
>
> http://ultimategerardm.blogspot.nl/2015/11/what-kind-of-box-is-wikipedia.html
>
> _______________________________________________
> Wiki-research-l mailing list
> [hidden email]
> https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
>
>
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] Quality issues

Mormegil
In reply to this post by Gerard Meijssen-3
On Fri, Nov 20, 2015 at 8:18 AM, Gerard Meijssen <[hidden email]>
wrote:

> When Wikipedia is a black box, not communicating about with the outside
> world, at some stage the situation becomes toxic. At this moment there are
> already those at Wikidata that argue not to bother about Wikipedia quality
> because in their view, Wikipedians do not care about its own quality.
>

Right. When some users blindly dump random data to Wikidata, not
communicating about with the outside world, at some stage the situation
becomes toxic. At this moment there are already those at Wikipedia that
argue not to bother about Wikidata quality because in their view,
Wikidatans do not care about its own quality.

For instance, take a look at
https://www.wikidata.org/wiki/User_talk:GerardM
https://www.wikidata.org/wiki/User_talk:GerardM/Archive_1

Erm
-- [[cs:User:Mormegil | Petr Kadlec]]
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] Quality issues

Richard Symonds-3
Folks, regardless of which views we hold, we're all on the same side - can
we try and be a little less acerbic please - it is Friday after all!

Richard Symonds
Wikimedia UK
0207 065 0992

Wikimedia UK is a Company Limited by Guarantee registered in England and
Wales, Registered No. 6741827. Registered Charity No.1144513. Registered
Office 4th Floor, Development House, 56-64 Leonard Street, London EC2A 4LT.
United Kingdom. Wikimedia UK is the UK chapter of a global Wikimedia
movement. The Wikimedia projects are run by the Wikimedia Foundation (who
operate Wikipedia, amongst other projects).

*Wikimedia UK is an independent non-profit charity with no legal control
over Wikipedia nor responsibility for its contents.*

On 20 November 2015 at 13:50, Petr Kadlec <[hidden email]> wrote:

> On Fri, Nov 20, 2015 at 8:18 AM, Gerard Meijssen <
> [hidden email]>
> wrote:
>
> > When Wikipedia is a black box, not communicating about with the outside
> > world, at some stage the situation becomes toxic. At this moment there
> are
> > already those at Wikidata that argue not to bother about Wikipedia
> quality
> > because in their view, Wikipedians do not care about its own quality.
> >
>
> Right. When some users blindly dump random data to Wikidata, not
> communicating about with the outside world, at some stage the situation
> becomes toxic. At this moment there are already those at Wikipedia that
> argue not to bother about Wikidata quality because in their view,
> Wikidatans do not care about its own quality.
>
> For instance, take a look at
> https://www.wikidata.org/wiki/User_talk:GerardM
> https://www.wikidata.org/wiki/User_talk:GerardM/Archive_1
>
> Erm
> -- [[cs:User:Mormegil | Petr Kadlec]]
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] Quality issues

Gerard Meijssen-3
In reply to this post by Jane Darnell
Hoi,
The difference between the use of quality images from Commons and
establishing what is correct is quite distinct. With Commons it is an
esthetic difference, with these lists it is about the credibility of the
data involved.
Thanks,
    GerardM

On 20 November 2015 at 09:53, Jane Darnell <[hidden email]> wrote:

> Gerard,
> I think this was always the case. Most Wikidatans are as at home on
> Wikipedia as they are on Commons. The issue you describe also happened to
> Commons - both communities feel the other is less focussed on quality. Many
> Commonists spend hours on high quality images and these are rarely picked
> up by Wikipedia unless a Commonist notices and does so in their own
> language. There is no requirement for Wikipedians to get to know any other
> project and this is normal wiki behavior. We don't want anyone to feel
> pressured to do anything they feel uncomfortable doing. It's already
> difficult to get Wikipedians to do small tasks like add catagories to their
> articles. The list of things necessary to create an acceptable article on
> Wikipedia just seems to get longer and longer, while the associated work
> for illustrations of that article or for data of that article is not even
> mentioned in current AfC policies on Wikipedia. I have thought about this,
> but I still think we need to break down the list of things necessary to
> make new short articles on Wikipedia, not extend the list. So in summary, I
> think that what you describe is normal predictable behavior for a
> "Wikipedia support" project such as Commons and Wikidata. This will change
> as more and more external users find out that Commons and Wikidata are
> valuable resources in and of themselves. This is already the case for many
> GLAMs which have found collaborations with Commons to be valuable
> experiences. I have high hopes this will become the case for Wikidata as
> well.
> Jane
>
> On Fri, Nov 20, 2015 at 8:18 AM, Gerard Meijssen <
> [hidden email]>
> wrote:
>
> > Hoi,
> > At Wikidata we often find issues with data imported from a Wikipedia.
> Lists
> > have been produced with these issues on the Wikipedia involved and
> arguably
> > they do present issues with the quality of Wikipedia or Wikidata for that
> > matter. So far hardly anything resulted from such outreach.
> >
> > When Wikipedia is a black box, not communicating about with the outside
> > world, at some stage the situation becomes toxic. At this moment there
> are
> > already those at Wikidata that argue not to bother about Wikipedia
> quality
> > because in their view, Wikipedians do not care about its own quality.
> >
> > Arguably known issues with quality are the easiest to solve.
> >
> > There are many ways to approach this subject. It is indeed a quality
> issue
> > both for Wikidata and Wikipedia. It can be seen as a research issue; how
> to
> > deal with quality and how do such mechanisms function if at all.
> >
> > I blogged about it..
> > Thanks,
> >      GerardM
> >
> >
> >
> http://ultimategerardm.blogspot.nl/2015/11/what-kind-of-box-is-wikipedia.html
> > _______________________________________________
> > Wikimedia-l mailing list, guidelines at:
> > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > [hidden email]
> > Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > <mailto:[hidden email]?subject=unsubscribe>
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] Quality issues

Gerard Meijssen-3
In reply to this post by Peter Southwood
Hoi,
So far such lists have been produced for bigger Wikipedias but essentially
it is potentially an issue for any and all Wikis that have data that may
exist on Wikidata or linked through Wikidata on external sources.
Thanks,
      GerardM

On 20 November 2015 at 12:33, Peter Southwood <[hidden email]>
wrote:

> Gerard,
> Who were you expecting would respond from the Wikipedias?
> Cheers,
> Peter
>
> -----Original Message-----
> From: Wikimedia-l [mailto:[hidden email]] On
> Behalf Of Gerard Meijssen
> Sent: Friday, 20 November 2015 9:18 AM
> To: Wikimedia Mailing List; Research into Wikimedia content and
> communities; WikiData-l
> Subject: [Wikimedia-l] Quality issues
>
> Hoi,
> At Wikidata we often find issues with data imported from a Wikipedia.
> Lists have been produced with these issues on the Wikipedia involved and
> arguably they do present issues with the quality of Wikipedia or Wikidata
> for that matter. So far hardly anything resulted from such outreach.
>
> When Wikipedia is a black box, not communicating about with the outside
> world, at some stage the situation becomes toxic. At this moment there are
> already those at Wikidata that argue not to bother about Wikipedia quality
> because in their view, Wikipedians do not care about its own quality.
>
> Arguably known issues with quality are the easiest to solve.
>
> There are many ways to approach this subject. It is indeed a quality issue
> both for Wikidata and Wikipedia. It can be seen as a research issue; how to
> deal with quality and how do such mechanisms function if at all.
>
> I blogged about it..
> Thanks,
>      GerardM
>
>
> http://ultimategerardm.blogspot.nl/2015/11/what-kind-of-box-is-wikipedia.html
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>
> -----
> No virus found in this message.
> Checked by AVG - www.avg.com
> Version: 2016.0.7227 / Virus Database: 4460/11032 - Release Date: 11/20/15
>
>
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] [Wiki-research-l] Quality issues

Gerard Meijssen-3
In reply to this post by WereSpielChequers-2
Hoi,
I have been working on Wikidata for almost two years on recent deaths. It
is one easy and obvious thing to signal recent deaths to all the WIkipedias
that have articles.It is quite similar to what you describe. It is dead
easy to produce such lists, not only for recent deaths but also for deaths
that differ from one source to the next.
Thanks,
      GerardM

On 20 November 2015 at 14:30, WereSpielChequers <[hidden email]
> wrote:

> My experience is that pretty much all Wikimedians care about quality,
> though some have different, even diametrically opposed views as to what
> quality means and which things are cosmetic or crucial.
>
> My experience of the sadly dormant death anomaly project
> <https://meta.wikimedia.org/wiki/Death_anomalies_table> was that people
> react positively to being told "here is a list of anomalies on your
> language wikipedia" especially if those anomalies are relatively serious.
> My experience of edits on many different languages is that wikipedians
> appreciate someone who improves articles, even if you don't speak their
> language. Dismissing any of our thousand wikis as a "black box" is I think
> less helpful.
>
> One of the great opportunities of Wikidata is to do the sort of data driven
> anomaly finding that we pioneered with the death anomalies report. But we
> always need to remember that there are cultural difference between wikis,
> and not just in such things as the age at which we assume people are dead.
> Diplomacy is a useful skill in cross wiki work.
>
> ~~~~
>
> On 20 November 2015 at 07:18, Gerard Meijssen <[hidden email]>
> wrote:
>
> > Hoi,
> > At Wikidata we often find issues with data imported from a Wikipedia.
> > Lists have been produced with these issues on the Wikipedia involved and
> > arguably they do present issues with the quality of Wikipedia or Wikidata
> > for that matter. So far hardly anything resulted from such outreach.
> >
> > When Wikipedia is a black box, not communicating about with the outside
> > world, at some stage the situation becomes toxic. At this moment there
> are
> > already those at Wikidata that argue not to bother about Wikipedia
> quality
> > because in their view, Wikipedians do not care about its own quality.
> >
> > Arguably known issues with quality are the easiest to solve.
> >
> > There are many ways to approach this subject. It is indeed a quality
> issue
> > both for Wikidata and Wikipedia. It can be seen as a research issue; how
> to
> > deal with quality and how do such mechanisms function if at all.
> >
> > I blogged about it..
> > Thanks,
> >      GerardM
> >
> >
> >
> http://ultimategerardm.blogspot.nl/2015/11/what-kind-of-box-is-wikipedia.html
> >
> > _______________________________________________
> > Wiki-research-l mailing list
> > [hidden email]
> > https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
> >
> >
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] Quality issues

Gerard Meijssen-3
In reply to this post by Mormegil
Hoi,
<grin> quality is different things </grin> I do care about quality but I do
not necessarily agree with you how to best achieve it. Arguably bots are
better and getting data into Wikidata than people. This means that the
error rate of bots is typically better than what people do. It is all in
the percentages.

I have always said that the best way to improve quality is by comparing
sources. When Wikidata has no data, it is arguably better to import data
from any source. When the quality is 90% correct, there is already 100%
more data. When 100% is compared with another source and 85% is the same,
you only have to check 15% and decide what is right. When you compare with
two distinct sources, the percentage that differs changes again.. :) In
this way it makes sense to check errors

It does not help when you state that either party has people that care or
do not care about quality. By providing a high likelihood that something is
problematic, you will learn who actually makes a difference. It however
started with having data to compare in the first place
Thanks,
      GerardM

On 20 November 2015 at 14:50, Petr Kadlec <[hidden email]> wrote:

> On Fri, Nov 20, 2015 at 8:18 AM, Gerard Meijssen <
> [hidden email]>
> wrote:
>
> > When Wikipedia is a black box, not communicating about with the outside
> > world, at some stage the situation becomes toxic. At this moment there
> are
> > already those at Wikidata that argue not to bother about Wikipedia
> quality
> > because in their view, Wikipedians do not care about its own quality.
> >
>
> Right. When some users blindly dump random data to Wikidata, not
> communicating about with the outside world, at some stage the situation
> becomes toxic. At this moment there are already those at Wikipedia that
> argue not to bother about Wikidata quality because in their view,
> Wikidatans do not care about its own quality.
>
> For instance, take a look at
> https://www.wikidata.org/wiki/User_talk:GerardM
> https://www.wikidata.org/wiki/User_talk:GerardM/Archive_1
>
> Erm
> -- [[cs:User:Mormegil | Petr Kadlec]]
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] Quality issues

Milos Rancic-2
Offtopic: Gerard, during the last half an hour or so, I am just
getting emails from you inside of this thread (including wiki-research
list). I thought my phone has a bug. It's useful to write a larger
email with addressing all the issues. Besides other things, with this
frequency, you'll spend your monthly email quota for this list the day
after tomorrow.

On Fri, Nov 20, 2015 at 11:35 PM, Gerard Meijssen
<[hidden email]> wrote:

> Hoi,
> <grin> quality is different things </grin> I do care about quality but I do
> not necessarily agree with you how to best achieve it. Arguably bots are
> better and getting data into Wikidata than people. This means that the
> error rate of bots is typically better than what people do. It is all in
> the percentages.
>
> I have always said that the best way to improve quality is by comparing
> sources. When Wikidata has no data, it is arguably better to import data
> from any source. When the quality is 90% correct, there is already 100%
> more data. When 100% is compared with another source and 85% is the same,
> you only have to check 15% and decide what is right. When you compare with
> two distinct sources, the percentage that differs changes again.. :) In
> this way it makes sense to check errors
>
> It does not help when you state that either party has people that care or
> do not care about quality. By providing a high likelihood that something is
> problematic, you will learn who actually makes a difference. It however
> started with having data to compare in the first place
> Thanks,
>       GerardM
>
> On 20 November 2015 at 14:50, Petr Kadlec <[hidden email]> wrote:
>
>> On Fri, Nov 20, 2015 at 8:18 AM, Gerard Meijssen <
>> [hidden email]>
>> wrote:
>>
>> > When Wikipedia is a black box, not communicating about with the outside
>> > world, at some stage the situation becomes toxic. At this moment there
>> are
>> > already those at Wikidata that argue not to bother about Wikipedia
>> quality
>> > because in their view, Wikipedians do not care about its own quality.
>> >
>>
>> Right. When some users blindly dump random data to Wikidata, not
>> communicating about with the outside world, at some stage the situation
>> becomes toxic. At this moment there are already those at Wikipedia that
>> argue not to bother about Wikidata quality because in their view,
>> Wikidatans do not care about its own quality.
>>
>> For instance, take a look at
>> https://www.wikidata.org/wiki/User_talk:GerardM
>> https://www.wikidata.org/wiki/User_talk:GerardM/Archive_1
>>
>> Erm
>> -- [[cs:User:Mormegil | Petr Kadlec]]
>> _______________________________________________
>> Wikimedia-l mailing list, guidelines at:
>> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
>> [hidden email]
>> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
>> <mailto:[hidden email]?subject=unsubscribe>
>>
> _______________________________________________
> Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>

_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] Quality issues

Gnangarra
In reply to this post by Gerard Meijssen-3
>
> ...
> *When 100% is compared with another source and 85% is the same,**you only
> have to check 15% and decide what is righ**t*....


​this very statement highlights one issue that ​

​will always be a problem between Wikidata and Wikipedias. Wikipedia, at
least in my 10 years of experience on en:wp is that when you have multiple
sources that differ you highlight the existence of those ​sources and the
conflict of information  we dont decide what is right or wrong.

On 21 November 2015 at 06:35, Gerard Meijssen <[hidden email]>
wrote:

> Hoi,
> <grin> quality is different things </grin> I do care about quality but I do
> not necessarily agree with you how to best achieve it. Arguably bots are
> better and getting data into Wikidata than people. This means that the
> error rate of bots is typically better than what people do. It is all in
> the percentages.
>
> I have always said that the best way to improve quality is by comparing
> sources. When Wikidata has no data, it is arguably better to import data
> from any source. When the quality is 90% correct, there is already 100%
> more data. When 100% is compared with another source and 85% is the same,
> you only have to check 15% and decide what is right. When you compare with
> two distinct sources, the percentage that differs changes again.. :) In
> this way it makes sense to check errors
>
> It does not help when you state that either party has people that care or
> do not care about quality. By providing a high likelihood that something is
> problematic, you will learn who actually makes a difference. It however
> started with having data to compare in the first place
> Thanks,
>       GerardM
>
> On 20 November 2015 at 14:50, Petr Kadlec <[hidden email]> wrote:
>
> > On Fri, Nov 20, 2015 at 8:18 AM, Gerard Meijssen <
> > [hidden email]>
> > wrote:
> >
> > > When Wikipedia is a black box, not communicating about with the outside
> > > world, at some stage the situation becomes toxic. At this moment there
> > are
> > > already those at Wikidata that argue not to bother about Wikipedia
> > quality
> > > because in their view, Wikipedians do not care about its own quality.
> > >
> >
> > Right. When some users blindly dump random data to Wikidata, not
> > communicating about with the outside world, at some stage the situation
> > becomes toxic. At this moment there are already those at Wikipedia that
> > argue not to bother about Wikidata quality because in their view,
> > Wikidatans do not care about its own quality.
> >
> > For instance, take a look at
> > https://www.wikidata.org/wiki/User_talk:GerardM
> > https://www.wikidata.org/wiki/User_talk:GerardM/Archive_1
> >
> > Erm
> > -- [[cs:User:Mormegil | Petr Kadlec]]
> > _______________________________________________
> > Wikimedia-l mailing list, guidelines at:
> > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > [hidden email]
> > Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > <mailto:[hidden email]?subject=unsubscribe>
> >
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>



--
GN.
President Wikimedia Australia
WMAU: http://www.wikimedia.org.au/wiki/User:Gnangarra
Photo Gallery: http://gnangarra.redbubble.com
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] [Wiki-research-l] Quality issues

Craig Franklin
In reply to this post by WereSpielChequers-2
Indeed, the things that make a Wikipedia article high quality (such as well
written and engaging prose) are not necessarily the same things that are
useful for a data-driven product like Wikidata.  When Wikidata offers
assistance to another project, and that assistance is not received
enthusiastically because the project feels it will not improve their own
quality metrics; that is not a "black box" communication problem, nor is it
anyone in particular's fault, that is an issue of differing priorities.

Cheers,
Craig

On 20 November 2015 at 23:30, WereSpielChequers <[hidden email]
> wrote:

> My experience is that pretty much all Wikimedians care about quality,
> though some have different, even diametrically opposed views as to what
> quality means and which things are cosmetic or crucial.
>
> My experience of the sadly dormant death anomaly project
> <https://meta.wikimedia.org/wiki/Death_anomalies_table> was that people
> react positively to being told "here is a list of anomalies on your
> language wikipedia" especially if those anomalies are relatively serious.
> My experience of edits on many different languages is that wikipedians
> appreciate someone who improves articles, even if you don't speak their
> language. Dismissing any of our thousand wikis as a "black box" is I think
> less helpful.
>
> One of the great opportunities of Wikidata is to do the sort of data driven
> anomaly finding that we pioneered with the death anomalies report. But we
> always need to remember that there are cultural difference between wikis,
> and not just in such things as the age at which we assume people are dead.
> Diplomacy is a useful skill in cross wiki work.
>
> ~~~~
>
> On 20 November 2015 at 07:18, Gerard Meijssen <[hidden email]>
> wrote:
>
> > Hoi,
> > At Wikidata we often find issues with data imported from a Wikipedia.
> > Lists have been produced with these issues on the Wikipedia involved and
> > arguably they do present issues with the quality of Wikipedia or Wikidata
> > for that matter. So far hardly anything resulted from such outreach.
> >
> > When Wikipedia is a black box, not communicating about with the outside
> > world, at some stage the situation becomes toxic. At this moment there
> are
> > already those at Wikidata that argue not to bother about Wikipedia
> quality
> > because in their view, Wikipedians do not care about its own quality.
> >
> > Arguably known issues with quality are the easiest to solve.
> >
> > There are many ways to approach this subject. It is indeed a quality
> issue
> > both for Wikidata and Wikipedia. It can be seen as a research issue; how
> to
> > deal with quality and how do such mechanisms function if at all.
> >
> > I blogged about it..
> > Thanks,
> >      GerardM
> >
> >
> >
> http://ultimategerardm.blogspot.nl/2015/11/what-kind-of-box-is-wikipedia.html
> >
> > _______________________________________________
> > Wiki-research-l mailing list
> > [hidden email]
> > https://lists.wikimedia.org/mailman/listinfo/wiki-research-l
> >
> >
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] Quality issues

Peter Southwood
In reply to this post by Gerard Meijssen-3
How are you notifying the Wikipedias/Wikipedians? Do you leave a message on the talk page of the relevant article?
Cheers,
Peter

-----Original Message-----
From: Wikimedia-l [mailto:[hidden email]] On Behalf Of Gerard Meijssen
Sent: Saturday, 21 November 2015 12:23 AM
To: Wikimedia Mailing List
Subject: Re: [Wikimedia-l] Quality issues

Hoi,
So far such lists have been produced for bigger Wikipedias but essentially it is potentially an issue for any and all Wikis that have data that may exist on Wikidata or linked through Wikidata on external sources.
Thanks,
      GerardM

On 20 November 2015 at 12:33, Peter Southwood <[hidden email]>
wrote:

> Gerard,
> Who were you expecting would respond from the Wikipedias?
> Cheers,
> Peter
>
> -----Original Message-----
> From: Wikimedia-l [mailto:[hidden email]] On
> Behalf Of Gerard Meijssen
> Sent: Friday, 20 November 2015 9:18 AM
> To: Wikimedia Mailing List; Research into Wikimedia content and
> communities; WikiData-l
> Subject: [Wikimedia-l] Quality issues
>
> Hoi,
> At Wikidata we often find issues with data imported from a Wikipedia.
> Lists have been produced with these issues on the Wikipedia involved
> and arguably they do present issues with the quality of Wikipedia or
> Wikidata for that matter. So far hardly anything resulted from such outreach.
>
> When Wikipedia is a black box, not communicating about with the
> outside world, at some stage the situation becomes toxic. At this
> moment there are already those at Wikidata that argue not to bother
> about Wikipedia quality because in their view, Wikipedians do not care about its own quality.
>
> Arguably known issues with quality are the easiest to solve.
>
> There are many ways to approach this subject. It is indeed a quality
> issue both for Wikidata and Wikipedia. It can be seen as a research
> issue; how to deal with quality and how do such mechanisms function if at all.
>
> I blogged about it..
> Thanks,
>      GerardM
>
>
> http://ultimategerardm.blogspot.nl/2015/11/what-kind-of-box-is-wikiped
> ia.html _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>
> -----
> No virus found in this message.
> Checked by AVG - www.avg.com
> Version: 2016.0.7227 / Virus Database: 4460/11032 - Release Date:
> 11/20/15
>
>
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>

-----
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2016.0.7227 / Virus Database: 4460/11036 - Release Date: 11/20/15


_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] Quality issues

Gerard Meijssen-3
Hoi,
That is indeed a problem. So far it has been lists, often well formatted
lists that do not have a workflow, are not updated regularly. I have added
these issues as a wishlist item to work on. [1]

You have to appreciate that when a list of problematic issues is listed
with over 100 items, it is no longer easy or obvious that you want to add
and follow 100 talk pages.This is one of the big differences between
Wikipedia think and Wikidata think. I care about a lot of data, data that
is linked. Analogous to the "Kevin Bacon steps of separation" I want all
items easily and obviously connected. <grin> That is another quality goal
for Wikidata </grin>.

Given the state of Wikipedia, most articles have an article, easy and
obvious tasks like fact checking and adding sources is exactly what we are
looking for for maintaining our community. Add relevance to the cocktail,
we know that these facts are likely to have issues, and you appreciate why
this may help us with our quality and with our community issues.
Thanks,
     GerardM


[1]
https://meta.wikimedia.org/wiki/2015_Community_Wishlist_Survey#Visibility_for_quality_issues

On 21 November 2015 at 07:11, Peter Southwood <[hidden email]>
wrote:

> How are you notifying the Wikipedias/Wikipedians? Do you leave a message
> on the talk page of the relevant article?
> Cheers,
> Peter
>
> -----Original Message-----
> From: Wikimedia-l [mailto:[hidden email]] On
> Behalf Of Gerard Meijssen
> Sent: Saturday, 21 November 2015 12:23 AM
> To: Wikimedia Mailing List
> Subject: Re: [Wikimedia-l] Quality issues
>
> Hoi,
> So far such lists have been produced for bigger Wikipedias but essentially
> it is potentially an issue for any and all Wikis that have data that may
> exist on Wikidata or linked through Wikidata on external sources.
> Thanks,
>       GerardM
>
> On 20 November 2015 at 12:33, Peter Southwood <
> [hidden email]>
> wrote:
>
> > Gerard,
> > Who were you expecting would respond from the Wikipedias?
> > Cheers,
> > Peter
> >
> > -----Original Message-----
> > From: Wikimedia-l [mailto:[hidden email]] On
> > Behalf Of Gerard Meijssen
> > Sent: Friday, 20 November 2015 9:18 AM
> > To: Wikimedia Mailing List; Research into Wikimedia content and
> > communities; WikiData-l
> > Subject: [Wikimedia-l] Quality issues
> >
> > Hoi,
> > At Wikidata we often find issues with data imported from a Wikipedia.
> > Lists have been produced with these issues on the Wikipedia involved
> > and arguably they do present issues with the quality of Wikipedia or
> > Wikidata for that matter. So far hardly anything resulted from such
> outreach.
> >
> > When Wikipedia is a black box, not communicating about with the
> > outside world, at some stage the situation becomes toxic. At this
> > moment there are already those at Wikidata that argue not to bother
> > about Wikipedia quality because in their view, Wikipedians do not care
> about its own quality.
> >
> > Arguably known issues with quality are the easiest to solve.
> >
> > There are many ways to approach this subject. It is indeed a quality
> > issue both for Wikidata and Wikipedia. It can be seen as a research
> > issue; how to deal with quality and how do such mechanisms function if
> at all.
> >
> > I blogged about it..
> > Thanks,
> >      GerardM
> >
> >
> > http://ultimategerardm.blogspot.nl/2015/11/what-kind-of-box-is-wikiped
> > ia.html _______________________________________________
> > Wikimedia-l mailing list, guidelines at:
> > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > [hidden email]
> > Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > <mailto:[hidden email]?subject=unsubscribe>
> >
> > -----
> > No virus found in this message.
> > Checked by AVG - www.avg.com
> > Version: 2016.0.7227 / Virus Database: 4460/11032 - Release Date:
> > 11/20/15
> >
> >
> > _______________________________________________
> > Wikimedia-l mailing list, guidelines at:
> > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > [hidden email]
> > Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > <mailto:[hidden email]?subject=unsubscribe>
> >
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>
> -----
> No virus found in this message.
> Checked by AVG - www.avg.com
> Version: 2016.0.7227 / Virus Database: 4460/11036 - Release Date: 11/20/15
>
>
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] Quality issues

Gerard Meijssen-3
In reply to this post by Gnangarra
Hoi,
You conflate two issues. First when facts differ, it should be possible to
explain why they differ. Only when there is no explanation particularly
when there are no sources, there is an issue. In come real sources. When
someone died on 7-5-1759 and another source has a different date, it may be
the difference between a Julian and a Gregorian date. When a source makes
this plain, one fact has been proven to be incorrect. When the date was
1759, it is obvious that the other date is more precise.. The point is very
much that Wikipedia values sources and so does Wikidata. USE THEM and find
that data sources may be wrong when they are. In this way we improve
quality.

Many data sources have data from the same origin. It does not follow that
without original sources they are all right. Quite the reverse. It does
however take humans to be bold, to determine where a booboo has been made.
Yes, we do decide what is right or wrong, we do this when we research an
issue and that is exactly what this is about. It all starts with
determining a source.

In the mean time, Wikidata is negligent in stating sources. The worst
example is in the "primary sources" tool. It is bad because it is brought
to us as the best work flow for adding uncertain data to Wikidata. So the
world is not perfect but hey it is a wiki :)
Thanks,
      GerardM

On 21 November 2015 at 00:32, Gnangarra <[hidden email]> wrote:

> >
> > ...
> > *When 100% is compared with another source and 85% is the same,**you only
> > have to check 15% and decide what is righ**t*....
>
>
> ​this very statement highlights one issue that ​
>
> ​will always be a problem between Wikidata and Wikipedias. Wikipedia, at
> least in my 10 years of experience on en:wp is that when you have multiple
> sources that differ you highlight the existence of those ​sources and the
> conflict of information  we dont decide what is right or wrong.
>
> On 21 November 2015 at 06:35, Gerard Meijssen <[hidden email]>
> wrote:
>
> > Hoi,
> > <grin> quality is different things </grin> I do care about quality but I
> do
> > not necessarily agree with you how to best achieve it. Arguably bots are
> > better and getting data into Wikidata than people. This means that the
> > error rate of bots is typically better than what people do. It is all in
> > the percentages.
> >
> > I have always said that the best way to improve quality is by comparing
> > sources. When Wikidata has no data, it is arguably better to import data
> > from any source. When the quality is 90% correct, there is already 100%
> > more data. When 100% is compared with another source and 85% is the same,
> > you only have to check 15% and decide what is right. When you compare
> with
> > two distinct sources, the percentage that differs changes again.. :) In
> > this way it makes sense to check errors
> >
> > It does not help when you state that either party has people that care or
> > do not care about quality. By providing a high likelihood that something
> is
> > problematic, you will learn who actually makes a difference. It however
> > started with having data to compare in the first place
> > Thanks,
> >       GerardM
> >
> > On 20 November 2015 at 14:50, Petr Kadlec <[hidden email]> wrote:
> >
> > > On Fri, Nov 20, 2015 at 8:18 AM, Gerard Meijssen <
> > > [hidden email]>
> > > wrote:
> > >
> > > > When Wikipedia is a black box, not communicating about with the
> outside
> > > > world, at some stage the situation becomes toxic. At this moment
> there
> > > are
> > > > already those at Wikidata that argue not to bother about Wikipedia
> > > quality
> > > > because in their view, Wikipedians do not care about its own quality.
> > > >
> > >
> > > Right. When some users blindly dump random data to Wikidata, not
> > > communicating about with the outside world, at some stage the situation
> > > becomes toxic. At this moment there are already those at Wikipedia that
> > > argue not to bother about Wikidata quality because in their view,
> > > Wikidatans do not care about its own quality.
> > >
> > > For instance, take a look at
> > > https://www.wikidata.org/wiki/User_talk:GerardM
> > > https://www.wikidata.org/wiki/User_talk:GerardM/Archive_1
> > >
> > > Erm
> > > -- [[cs:User:Mormegil | Petr Kadlec]]
> > > _______________________________________________
> > > Wikimedia-l mailing list, guidelines at:
> > > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > > [hidden email]
> > > Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > > <mailto:[hidden email]?subject=unsubscribe>
> > >
> > _______________________________________________
> > Wikimedia-l mailing list, guidelines at:
> > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > [hidden email]
> > Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > <mailto:[hidden email]?subject=unsubscribe>
> >
>
>
>
> --
> GN.
> President Wikimedia Australia
> WMAU: http://www.wikimedia.org.au/wiki/User:Gnangarra
> Photo Gallery: http://gnangarra.redbubble.com
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] Quality issues

Peter Southwood
In reply to this post by Gerard Meijssen-3
The problem may simply be that the information is not coming to the attention of the people who care, as they don't know that it exists or where to find it. The normal place to put information relating to improvement of an article is on the article talk page, and that is where Wikipedians will expect to find it.
Cheers,
Peter

-----Original Message-----
From: Wikimedia-l [mailto:[hidden email]] On Behalf Of Gerard Meijssen
Sent: Saturday, 21 November 2015 9:57 AM
To: Wikimedia Mailing List
Subject: Re: [Wikimedia-l] Quality issues

Hoi,
That is indeed a problem. So far it has been lists, often well formatted lists that do not have a workflow, are not updated regularly. I have added these issues as a wishlist item to work on. [1]

You have to appreciate that when a list of problematic issues is listed with over 100 items, it is no longer easy or obvious that you want to add and follow 100 talk pages.This is one of the big differences between Wikipedia think and Wikidata think. I care about a lot of data, data that is linked. Analogous to the "Kevin Bacon steps of separation" I want all items easily and obviously connected. <grin> That is another quality goal for Wikidata </grin>.

Given the state of Wikipedia, most articles have an article, easy and obvious tasks like fact checking and adding sources is exactly what we are looking for for maintaining our community. Add relevance to the cocktail, we know that these facts are likely to have issues, and you appreciate why this may help us with our quality and with our community issues.
Thanks,
     GerardM


[1]
https://meta.wikimedia.org/wiki/2015_Community_Wishlist_Survey#Visibility_for_quality_issues

On 21 November 2015 at 07:11, Peter Southwood <[hidden email]>
wrote:

> How are you notifying the Wikipedias/Wikipedians? Do you leave a
> message on the talk page of the relevant article?
> Cheers,
> Peter
>
> -----Original Message-----
> From: Wikimedia-l [mailto:[hidden email]] On
> Behalf Of Gerard Meijssen
> Sent: Saturday, 21 November 2015 12:23 AM
> To: Wikimedia Mailing List
> Subject: Re: [Wikimedia-l] Quality issues
>
> Hoi,
> So far such lists have been produced for bigger Wikipedias but
> essentially it is potentially an issue for any and all Wikis that have
> data that may exist on Wikidata or linked through Wikidata on external sources.
> Thanks,
>       GerardM
>
> On 20 November 2015 at 12:33, Peter Southwood <
> [hidden email]>
> wrote:
>
> > Gerard,
> > Who were you expecting would respond from the Wikipedias?
> > Cheers,
> > Peter
> >
> > -----Original Message-----
> > From: Wikimedia-l [mailto:[hidden email]]
> > On Behalf Of Gerard Meijssen
> > Sent: Friday, 20 November 2015 9:18 AM
> > To: Wikimedia Mailing List; Research into Wikimedia content and
> > communities; WikiData-l
> > Subject: [Wikimedia-l] Quality issues
> >
> > Hoi,
> > At Wikidata we often find issues with data imported from a Wikipedia.
> > Lists have been produced with these issues on the Wikipedia involved
> > and arguably they do present issues with the quality of Wikipedia or
> > Wikidata for that matter. So far hardly anything resulted from such
> outreach.
> >
> > When Wikipedia is a black box, not communicating about with the
> > outside world, at some stage the situation becomes toxic. At this
> > moment there are already those at Wikidata that argue not to bother
> > about Wikipedia quality because in their view, Wikipedians do not
> > care
> about its own quality.
> >
> > Arguably known issues with quality are the easiest to solve.
> >
> > There are many ways to approach this subject. It is indeed a quality
> > issue both for Wikidata and Wikipedia. It can be seen as a research
> > issue; how to deal with quality and how do such mechanisms function
> > if
> at all.
> >
> > I blogged about it..
> > Thanks,
> >      GerardM
> >
> >
> > http://ultimategerardm.blogspot.nl/2015/11/what-kind-of-box-is-wikip
> > ed ia.html _______________________________________________
> > Wikimedia-l mailing list, guidelines at:
> > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > [hidden email]
> > Unsubscribe:
> > https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > <mailto:[hidden email]?subject=unsubscribe>
> >
> > -----
> > No virus found in this message.
> > Checked by AVG - www.avg.com
> > Version: 2016.0.7227 / Virus Database: 4460/11032 - Release Date:
> > 11/20/15
> >
> >
> > _______________________________________________
> > Wikimedia-l mailing list, guidelines at:
> > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > [hidden email]
> > Unsubscribe:
> > https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > <mailto:[hidden email]?subject=unsubscribe>
> >
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>
> -----
> No virus found in this message.
> Checked by AVG - www.avg.com
> Version: 2016.0.7227 / Virus Database: 4460/11036 - Release Date:
> 11/20/15
>
>
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>

-----
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2016.0.7227 / Virus Database: 4460/11038 - Release Date: 11/21/15


_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] Quality issues

Gnangarra
In reply to this post by Gerard Meijssen-3
>
> Many data sources have data from the same origin. It does not follow that
> without original sources they are all right. Quite the reverse. It does
> however take humans to be bold, to determine where a booboo has been made.
> Yes, we do decide what is right or wrong,


​No we dont decide what is right or wrong, en:wp has very specific core
policies about this

   - ​Original research - we dont draw conclusions from available data
   - NPOV​ -
   *which means presenting information without editorial bias*,
   ​ the moment we make that decision about whats right  we exceed the
   boundaries of our core pillars.... ​dont know, uncertain or conflicting
   information means exactly that we dont get to choose what we think is right


​The data article writers work with isnt black and white and its definitely
not set in stone Wikipedia content is a constant evolving collation of
knowledge, we should be careful when ever we put in place a process that
makes information definitive because people become reluctant to add to that
and they are even less likely to challenge something that has been cast in
stone already regardless of the inaccuracy of that casting .  We see it
within Wikipedia when articles are elevated to FA status with the number of
editors who fiercely defend that current/correct version against any
changes regardless of the merit in the information being added with
comments like "discuss it on talk page first" "revert good faith edit"


the more disjointed knowledge becomes the harder it is to keep it current,
accurate the more isolated that knowledge. Then power over making changes
takes precedence over productivity, accuracy and openness

On 21 November 2015 at 16:12, Gerard Meijssen <[hidden email]>
wrote:

> Hoi,
> You conflate two issues. First when facts differ, it should be possible to
> explain why they differ. Only when there is no explanation particularly
> when there are no sources, there is an issue. In come real sources. When
> someone died on 7-5-1759 and another source has a different date, it may be
> the difference between a Julian and a Gregorian date. When a source makes
> this plain, one fact has been proven to be incorrect. When the date was
> 1759, it is obvious that the other date is more precise.. The point is very
> much that Wikipedia values sources and so does Wikidata. USE THEM and find
> that data sources may be wrong when they are. In this way we improve
> quality.
>
> Many data sources have data from the same origin. It does not follow that
> without original sources they are all right. Quite the reverse. It does
> however take humans to be bold, to determine where a booboo has been made.
> Yes, we do decide what is right or wrong, we do this when we research an
> issue and that is exactly what this is about. It all starts with
> determining a source.
>
> In the mean time, Wikidata is negligent in stating sources. The worst
> example is in the "primary sources" tool. It is bad because it is brought
> to us as the best work flow for adding uncertain data to Wikidata. So the
> world is not perfect but hey it is a wiki :)
> Thanks,
>       GerardM
>
> On 21 November 2015 at 00:32, Gnangarra <[hidden email]> wrote:
>
> > >
> > > ...
> > > *When 100% is compared with another source and 85% is the same,**you
> only
> > > have to check 15% and decide what is righ**t*....
> >
> >
> > ​this very statement highlights one issue that ​
> >
> > ​will always be a problem between Wikidata and Wikipedias. Wikipedia, at
> > least in my 10 years of experience on en:wp is that when you have
> multiple
> > sources that differ you highlight the existence of those ​sources and the
> > conflict of information  we dont decide what is right or wrong.
> >
> > On 21 November 2015 at 06:35, Gerard Meijssen <[hidden email]
> >
> > wrote:
> >
> > > Hoi,
> > > <grin> quality is different things </grin> I do care about quality but
> I
> > do
> > > not necessarily agree with you how to best achieve it. Arguably bots
> are
> > > better and getting data into Wikidata than people. This means that the
> > > error rate of bots is typically better than what people do. It is all
> in
> > > the percentages.
> > >
> > > I have always said that the best way to improve quality is by comparing
> > > sources. When Wikidata has no data, it is arguably better to import
> data
> > > from any source. When the quality is 90% correct, there is already 100%
> > > more data. When 100% is compared with another source and 85% is the
> same,
> > > you only have to check 15% and decide what is right. When you compare
> > with
> > > two distinct sources, the percentage that differs changes again.. :) In
> > > this way it makes sense to check errors
> > >
> > > It does not help when you state that either party has people that care
> or
> > > do not care about quality. By providing a high likelihood that
> something
> > is
> > > problematic, you will learn who actually makes a difference. It however
> > > started with having data to compare in the first place
> > > Thanks,
> > >       GerardM
> > >
> > > On 20 November 2015 at 14:50, Petr Kadlec <[hidden email]>
> wrote:
> > >
> > > > On Fri, Nov 20, 2015 at 8:18 AM, Gerard Meijssen <
> > > > [hidden email]>
> > > > wrote:
> > > >
> > > > > When Wikipedia is a black box, not communicating about with the
> > outside
> > > > > world, at some stage the situation becomes toxic. At this moment
> > there
> > > > are
> > > > > already those at Wikidata that argue not to bother about Wikipedia
> > > > quality
> > > > > because in their view, Wikipedians do not care about its own
> quality.
> > > > >
> > > >
> > > > Right. When some users blindly dump random data to Wikidata, not
> > > > communicating about with the outside world, at some stage the
> situation
> > > > becomes toxic. At this moment there are already those at Wikipedia
> that
> > > > argue not to bother about Wikidata quality because in their view,
> > > > Wikidatans do not care about its own quality.
> > > >
> > > > For instance, take a look at
> > > > https://www.wikidata.org/wiki/User_talk:GerardM
> > > > https://www.wikidata.org/wiki/User_talk:GerardM/Archive_1
> > > >
> > > > Erm
> > > > -- [[cs:User:Mormegil | Petr Kadlec]]
> > > > _______________________________________________
> > > > Wikimedia-l mailing list, guidelines at:
> > > > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > > > [hidden email]
> > > > Unsubscribe:
> https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > > > <mailto:[hidden email]?subject=unsubscribe>
> > > >
> > > _______________________________________________
> > > Wikimedia-l mailing list, guidelines at:
> > > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > > [hidden email]
> > > Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > > <mailto:[hidden email]?subject=unsubscribe>
> > >
> >
> >
> >
> > --
> > GN.
> > President Wikimedia Australia
> > WMAU: http://www.wikimedia.org.au/wiki/User:Gnangarra
> > Photo Gallery: http://gnangarra.redbubble.com
> > _______________________________________________
> > Wikimedia-l mailing list, guidelines at:
> > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > [hidden email]
> > Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > <mailto:[hidden email]?subject=unsubscribe>
> >
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>



--
GN.
President Wikimedia Australia
WMAU: http://www.wikimedia.org.au/wiki/User:Gnangarra
Photo Gallery: http://gnangarra.redbubble.com
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] Quality issues

Gerard Meijssen-3
Hoi,
I respect the policy of Wikipedia. However, when multiple Wikipedias differ
and when there is no sourcing does this policy hold? When Wikidata has no
attributable sources but multiple statements is it not conceivable that
things are easy and obvious.. that they are wrong?

When you talk about the FA status of articles, you are considering
something totally alien to what is at stake. Typically we do not have
credible sources at Wikidata and typically there is an issue with the data.

When Wikidata is as mature as en.wp we will have on average 10 statements
for every item. Currently half of our items have at most two statements. We
do find issues in any source by comparing them. It does make sense to make
this effort. It is an obvious way of improving quality in all of our
projects and even beyond that.
Thanks,
     GerardM

On 21 November 2015 at 10:26, Gnangarra <[hidden email]> wrote:

> >
> > Many data sources have data from the same origin. It does not follow that
> > without original sources they are all right. Quite the reverse. It does
> > however take humans to be bold, to determine where a booboo has been
> made.
> > Yes, we do decide what is right or wrong,
>
>
> ​No we dont decide what is right or wrong, en:wp has very specific core
> policies about this
>
>    - ​Original research - we dont draw conclusions from available data
>    - NPOV​ -
>    *which means presenting information without editorial bias*,
>    ​ the moment we make that decision about whats right  we exceed the
>    boundaries of our core pillars.... ​dont know, uncertain or conflicting
>    information means exactly that we dont get to choose what we think is
> right
>
>
> ​The data article writers work with isnt black and white and its definitely
> not set in stone Wikipedia content is a constant evolving collation of
> knowledge, we should be careful when ever we put in place a process that
> makes information definitive because people become reluctant to add to that
> and they are even less likely to challenge something that has been cast in
> stone already regardless of the inaccuracy of that casting .  We see it
> within Wikipedia when articles are elevated to FA status with the number of
> editors who fiercely defend that current/correct version against any
> changes regardless of the merit in the information being added with
> comments like "discuss it on talk page first" "revert good faith edit"
>
>
> the more disjointed knowledge becomes the harder it is to keep it current,
> accurate the more isolated that knowledge. Then power over making changes
> takes precedence over productivity, accuracy and openness
>
> On 21 November 2015 at 16:12, Gerard Meijssen <[hidden email]>
> wrote:
>
> > Hoi,
> > You conflate two issues. First when facts differ, it should be possible
> to
> > explain why they differ. Only when there is no explanation particularly
> > when there are no sources, there is an issue. In come real sources. When
> > someone died on 7-5-1759 and another source has a different date, it may
> be
> > the difference between a Julian and a Gregorian date. When a source makes
> > this plain, one fact has been proven to be incorrect. When the date was
> > 1759, it is obvious that the other date is more precise.. The point is
> very
> > much that Wikipedia values sources and so does Wikidata. USE THEM and
> find
> > that data sources may be wrong when they are. In this way we improve
> > quality.
> >
> > Many data sources have data from the same origin. It does not follow that
> > without original sources they are all right. Quite the reverse. It does
> > however take humans to be bold, to determine where a booboo has been
> made.
> > Yes, we do decide what is right or wrong, we do this when we research an
> > issue and that is exactly what this is about. It all starts with
> > determining a source.
> >
> > In the mean time, Wikidata is negligent in stating sources. The worst
> > example is in the "primary sources" tool. It is bad because it is brought
> > to us as the best work flow for adding uncertain data to Wikidata. So the
> > world is not perfect but hey it is a wiki :)
> > Thanks,
> >       GerardM
> >
> > On 21 November 2015 at 00:32, Gnangarra <[hidden email]> wrote:
> >
> > > >
> > > > ...
> > > > *When 100% is compared with another source and 85% is the same,**you
> > only
> > > > have to check 15% and decide what is righ**t*....
> > >
> > >
> > > ​this very statement highlights one issue that ​
> > >
> > > ​will always be a problem between Wikidata and Wikipedias. Wikipedia,
> at
> > > least in my 10 years of experience on en:wp is that when you have
> > multiple
> > > sources that differ you highlight the existence of those ​sources and
> the
> > > conflict of information  we dont decide what is right or wrong.
> > >
> > > On 21 November 2015 at 06:35, Gerard Meijssen <
> [hidden email]
> > >
> > > wrote:
> > >
> > > > Hoi,
> > > > <grin> quality is different things </grin> I do care about quality
> but
> > I
> > > do
> > > > not necessarily agree with you how to best achieve it. Arguably bots
> > are
> > > > better and getting data into Wikidata than people. This means that
> the
> > > > error rate of bots is typically better than what people do. It is all
> > in
> > > > the percentages.
> > > >
> > > > I have always said that the best way to improve quality is by
> comparing
> > > > sources. When Wikidata has no data, it is arguably better to import
> > data
> > > > from any source. When the quality is 90% correct, there is already
> 100%
> > > > more data. When 100% is compared with another source and 85% is the
> > same,
> > > > you only have to check 15% and decide what is right. When you compare
> > > with
> > > > two distinct sources, the percentage that differs changes again.. :)
> In
> > > > this way it makes sense to check errors
> > > >
> > > > It does not help when you state that either party has people that
> care
> > or
> > > > do not care about quality. By providing a high likelihood that
> > something
> > > is
> > > > problematic, you will learn who actually makes a difference. It
> however
> > > > started with having data to compare in the first place
> > > > Thanks,
> > > >       GerardM
> > > >
> > > > On 20 November 2015 at 14:50, Petr Kadlec <[hidden email]>
> > wrote:
> > > >
> > > > > On Fri, Nov 20, 2015 at 8:18 AM, Gerard Meijssen <
> > > > > [hidden email]>
> > > > > wrote:
> > > > >
> > > > > > When Wikipedia is a black box, not communicating about with the
> > > outside
> > > > > > world, at some stage the situation becomes toxic. At this moment
> > > there
> > > > > are
> > > > > > already those at Wikidata that argue not to bother about
> Wikipedia
> > > > > quality
> > > > > > because in their view, Wikipedians do not care about its own
> > quality.
> > > > > >
> > > > >
> > > > > Right. When some users blindly dump random data to Wikidata, not
> > > > > communicating about with the outside world, at some stage the
> > situation
> > > > > becomes toxic. At this moment there are already those at Wikipedia
> > that
> > > > > argue not to bother about Wikidata quality because in their view,
> > > > > Wikidatans do not care about its own quality.
> > > > >
> > > > > For instance, take a look at
> > > > > https://www.wikidata.org/wiki/User_talk:GerardM
> > > > > https://www.wikidata.org/wiki/User_talk:GerardM/Archive_1
> > > > >
> > > > > Erm
> > > > > -- [[cs:User:Mormegil | Petr Kadlec]]
> > > > > _______________________________________________
> > > > > Wikimedia-l mailing list, guidelines at:
> > > > > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > > > > [hidden email]
> > > > > Unsubscribe:
> > https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > > > > <mailto:[hidden email]
> ?subject=unsubscribe>
> > > > >
> > > > _______________________________________________
> > > > Wikimedia-l mailing list, guidelines at:
> > > > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > > > [hidden email]
> > > > Unsubscribe:
> https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > > > <mailto:[hidden email]?subject=unsubscribe>
> > > >
> > >
> > >
> > >
> > > --
> > > GN.
> > > President Wikimedia Australia
> > > WMAU: http://www.wikimedia.org.au/wiki/User:Gnangarra
> > > Photo Gallery: http://gnangarra.redbubble.com
> > > _______________________________________________
> > > Wikimedia-l mailing list, guidelines at:
> > > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > > [hidden email]
> > > Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > > <mailto:[hidden email]?subject=unsubscribe>
> > >
> > _______________________________________________
> > Wikimedia-l mailing list, guidelines at:
> > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > [hidden email]
> > Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > <mailto:[hidden email]?subject=unsubscribe>
> >
>
>
>
> --
> GN.
> President Wikimedia Australia
> WMAU: http://www.wikimedia.org.au/wiki/User:Gnangarra
> Photo Gallery: http://gnangarra.redbubble.com
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
Reply | Threaded
Open this post in threaded view
|

Re: [Wikimedia-l] Quality issues

Gnangarra
agree getting information in is in and of itself a good starting point but
ignoring the lessons learnt in other project in doing so is only creating
more work for those that follow. Having less clear policy about sources and
allowing unsourced information is only going to put Wikidata behind
Wikipedia in quality, in doing so its not going to endear WikiData
information to Wikipedians which in turn Wikipedians as they get data just
arent going to go that extra step to share no matter how easy the step is
to take

On 21 November 2015 at 19:13, Gerard Meijssen <[hidden email]>
wrote:

> Hoi,
> I respect the policy of Wikipedia. However, when multiple Wikipedias differ
> and when there is no sourcing does this policy hold? When Wikidata has no
> attributable sources but multiple statements is it not conceivable that
> things are easy and obvious.. that they are wrong?
>
> When you talk about the FA status of articles, you are considering
> something totally alien to what is at stake. Typically we do not have
> credible sources at Wikidata and typically there is an issue with the data.
>
> When Wikidata is as mature as en.wp we will have on average 10 statements
> for every item. Currently half of our items have at most two statements. We
> do find issues in any source by comparing them. It does make sense to make
> this effort. It is an obvious way of improving quality in all of our
> projects and even beyond that.
> Thanks,
>      GerardM
>
> On 21 November 2015 at 10:26, Gnangarra <[hidden email]> wrote:
>
> > >
> > > Many data sources have data from the same origin. It does not follow
> that
> > > without original sources they are all right. Quite the reverse. It does
> > > however take humans to be bold, to determine where a booboo has been
> > made.
> > > Yes, we do decide what is right or wrong,
> >
> >
> > ​No we dont decide what is right or wrong, en:wp has very specific core
> > policies about this
> >
> >    - ​Original research - we dont draw conclusions from available data
> >    - NPOV​ -
> >    *which means presenting information without editorial bias*,
> >    ​ the moment we make that decision about whats right  we exceed the
> >    boundaries of our core pillars.... ​dont know, uncertain or
> conflicting
> >    information means exactly that we dont get to choose what we think is
> > right
> >
> >
> > ​The data article writers work with isnt black and white and its
> definitely
> > not set in stone Wikipedia content is a constant evolving collation of
> > knowledge, we should be careful when ever we put in place a process that
> > makes information definitive because people become reluctant to add to
> that
> > and they are even less likely to challenge something that has been cast
> in
> > stone already regardless of the inaccuracy of that casting .  We see it
> > within Wikipedia when articles are elevated to FA status with the number
> of
> > editors who fiercely defend that current/correct version against any
> > changes regardless of the merit in the information being added with
> > comments like "discuss it on talk page first" "revert good faith edit"
> >
> >
> > the more disjointed knowledge becomes the harder it is to keep it
> current,
> > accurate the more isolated that knowledge. Then power over making changes
> > takes precedence over productivity, accuracy and openness
> >
> > On 21 November 2015 at 16:12, Gerard Meijssen <[hidden email]
> >
> > wrote:
> >
> > > Hoi,
> > > You conflate two issues. First when facts differ, it should be possible
> > to
> > > explain why they differ. Only when there is no explanation particularly
> > > when there are no sources, there is an issue. In come real sources.
> When
> > > someone died on 7-5-1759 and another source has a different date, it
> may
> > be
> > > the difference between a Julian and a Gregorian date. When a source
> makes
> > > this plain, one fact has been proven to be incorrect. When the date was
> > > 1759, it is obvious that the other date is more precise.. The point is
> > very
> > > much that Wikipedia values sources and so does Wikidata. USE THEM and
> > find
> > > that data sources may be wrong when they are. In this way we improve
> > > quality.
> > >
> > > Many data sources have data from the same origin. It does not follow
> that
> > > without original sources they are all right. Quite the reverse. It does
> > > however take humans to be bold, to determine where a booboo has been
> > made.
> > > Yes, we do decide what is right or wrong, we do this when we research
> an
> > > issue and that is exactly what this is about. It all starts with
> > > determining a source.
> > >
> > > In the mean time, Wikidata is negligent in stating sources. The worst
> > > example is in the "primary sources" tool. It is bad because it is
> brought
> > > to us as the best work flow for adding uncertain data to Wikidata. So
> the
> > > world is not perfect but hey it is a wiki :)
> > > Thanks,
> > >       GerardM
> > >
> > > On 21 November 2015 at 00:32, Gnangarra <[hidden email]> wrote:
> > >
> > > > >
> > > > > ...
> > > > > *When 100% is compared with another source and 85% is the
> same,**you
> > > only
> > > > > have to check 15% and decide what is righ**t*....
> > > >
> > > >
> > > > ​this very statement highlights one issue that ​
> > > >
> > > > ​will always be a problem between Wikidata and Wikipedias. Wikipedia,
> > at
> > > > least in my 10 years of experience on en:wp is that when you have
> > > multiple
> > > > sources that differ you highlight the existence of those ​sources and
> > the
> > > > conflict of information  we dont decide what is right or wrong.
> > > >
> > > > On 21 November 2015 at 06:35, Gerard Meijssen <
> > [hidden email]
> > > >
> > > > wrote:
> > > >
> > > > > Hoi,
> > > > > <grin> quality is different things </grin> I do care about quality
> > but
> > > I
> > > > do
> > > > > not necessarily agree with you how to best achieve it. Arguably
> bots
> > > are
> > > > > better and getting data into Wikidata than people. This means that
> > the
> > > > > error rate of bots is typically better than what people do. It is
> all
> > > in
> > > > > the percentages.
> > > > >
> > > > > I have always said that the best way to improve quality is by
> > comparing
> > > > > sources. When Wikidata has no data, it is arguably better to import
> > > data
> > > > > from any source. When the quality is 90% correct, there is already
> > 100%
> > > > > more data. When 100% is compared with another source and 85% is the
> > > same,
> > > > > you only have to check 15% and decide what is right. When you
> compare
> > > > with
> > > > > two distinct sources, the percentage that differs changes again..
> :)
> > In
> > > > > this way it makes sense to check errors
> > > > >
> > > > > It does not help when you state that either party has people that
> > care
> > > or
> > > > > do not care about quality. By providing a high likelihood that
> > > something
> > > > is
> > > > > problematic, you will learn who actually makes a difference. It
> > however
> > > > > started with having data to compare in the first place
> > > > > Thanks,
> > > > >       GerardM
> > > > >
> > > > > On 20 November 2015 at 14:50, Petr Kadlec <[hidden email]>
> > > wrote:
> > > > >
> > > > > > On Fri, Nov 20, 2015 at 8:18 AM, Gerard Meijssen <
> > > > > > [hidden email]>
> > > > > > wrote:
> > > > > >
> > > > > > > When Wikipedia is a black box, not communicating about with the
> > > > outside
> > > > > > > world, at some stage the situation becomes toxic. At this
> moment
> > > > there
> > > > > > are
> > > > > > > already those at Wikidata that argue not to bother about
> > Wikipedia
> > > > > > quality
> > > > > > > because in their view, Wikipedians do not care about its own
> > > quality.
> > > > > > >
> > > > > >
> > > > > > Right. When some users blindly dump random data to Wikidata, not
> > > > > > communicating about with the outside world, at some stage the
> > > situation
> > > > > > becomes toxic. At this moment there are already those at
> Wikipedia
> > > that
> > > > > > argue not to bother about Wikidata quality because in their view,
> > > > > > Wikidatans do not care about its own quality.
> > > > > >
> > > > > > For instance, take a look at
> > > > > > https://www.wikidata.org/wiki/User_talk:GerardM
> > > > > > https://www.wikidata.org/wiki/User_talk:GerardM/Archive_1
> > > > > >
> > > > > > Erm
> > > > > > -- [[cs:User:Mormegil | Petr Kadlec]]
> > > > > > _______________________________________________
> > > > > > Wikimedia-l mailing list, guidelines at:
> > > > > > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > > > > > [hidden email]
> > > > > > Unsubscribe:
> > > https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > > > > > <mailto:[hidden email]
> > ?subject=unsubscribe>
> > > > > >
> > > > > _______________________________________________
> > > > > Wikimedia-l mailing list, guidelines at:
> > > > > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > > > > [hidden email]
> > > > > Unsubscribe:
> > https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > > > > <mailto:[hidden email]
> ?subject=unsubscribe>
> > > > >
> > > >
> > > >
> > > >
> > > > --
> > > > GN.
> > > > President Wikimedia Australia
> > > > WMAU: http://www.wikimedia.org.au/wiki/User:Gnangarra
> > > > Photo Gallery: http://gnangarra.redbubble.com
> > > > _______________________________________________
> > > > Wikimedia-l mailing list, guidelines at:
> > > > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > > > [hidden email]
> > > > Unsubscribe:
> https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > > > <mailto:[hidden email]?subject=unsubscribe>
> > > >
> > > _______________________________________________
> > > Wikimedia-l mailing list, guidelines at:
> > > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > > [hidden email]
> > > Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > > <mailto:[hidden email]?subject=unsubscribe>
> > >
> >
> >
> >
> > --
> > GN.
> > President Wikimedia Australia
> > WMAU: http://www.wikimedia.org.au/wiki/User:Gnangarra
> > Photo Gallery: http://gnangarra.redbubble.com
> > _______________________________________________
> > Wikimedia-l mailing list, guidelines at:
> > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > [hidden email]
> > Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > <mailto:[hidden email]?subject=unsubscribe>
> >
> _______________________________________________
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> [hidden email]
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> <mailto:[hidden email]?subject=unsubscribe>
>



--
GN.
President Wikimedia Australia
WMAU: http://www.wikimedia.org.au/wiki/User:Gnangarra
Photo Gallery: http://gnangarra.redbubble.com
_______________________________________________
Wikimedia-l mailing list, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
[hidden email]
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, <mailto:[hidden email]?subject=unsubscribe>
1234 ... 9