Re: [Wikimedia-l] Wikidata training in Oceania, SE Asia (Was: Quality issues)

[Gerard Meijssen]

Hoi,
I am really pleased that this is going to happen. We can do with much more
information about Australia and Indonesia in Wikidata.
Thanks,
 GerardM

On 21 December 2015 at 08:26, Gnangarra  wrote:

> This is a followup considered from the now defunct WMAU conference in QLD
> with that falling through the committee decided to continue this part as it
> was one of the important features that needs to be better understood
> locally.
>
> Details will be sent out after the silly season in short it will start in
> mid febraury in Melbourne coinciding with a libraries conference there,
> from there upto sydney, then onto canberra to coincide with a conference
> there then across to Perth from Perth the person will be heading home via
> Dubai.  There is a mini conference planned in Perth that WMID would be
> welcome to send  people to with a workshop focusing on experienced
> contributors to other projects on the next day as well.
>
> Full details will be  published in Early January
>
> On 21 December 2015 at 10:08, John Mark Vandenberg 
> wrote:
>
> > On Mon, Dec 21, 2015 at 11:22 AM, Gnangarra  wrote:
> > > This is going nowhere, one of the big issues is that there is a lack of
> > > understanding on how WikiData works and whats it purpose is.
> > >
> > >
> > > Wikimedia Australia solution is invest community money into bringing
> > > someone who has contributed to Wikidata to Australia to do a series of
> > > talks and workshops around the country over a three week period.
> >
> > Ah, that sounds interesting.  Could you provide more information about
> > that please?  I havent seen this mentioned on the public mailing lists
> > before.
> >
> > Wikimedia Indonesia is doing a Wikidata project next year and lapping
> > up all the training they can get.  I am sure they would be interested
> > in having a trainer stop over in Jakarta on the way to or from
> > Australia.
> >
> > --
> > John Vandenberg
> >
> > ___
> > Wikimedia-l mailing list, guidelines at:
> > https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> > New messages to: Wikimedia-l@lists.wikimedia.org
> > Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> > 
>
>
>
>
> --
> 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
> New messages to: Wikimedia-l@lists.wikimedia.org
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> 
>
___
Wikimedia-l mailing list, guidelines at: 
https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
New messages to: Wikimedia-l@lists.wikimedia.org
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, 

Re: [Wikimedia-l] Wikidata training in Oceania, SE Asia (Was: Quality issues)

[Gnangarra]

This is a followup considered from the now defunct WMAU conference in QLD
with that falling through the committee decided to continue this part as it
was one of the important features that needs to be better understood
locally.

Details will be sent out after the silly season in short it will start in
mid febraury in Melbourne coinciding with a libraries conference there,
from there upto sydney, then onto canberra to coincide with a conference
there then across to Perth from Perth the person will be heading home via
Dubai.  There is a mini conference planned in Perth that WMID would be
welcome to send  people to with a workshop focusing on experienced
contributors to other projects on the next day as well.

Full details will be  published in Early January

On 21 December 2015 at 10:08, John Mark Vandenberg  wrote:

> On Mon, Dec 21, 2015 at 11:22 AM, Gnangarra  wrote:
> > This is going nowhere, one of the big issues is that there is a lack of
> > understanding on how WikiData works and whats it purpose is.
> >
> >
> > Wikimedia Australia solution is invest community money into bringing
> > someone who has contributed to Wikidata to Australia to do a series of
> > talks and workshops around the country over a three week period.
>
> Ah, that sounds interesting.  Could you provide more information about
> that please?  I havent seen this mentioned on the public mailing lists
> before.
>
> Wikimedia Indonesia is doing a Wikidata project next year and lapping
> up all the training they can get.  I am sure they would be interested
> in having a trainer stop over in Jakarta on the way to or from
> Australia.
>
> --
> John Vandenberg
>
> ___
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> New messages to: Wikimedia-l@lists.wikimedia.org
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> 




-- 
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
New messages to: Wikimedia-l@lists.wikimedia.org
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, 

[Wikimedia-l] Wikidata training in Oceania, SE Asia (Was: Quality issues)

[John Mark Vandenberg]

On Mon, Dec 21, 2015 at 11:22 AM, Gnangarra  wrote:
> This is going nowhere, one of the big issues is that there is a lack of
> understanding on how WikiData works and whats it purpose is.
>
>
> Wikimedia Australia solution is invest community money into bringing
> someone who has contributed to Wikidata to Australia to do a series of
> talks and workshops around the country over a three week period.

Ah, that sounds interesting.  Could you provide more information about
that please?  I havent seen this mentioned on the public mailing lists
before.

Wikimedia Indonesia is doing a Wikidata project next year and lapping
up all the training they can get.  I am sure they would be interested
in having a trainer stop over in Jakarta on the way to or from
Australia.

-- 
John Vandenberg

___
Wikimedia-l mailing list, guidelines at: 
https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
New messages to: Wikimedia-l@lists.wikimedia.org
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, 

Re: [Wikimedia-l] Quality issues

[Gnangarra]

This is going nowhere, one of the big issues is that there is a lack of
understanding on how WikiData works and whats it purpose is.


Wikimedia Australia solution is invest community money into bringing
someone who has contributed to Wikidata to Australia to do a series of
talks and workshops around the country over a three week period.


Perhaps Wikidata community could improve the wider Wikimedian communities
understanding by doing the same on a larger scale get out to where the
contributors are, answer question explain how it works and do workshops
aimed directly at existing Wikiproject contributors. obviously its a lot
simpler and cheaper in places like Europe/US where there are already active
contributors who have the knowledge

On 21 December 2015 at 08:00, Fabian Flöck  wrote:

> Not a contribution to the discussion at large, but I had the same problem
> Andreas is mentioning a couple of days ago when doing some of my first WD
> edits and adding a reference. I had no idea what to chose in that property
> field and it only showed me “instance of” and “subclass of” as a cold
> start. (I guess your average new editor might even wonder why to enter a
> property at all there - they would probably expect a single field to enter
> the source.)  So I just went to some other items and checked how it was
> done there, which is not optimal. A pre-selection of relevant properties
> (maybe most used in other items?) in the type-ahead would be nice. And
> maybe a small explanation of what the property for references means
> (something like “specifies type of reference” ?). I was also unsure if and
> when to ever use “imported from” in that field (i.e., if I got the fact
> from a Wikipedia page, but no primary source exists) or if that was
> reserved for machine imports.
>
> Fabian
>
>
> > Date: Sun, 20 Dec 2015 15:59:58 +
> > From: Andreas Kolbe mailto:jayen...@gmail.com>>
> > To: Wikimedia Mailing List  wikimedia-l@lists.wikimedia.org>>
> > Subject: Re: [Wikimedia-l] Quality issues
> >>
> > Just try it, Lydia. Click "add" in subsidiaries in
> > https://www.wikidata.org/wiki/Q37156 <
> https://www.wikidata.org/wiki/Q37156> -- enter a company name, and then
> > click "add reference". When I do that, the text field contains a
> greyed-out
> > "property", and the drop-down shows the unhelpful items I mentioned
> above.
> >
> > And it would be good if the help text actually *asked* people to cite a
> > reference.
> ___
> Wikimedia-l mailing list, guidelines at:
> https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
> New messages to: Wikimedia-l@lists.wikimedia.org
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l,
> 
>



-- 
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
New messages to: Wikimedia-l@lists.wikimedia.org
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, 

Re: [Wikimedia-l] Quality issues

[Fabian Flöck]

Not a contribution to the discussion at large, but I had the same problem 
Andreas is mentioning a couple of days ago when doing some of my first WD edits 
and adding a reference. I had no idea what to chose in that property field and 
it only showed me “instance of” and “subclass of” as a cold start. (I guess 
your average new editor might even wonder why to enter a property at all there 
- they would probably expect a single field to enter the source.)  So I just 
went to some other items and checked how it was done there, which is not 
optimal. A pre-selection of relevant properties (maybe most used in other 
items?) in the type-ahead would be nice. And maybe a small explanation of what 
the property for references means (something like “specifies type of reference” 
?). I was also unsure if and when to ever use “imported from” in that field 
(i.e., if I got the fact from a Wikipedia page, but no primary source exists) 
or if that was reserved for machine imports.

Fabian


> Date: Sun, 20 Dec 2015 15:59:58 +
> From: Andreas Kolbe mailto:jayen...@gmail.com>>
> To: Wikimedia Mailing List  >
> Subject: Re: [Wikimedia-l] Quality issues
>> 
> Just try it, Lydia. Click "add" in subsidiaries in
> https://www.wikidata.org/wiki/Q37156  
> -- enter a company name, and then
> click "add reference". When I do that, the text field contains a greyed-out
> "property", and the drop-down shows the unhelpful items I mentioned above.
> 
> And it would be good if the help text actually *asked* people to cite a
> reference.
___
Wikimedia-l mailing list, guidelines at: 
https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
New messages to: Wikimedia-l@lists.wikimedia.org
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, 

Re: [Wikimedia-l] Quality issues

[geni]

On 20 December 2015 at 15:59, Andreas Kolbe  wrote:

>
> Can you tell me just whose interests it serves if re-users do not have to
> indicate that the data they're showing their users come from Wikidata? Max
> Klein mused that the big search engines might be paying for Wikidata "to
> remove a blemish on their perceived omniscience", because they can present
> Wikidata content as though they had compiled it themselves.[1] That is at
> least a plausible line of thought; but whom else does it serve?
>
>
Anyone who doesn't want to spend way too much of their time worrying about
copyright law.



>
> So, who in society benefits, other than (arguably) the big commercial
> search engines?
>


The big search engine companies could make their in house version with
little effort. An open version means that they have to put up with everyone
else having access to it.

-- 
geni
___
Wikimedia-l mailing list, guidelines at: 
https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
New messages to: Wikimedia-l@lists.wikimedia.org
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, 

Re: [Wikimedia-l] [Wikitech-l] IRC office hours: Shared hosting

[Brian Wolff]

On 12/20/15, James Salsman  wrote:
> Were there any objections to my request below?
>

Yes. As MaxSem said earlier[1], its basically being ignored as being
totally irrelevant to the topic at hand. (To be clear: Third-party
does not mean people who are doing work on Wikimedia sites that aren't
WMF. Third party = Wikis that have nothing to do with Wikimedia wikis
(e.g. wikia, wikihow, uncyclopedia etc))

If you want to get Dispenser his hard disk space, you should take it
up with the labs people, or at the very least some thread where it
would be on-topic.

> Can we also please hire additional database, system, and if necessary
> network administration support to make sure that the third party spam
> prevention bot infrastructure is supported more robustly in the future?

Then by definition it wouldn't be a third-party spam framework if WMF
was running it.

--
-bawolff

[1] https://lists.wikimedia.org/pipermail/wikitech-l/2015-December/084326.html
[Linking because this thread is super-cross posted, and some people
are going to be confused as to what I'm referring to]

___
Wikimedia-l mailing list, guidelines at: 
https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
New messages to: Wikimedia-l@lists.wikimedia.org
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, 

Re: [Wikimedia-l] Community Wishlist Survey: Top 10 wishes!

[Trillium Corsage]

17.12.2015, 01:26, "Sam Klein" :
> Thanks to all for organizing the survey and for sharing!



> And a mentor-friendly feature I've wanted for a long time:
> #10. Add a user watchlist

That's not only mentor-friendly, it's hounder-friendly and harrasser-friendly. 
Really, you should have a look at the seamier side of (at least) English 
WIkipedia. There are people, particularly among the administrative set, with 
null interest in articles but consumed with interpersonal conflict and 
targeting of others. It's not a small problem.

Trillium Corsage

___
Wikimedia-l mailing list, guidelines at: 
https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
New messages to: Wikimedia-l@lists.wikimedia.org
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, 

Re: [Wikimedia-l] IRC office hours: Shared hosting

[James Salsman]

Were there any objections to my request below?

Can we also please hire additional database, system, and if necessary
network administration support to make sure that the third party spam
prevention bot infrastructure is supported more robustly in the future?

On Monday, December 14, 2015, James Salsman  wrote:

> Hi Giles,
>
> I regret I will probably not be available for the IRC office hours as
> scheduled.
>
> In the discussion of shared hosting, I worry that en:User:Dispenser's
> reflinks project, which requires a 20 TB cache, is being forgotten
> again. He tried to host it himself, but it's offline again. This data
> is essential in maintaining an audit trail of references as long as
> the Internet Archive respects robots.txt retroactively, allowing those
> who inherit domains to censor them, even if they have already been
> used as a reference in Wikipedia. Keeping the cache is absolutely a
> fair use right in the US, in both statutory and case law, and it is
> essential to be able to track down patterns of attempts at deceptive
> editing to address quality concerns around deliberately biased editing
> such as paid editing. Because of the sensitivity of this goal, the
> Foundation should certainly bear the risk of hosting the reflinks
> cache. However, in the past, 20 TB was considered excessive, even
> though the cost was shown to be less than $5000 without whatever Dell
> NSA-enabled hardware you usually buy.
>
> Would you please reach out to en:User:Dispenser and offer them the
> 20TB hosting solution they need for the Foundation to bear the risk of
> the reflinks cache?  Thank you for your kind consideration.
>
> Best regards,
> Jim
>
___
Wikimedia-l mailing list, guidelines at: 
https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
New messages to: Wikimedia-l@lists.wikimedia.org
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, 

Re: [Wikimedia-l] Quality issues

[Andreas Kolbe]

On Sun, Dec 20, 2015 at 1:38 PM, Lydia Pintscher <
lydia.pintsc...@wikimedia.de> wrote:

> > At any rate, this is what I did. After I clicked "add reference", I got a
> > new field that came with a "property" drop down menu pre-populated with
> > "sex or gender", "date of birth", "given name", "occupation", "country of
> > citizenship", "GND identifier" and "image", none of which are remotely
> > relevant to entering a reference.
>
> Those should not have shown up for references and I am not aware of
> issues with that. Which statement was this specifically? The
> suggestions are not always perfect but at least the distinction
> between properties in the main part of the statement and its
> references should work very well.
>


Just try it, Lydia. Click "add" in subsidiaries in
https://www.wikidata.org/wiki/Q37156 -- enter a company name, and then
click "add reference". When I do that, the text field contains a greyed-out
"property", and the drop-down shows the unhelpful items I mentioned above.

And it would be good if the help text actually *asked* people to cite a
reference.



> > Use a licence that requires re-users to mention "Wikidata" on their
> sites,
> > ideally with a link to the Wikidata disclaimer, and you won't have to do
> > any education at all, and at the same time you'll have done a great thing
> > for transparency of data provenance on the internet.
> >
> > Moreover, you will have ensured that hundreds of millions of Internet
> users
> > are told where they can find Wikidata and edit it. Surely, if you
> actually
> > *want* to have human beings visiting and editing your wiki, that's in
> your
> > interest?
>
> I think we have to agree to disagree on the licensing part and what is
> best for Wikidata there. Yes I do want people to come to Wikidata but
> I do not want the license to be our forceful stick to achieve this. We
> have to work to build a project that people want to come to and
> contribute to. And we can do it as the number of editors for example
> shows.



Can you tell me just whose interests it serves if re-users do not have to
indicate that the data they're showing their users come from Wikidata? Max
Klein mused that the big search engines might be paying for Wikidata "to
remove a blemish on their perceived omniscience", because they can present
Wikidata content as though they had compiled it themselves.[1] That is at
least a plausible line of thought; but whom else does it serve?

It does not serve the end user, because they are left in the dark about the
provenance of the data. Moreover, they may not understand that these are
crowdsourced data, to which certain caveats always apply.

It does not serve Wikidata's interests, because many consumers of Wikidata
content who might otherwise come to edit the wiki, correct errors, refine
information and so on, will lack the bridge that would take them there.

We are a non-profit. The public good, the benefit to society, should be our
only concern.

So, who in society benefits, other than (arguably) the big commercial
search engines? Please explain.

Andreas

[1] http://hblog.org/
___
Wikimedia-l mailing list, guidelines at: 
https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
New messages to: Wikimedia-l@lists.wikimedia.org
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, 

Re: [Wikimedia-l] Quality issues

[geni]

On 20 December 2015 at 13:18, Andreas Kolbe  wrote:

> Lydia,
>
> I can only relate my impressions to you. The first two items I looked at
> (Jerusalem and Obama) happened to be protected, so on my first visit I was
> completely non-plussed as to how to edit anything on Wikidata.
>

Both are semied on en. I think this mostly shows your ignorance of
protection patterns. The first things you think of will pretty much always
be protected since they are the ones that attract a lot of vandalism. You
either use special:random or something closer to your personal interests.


-- 
geni
___
Wikimedia-l mailing list, guidelines at: 
https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
New messages to: Wikimedia-l@lists.wikimedia.org
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, 

Re: [Wikimedia-l] Mid-December Fundraiser Update

[Andreas Kolbe]

Thanks, Megan. It's been good to see some of these suggestions implemented.

Judging by the daily figures on https://frdata.wikimedia.org/ it looks like
you will end up with a fairly similar result to last year, nudging $30m.

Andreas

On Wed, Dec 16, 2015 at 10:42 AM, Megan Hernandez 
wrote:

> Hi everyone,
>
>
> The fundraising team is wrapping up the second week of the December
> campaign and we’d like to share an update with you--where we are, what
> we’ve changed, and what you can do to get involved.
>
>
> WHERE WE ARE:
>
> We’ve passed the halfway mark to the $25 million campaign goal.  So far,
> we’ve raised roughly $18 million (a preliminary total that is quickly
> changing and has not been reconciled with official totals from the finance
> department).   Banners are running on desktop and mobile devices. We are
> also sending emails to past donors asking if they would give again this
> year.  We’re monitoring the trends daily and look forward to sharing a
> post-campaign analysis with you.   We will post an update on when we will
> be able to end the campaign when we have a clearer picture.
>
>
> WHAT WE’VE CHANGED:
>
> Over the past several months, staff members and volunteers have provided
> both critical and generative feedback and new fundraising banner ideas.
> Their help has been very valuable. Many of these new messages have been
> tried in pre-campaign tests and in the last two weeks. Thank you to
> everyone who has shared their time and ideas!  A few message highlights:
>
>
>
>-
>
>We are no longer using the line "keep Wikipedia online and ad-free." It
>has been changed to “keep our work going another year.”
>-
>
>New text in the current banner: “We believe that knowledge is a
>foundation. It is a foundation for human potential, for freedom, for
>opportunity. We believe everyone should have access to knowledge—for
> free,
>without restriction, without limitation.”
>-
>
>"We survive on donations" has been changed to "We're sustained by
>donations"
>-
>
>"Please help us end the fundraiser and get back to improving Wikipedia"
>has been changed to "Please help us end the fundraiser and improve
>Wikipedia."
>-
>
>We have removed the persistent reminder from large banners. The reminder
>is still included in the small banners, which is consistent with the
> same
>style banners from the 2013 and 2014 campaigns.
>-
>
>The coffee cup image has been removed from banners
>
>
>
> We have also run some initial tests with new messaging that show
> encouraging results. We're still working on more messages and sorting out
> how we'll incorporate new ideas into the overall banner, but here are some
> sentences we’re testing:
>
>
>
>-
>
>“Wikipedia has become nothing short of a global public library.”
>-
>
>“We don’t run ads. We respect your privacy. We don’t sell your data.”
>-
>
>“Wikipedia exists to verify, protect, and share the combined knowledge
>of humanity.”
>-
>
>"There is nothing else on the internet like Wikipedia. We are a global
>information hub with 15 billion page views a month, written by a
> community
>of volunteers with a passion for knowledge."
>-
>
>"The information in Wikipedia is constantly growing, but we need your
>help to keep up with rapidly changing technology. Wikipedia is like a
>library or public park created by a passionate community."
>-
>
>"Wikipedia changes every second of every day. We just hit 5 million
>English articles, but we need to continually grow to serve our readers."
>-
>
>"Please become a volunteer or a donor to help end us the fundraiser and
>improve Wikipedia. Thank you."
>
>
> With just a couple weeks left to go in 2015, the team is working hard to
> make improvements. We’d love your help.
>
>
> HOW TO GET INVOLVED:
>
>-
>
>To file a bug report or technical issue, please create a phabricator
>ticket
><
> https://phabricator.wikimedia.org/maniphest/task/create/?template=118862>
>or email problemsdonat...@wikimedia.org
>-
>
>To get up-to-date on the the latest reader survey, read the full report
>on commons
><
> https://upload.wikimedia.org/wikipedia/commons/2/25/Wikimedia_Reader_Survey_November_2015.pdf
> >
>-
>
>To read the latest news from the team, see the fundraising meta page
>
>-
>
>To suggest another banner idea, visit the test ideas meta page
>
>-
>
>To learn more about the fundraising program and last year’s campaign,
>see the 2014-15 fundraising report
>
>
>
> Thank you to everyone for your support and to the fundraising team for
> working so collaboratively and such long hours throughout this campaign.
> And thank you to all those

Re: [Wikimedia-l] Quality issues

[Gnangarra]

there is a compromise license cc-by without the sa

On 20 December 2015 at 21:38, Lydia Pintscher 
wrote:

> On Sun, Dec 20, 2015 at 2:18 PM, Andreas Kolbe  wrote:
> > Lydia,
> >
> > I can only relate my impressions to you. The first two items I looked at
>
> Now we're getting somewhere ;-)
>
> > (Jerusalem and Obama) happened to be protected, so on my first visit I
> was
> > completely non-plussed as to how to edit anything on Wikidata. I never
> > noticed the lock icon (whereas I would have noticed, say, a coloured box
> at
> > the top of the page informing me that the item is locked). If I had been
> > just a random user, I would not have been back.
>
> Ok. I think we can make the icon more colorful for example to draw
> more attention to it. Mind you the icon is on-line with what you see
> on Wikipedia as well. That is why we have it.
>
> > Once I got over that one, I found the order in which statements are
> listed
> > completely confusing. I would have expected them to follow some logical
> > order, but it seems they are permanently *listed in the order in which
> they
> > were added to Wikidata*. So someone's date of birth can be the last
> > statement on a Wikidata page, or the first.
> >
> > Compare for example the location of the date of birth for Angela Merkel
> in
> > https://www.wikidata.org/wiki/Q567 to the location of Barack Obama's
> date
> > of birth in https://www.wikidata.org/wiki/Q76
> >
> > I tried to figure out a way to change the order, but couldn't find one.
> > Again, profoundly demotivating. Machines may not be bothered by this,
> > because they can instantly find what they are looking for, but people
> are.
> >
> > It might help to establish a default order for statements that makes
> > logical sense to a human being, and that people can become used to.
>
> Yes that is indeed one of the problems we have identified for quite
> some time already. It is high on the list for 2016. I hope we get to
> it in Q1.
>
> > As for actual editing, a few weeks ago, figuring out how to add an IBM
> > subsidiary to the IBM item, with a reference, must have taken me
> something
> > like half an hour. I read Wikidata:Introduction, learned about
> properties,
> > and then checked Help:Editing, which contained *nothing* about adding
> > properties. The word is not even mentioned.
>
> Ok so on-wiki documentation is not good enough. Point taken. It has
> been written by editors who are familiar with Wikidata. Giving
> feedback on the talk pages for those help pages would be valuable.
>
> > After clicking "add" in the *existing* subsidiaries statement for IBM
> item,
> > I saw a question mark icon with a "help text" that reads,
> >
> > ---o0o---
> >
> > Enter a value corresponding to the property named "subsidiaries". If the
> > property has no designated value or the actual value is not known, you
> may
> > choose an alternative to specifying a custom value by clicking the icon
> > next to the value input box.
> >
> > ---o0o---
> >
> > I didn't find this text helpful at all. It could have simply said, "Enter
> > the name of the subsidiary in the text box, and then add a reference."
>
> It is not as easy as that unfortunately. Potentially no item exists
> for that subsidiary and then you need to create one. Also the
> explanation for no-value and some-value in the text is important
> (though we need to improve the UI for them). But point taken we can
> improve this message.
>
> > At any rate, this is what I did. After I clicked "add reference", I got a
> > new field that came with a "property" drop down menu pre-populated with
> > "sex or gender", "date of birth", "given name", "occupation", "country of
> > citizenship", "GND identifier" and "image", none of which are remotely
> > relevant to entering a reference.
>
> Those should not have shown up for references and I am not aware of
> issues with that. Which statement was this specifically? The
> suggestions are not always perfect but at least the distinction
> between properties in the main part of the statement and its
> references should work very well.
>
> > The single property that would be most useful to list in that drop down
> > menu when people have said they want to add a reference is "reference
> URL".
> > But it's not included. If newbies don't know this property exists, how
> are
> > they supposed to discover it? Somehow I got there, but it was not
> enjoyable.
>
> As above this should have shown up.
>
> > These are indeed all user interface issues, and quite separate from the
> > other aspects we have been talking about. But they contribute to making
> > this wiki less attractive as a site that ordinary people might want to
> > contribute to manually, on a casual basis.
> >
> > Yes, if you are sufficiently motivated, you can figure things out. But as
> > things stand, I didn't find it inviting.
>
> Sure. As I said we still have quite some work to do and feedback such
> as the above is what will help us make it better.
>
> > On Sun, Dec 2

Re: [Wikimedia-l] Quality issues

[Lydia Pintscher]

On Sun, Dec 20, 2015 at 2:18 PM, Andreas Kolbe  wrote:
> Lydia,
>
> I can only relate my impressions to you. The first two items I looked at

Now we're getting somewhere ;-)

> (Jerusalem and Obama) happened to be protected, so on my first visit I was
> completely non-plussed as to how to edit anything on Wikidata. I never
> noticed the lock icon (whereas I would have noticed, say, a coloured box at
> the top of the page informing me that the item is locked). If I had been
> just a random user, I would not have been back.

Ok. I think we can make the icon more colorful for example to draw
more attention to it. Mind you the icon is on-line with what you see
on Wikipedia as well. That is why we have it.

> Once I got over that one, I found the order in which statements are listed
> completely confusing. I would have expected them to follow some logical
> order, but it seems they are permanently *listed in the order in which they
> were added to Wikidata*. So someone's date of birth can be the last
> statement on a Wikidata page, or the first.
>
> Compare for example the location of the date of birth for Angela Merkel in
> https://www.wikidata.org/wiki/Q567 to the location of Barack Obama's date
> of birth in https://www.wikidata.org/wiki/Q76
>
> I tried to figure out a way to change the order, but couldn't find one.
> Again, profoundly demotivating. Machines may not be bothered by this,
> because they can instantly find what they are looking for, but people are.
>
> It might help to establish a default order for statements that makes
> logical sense to a human being, and that people can become used to.

Yes that is indeed one of the problems we have identified for quite
some time already. It is high on the list for 2016. I hope we get to
it in Q1.

> As for actual editing, a few weeks ago, figuring out how to add an IBM
> subsidiary to the IBM item, with a reference, must have taken me something
> like half an hour. I read Wikidata:Introduction, learned about properties,
> and then checked Help:Editing, which contained *nothing* about adding
> properties. The word is not even mentioned.

Ok so on-wiki documentation is not good enough. Point taken. It has
been written by editors who are familiar with Wikidata. Giving
feedback on the talk pages for those help pages would be valuable.

> After clicking "add" in the *existing* subsidiaries statement for IBM item,
> I saw a question mark icon with a "help text" that reads,
>
> ---o0o---
>
> Enter a value corresponding to the property named "subsidiaries". If the
> property has no designated value or the actual value is not known, you may
> choose an alternative to specifying a custom value by clicking the icon
> next to the value input box.
>
> ---o0o---
>
> I didn't find this text helpful at all. It could have simply said, "Enter
> the name of the subsidiary in the text box, and then add a reference."

It is not as easy as that unfortunately. Potentially no item exists
for that subsidiary and then you need to create one. Also the
explanation for no-value and some-value in the text is important
(though we need to improve the UI for them). But point taken we can
improve this message.

> At any rate, this is what I did. After I clicked "add reference", I got a
> new field that came with a "property" drop down menu pre-populated with
> "sex or gender", "date of birth", "given name", "occupation", "country of
> citizenship", "GND identifier" and "image", none of which are remotely
> relevant to entering a reference.

Those should not have shown up for references and I am not aware of
issues with that. Which statement was this specifically? The
suggestions are not always perfect but at least the distinction
between properties in the main part of the statement and its
references should work very well.

> The single property that would be most useful to list in that drop down
> menu when people have said they want to add a reference is "reference URL".
> But it's not included. If newbies don't know this property exists, how are
> they supposed to discover it? Somehow I got there, but it was not enjoyable.

As above this should have shown up.

> These are indeed all user interface issues, and quite separate from the
> other aspects we have been talking about. But they contribute to making
> this wiki less attractive as a site that ordinary people might want to
> contribute to manually, on a casual basis.
>
> Yes, if you are sufficiently motivated, you can figure things out. But as
> things stand, I didn't find it inviting.

Sure. As I said we still have quite some work to do and feedback such
as the above is what will help us make it better.

> On Sun, Dec 20, 2015 at 11:25 AM, Lydia Pintscher <
> lydia.pintsc...@wikimedia.de> wrote:
>
>> That is why I am working with re-users of Wikidata's data on this.
>> They can link to Wikidata. They can build ways to let their users edit
>> in-place. inventaire and Histropedia are two projects that show the
>> start of this. As I w

Re: [Wikimedia-l] Quality issues

[Andreas Kolbe]

Lydia,

I can only relate my impressions to you. The first two items I looked at
(Jerusalem and Obama) happened to be protected, so on my first visit I was
completely non-plussed as to how to edit anything on Wikidata. I never
noticed the lock icon (whereas I would have noticed, say, a coloured box at
the top of the page informing me that the item is locked). If I had been
just a random user, I would not have been back.

Once I got over that one, I found the order in which statements are listed
completely confusing. I would have expected them to follow some logical
order, but it seems they are permanently *listed in the order in which they
were added to Wikidata*. So someone's date of birth can be the last
statement on a Wikidata page, or the first.

Compare for example the location of the date of birth for Angela Merkel in
https://www.wikidata.org/wiki/Q567 to the location of Barack Obama's date
of birth in https://www.wikidata.org/wiki/Q76

I tried to figure out a way to change the order, but couldn't find one.
Again, profoundly demotivating. Machines may not be bothered by this,
because they can instantly find what they are looking for, but people are.

It might help to establish a default order for statements that makes
logical sense to a human being, and that people can become used to.

As for actual editing, a few weeks ago, figuring out how to add an IBM
subsidiary to the IBM item, with a reference, must have taken me something
like half an hour. I read Wikidata:Introduction, learned about properties,
and then checked Help:Editing, which contained *nothing* about adding
properties. The word is not even mentioned.

After clicking "add" in the *existing* subsidiaries statement for IBM item,
I saw a question mark icon with a "help text" that reads,

---o0o---

Enter a value corresponding to the property named "subsidiaries". If the
property has no designated value or the actual value is not known, you may
choose an alternative to specifying a custom value by clicking the icon
next to the value input box.

---o0o---

I didn't find this text helpful at all. It could have simply said, "Enter
the name of the subsidiary in the text box, and then add a reference."

At any rate, this is what I did. After I clicked "add reference", I got a
new field that came with a "property" drop down menu pre-populated with
"sex or gender", "date of birth", "given name", "occupation", "country of
citizenship", "GND identifier" and "image", none of which are remotely
relevant to entering a reference.

The single property that would be most useful to list in that drop down
menu when people have said they want to add a reference is "reference URL".
But it's not included. If newbies don't know this property exists, how are
they supposed to discover it? Somehow I got there, but it was not enjoyable.

These are indeed all user interface issues, and quite separate from the
other aspects we have been talking about. But they contribute to making
this wiki less attractive as a site that ordinary people might want to
contribute to manually, on a casual basis.

Yes, if you are sufficiently motivated, you can figure things out. But as
things stand, I didn't find it inviting.


On Sun, Dec 20, 2015 at 11:25 AM, Lydia Pintscher <
lydia.pintsc...@wikimedia.de> wrote:

> That is why I am working with re-users of Wikidata's data on this.
> They can link to Wikidata. They can build ways to let their users edit
> in-place. inventaire and Histropedia are two projects that show the
> start of this. As I wrote in my Signpost piece it needs work and
> education that is ongoing.
>


Use a licence that requires re-users to mention "Wikidata" on their sites,
ideally with a link to the Wikidata disclaimer, and you won't have to do
any education at all, and at the same time you'll have done a great thing
for transparency of data provenance on the internet.

Moreover, you will have ensured that hundreds of millions of Internet users
are told where they can find Wikidata and edit it. Surely, if you actually
*want* to have human beings visiting and editing your wiki, that's in your
interest?

Andreas
___
Wikimedia-l mailing list, guidelines at: 
https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines
New messages to: Wikimedia-l@lists.wikimedia.org
Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/wikimedia-l, 

Re: [Wikimedia-l] Quality issues

[Andrea Zanni]

I second all Lydia's answers.
Also, I do think that there is a huge difference between usability/UX
issues and core, fundamental, systemic issues.
I personally think, Andreas, that you are displaying usability issues,
which are solvable (not easy, and not trivial, but at least can be fixed).

Regarding the CC0 vs CC-BY-SA problem, I don't think a single switch
between license would solve all the attribution problem: it hasn't solved
propagation of errors in the past with Wikipedia, I don't really get how it
could solve propagation of errors for Wikidata (we do know, though, that it
would bring a hell of issues for Wikidata itaself).

Aubrey

On Sun, Dec 20, 2015 at 12:25 PM, Lydia Pintscher <
lydia.pintsc...@wikimedia.de> wrote:

> On Fri, Dec 18, 2015 at 4:06 PM, Andreas Kolbe  wrote:
> > On Fri, Dec 18, 2015 at 11:28 AM, Andrea Zanni  >
> > wrote:
> >
> >> Andreas, you apparently did not read the following sentence:
> >> "Of course, the opposite is also true: it's a single point of openness,
> >> correction, information. "
> >>
> >
> > Andrea,
> >
> > I understand and appreciate your point, but I would like you to consider
> > that what you say may be less true of Wikidata than it is for other
> > Wikimedia wikis, for several reasons:
> >
> > Wikipedia, Wiktionary etc. are functionally open and correctable because
> > people by and large view their content on Wikipedia, Wiktionary etc.
> itself
> > (or in places where the provenance is clearly indicated, thanks to CC
> > BY-SA). The place where you read it is the same place where you can edit
> > it. There is an "Edit" tab, and it really *is* easy to change the
> content.
> > (It is certainly easy to correct a typo, which is how many of us
> started.)
>
> You are used to the edit tab being there. Someone recently said on
> Twitter this is the most displayed invisible link on the internet. All
> a matter of perspective and what we are used to ;-)
>
> > With Wikidata, this is different. Wikidata, as a semantic wiki, is
> designed
> > to be read by machines. These machines don't edit, they *propagate*.
> > Wikidata is not a site that end users--human beings--will browse and
> > consult the way people consult Wikipedia, Wiktionary, Commons, etc.
>
> Machines (with people behind them) _do_ edit Wikidata. Wikidata is
> designed to be read and written my both humans and machines. And it is
> used that way.
>
> > Wikidata is, or will be, of interest mostly to re-users--search engines
> and
> > other intermediaries who will use its machine-readable data as an input
> to
> > build and design their own content. And when they use Wikidata as an
> input,
> > they don't have to acknowledge the source.
> >
> > Allowing unattributed re-use may *seem* more open. But I contend that in
> > practice it makes Wikidata *less* open as a wiki: because when people
> don't
> > know where the information comes from, they are also unable to contribute
> > at source. The underlying Wikimedia project effectively becomes invisible
> > to them, a closed book.
> >
> > That is not good for a crowdsourced project from multiple points of view.
> >
> > Firstly, it impedes recruitment. Far fewer consumers of Wikidata
> > information will become Wikidata editors, because they will typically
> find
> > Wikidata content on other sites where Wikidata is not even mentioned.
>
> That is why I am working with re-users of Wikidata's data on this.
> They can link to Wikidata. They can build ways to let their users edit
> in-place. inventaire and Histropedia are two projects that show the
> start of this. As I wrote in my Signpost piece it needs work and
> education that is ongoing.
>
> > Secondly, it reduces transparency. Data provenance is important, as Mark
> > Graham and Heather Ford have pointed out.
> >
> > Thirdly, it fails to encourage appropriate vigilance in the consumer.
> (The
> > error propagation problems I've described in this thread all involved
> > unattributed re-use of Wikimedia content.)
> >
> > There are other reasons why Wikidata is less open, besides CC0 and the
> lack
> > of attribution.
> >
> > Wikidata is the least user-friendly Wikimedia wiki. The hurdle that
> > newbies--even experienced Wikimedians--have to overcome to contribute is
> an
> > order of magnitude higher than it is for other Wikimedia projects.
>
> Granted Wikidata isn't the most userfriendly at this point - which is
> why we are working on improvements in that area. Some of them have
> gone live just the other week. More will go live in January.
>
> > For a start, there is no Edit tab at the top of the page. When you go to
> > Barack Obama's entry in Wikidata[1] for example, the word "Edit" is not
> to
> > be found anywhere on the page. It does not look like a page you can edit
> > (and indeed, members of the public can't edit it).
>
> Now please go to any other page that is not protected. It has edit
> links plastered all over it. Editing there is much much more obvious
> than on Wikipedia.
> I really encour

Re: [Wikimedia-l] Quality issues

[Lydia Pintscher]

On Fri, Dec 18, 2015 at 4:06 PM, Andreas Kolbe  wrote:
> On Fri, Dec 18, 2015 at 11:28 AM, Andrea Zanni 
> wrote:
>
>> Andreas, you apparently did not read the following sentence:
>> "Of course, the opposite is also true: it's a single point of openness,
>> correction, information. "
>>
>
> Andrea,
>
> I understand and appreciate your point, but I would like you to consider
> that what you say may be less true of Wikidata than it is for other
> Wikimedia wikis, for several reasons:
>
> Wikipedia, Wiktionary etc. are functionally open and correctable because
> people by and large view their content on Wikipedia, Wiktionary etc. itself
> (or in places where the provenance is clearly indicated, thanks to CC
> BY-SA). The place where you read it is the same place where you can edit
> it. There is an "Edit" tab, and it really *is* easy to change the content.
> (It is certainly easy to correct a typo, which is how many of us started.)

You are used to the edit tab being there. Someone recently said on
Twitter this is the most displayed invisible link on the internet. All
a matter of perspective and what we are used to ;-)

> With Wikidata, this is different. Wikidata, as a semantic wiki, is designed
> to be read by machines. These machines don't edit, they *propagate*.
> Wikidata is not a site that end users--human beings--will browse and
> consult the way people consult Wikipedia, Wiktionary, Commons, etc.

Machines (with people behind them) _do_ edit Wikidata. Wikidata is
designed to be read and written my both humans and machines. And it is
used that way.

> Wikidata is, or will be, of interest mostly to re-users--search engines and
> other intermediaries who will use its machine-readable data as an input to
> build and design their own content. And when they use Wikidata as an input,
> they don't have to acknowledge the source.
>
> Allowing unattributed re-use may *seem* more open. But I contend that in
> practice it makes Wikidata *less* open as a wiki: because when people don't
> know where the information comes from, they are also unable to contribute
> at source. The underlying Wikimedia project effectively becomes invisible
> to them, a closed book.
>
> That is not good for a crowdsourced project from multiple points of view.
>
> Firstly, it impedes recruitment. Far fewer consumers of Wikidata
> information will become Wikidata editors, because they will typically find
> Wikidata content on other sites where Wikidata is not even mentioned.

That is why I am working with re-users of Wikidata's data on this.
They can link to Wikidata. They can build ways to let their users edit
in-place. inventaire and Histropedia are two projects that show the
start of this. As I wrote in my Signpost piece it needs work and
education that is ongoing.

> Secondly, it reduces transparency. Data provenance is important, as Mark
> Graham and Heather Ford have pointed out.
>
> Thirdly, it fails to encourage appropriate vigilance in the consumer. (The
> error propagation problems I've described in this thread all involved
> unattributed re-use of Wikimedia content.)
>
> There are other reasons why Wikidata is less open, besides CC0 and the lack
> of attribution.
>
> Wikidata is the least user-friendly Wikimedia wiki. The hurdle that
> newbies--even experienced Wikimedians--have to overcome to contribute is an
> order of magnitude higher than it is for other Wikimedia projects.

Granted Wikidata isn't the most userfriendly at this point - which is
why we are working on improvements in that area. Some of them have
gone live just the other week. More will go live in January.

> For a start, there is no Edit tab at the top of the page. When you go to
> Barack Obama's entry in Wikidata[1] for example, the word "Edit" is not to
> be found anywhere on the page. It does not look like a page you can edit
> (and indeed, members of the public can't edit it).

Now please go to any other page that is not protected. It has edit
links plastered all over it. Editing there is much much more obvious
than on Wikipedia.
I really encourage you to actually go and edit on Wikidata for longer
than 2 minutes.

> It took me a while to figure out that the item is protected (just like the
> Jerusalem item).

We have a lock icon in the top right corner to indicate protected
items like this.

> In other Wikimedia wikis that do have an "Edit" tab, that tab changes to
> "View source" if the page is protected, giving a visual indication of the
> page's status that people--Wikimedia insiders at least--can recognise.
>
> Unprotected Wikidata items do have "edit" and "add" links, but they are
> less prominent. (The "add" link for adding new properties is hidden away at
> the very bottom of the page.) And when you do click "edit" or "add", it is
> not obvious what you are supposed to do, the way it is in text-based wikis.

It is not a text-based wiki. So yes some things work differently. That
doesn't necessarily mean they are worse. I dispute your claim that the