Le 01/12/2017 à 09:34, Markus Kroetzsch a écrit :
Well, I guess this is a dilemma that many sociologists and
anthropologists have to deal with. That's a really hard epistemic
problem you are raising here, and I don't think this list is the place
to discuss it extensively. So to make it short, I fully agree that your
concern is legitimate, but if your implied conclusion is that it would
be better to do nothing rather than going into a difficult epistemic
position, I don't share this conclusion. Also, to my mind belief in
absolute objectiveness is only delusion. I prefer to expose clearly what
I can myself identify as my starting point of view and let audience take
my biases into account rather than pretending that I aim presenting the
ultimate objective truth.
You are in an impossible position. Either you want to be an objective
researcher who tries to reconstruct past events as they happened, or
you are pursuing an agenda to criticise and change some aspects of
Wikidata. The way you do it, you are making yourself part of the
debate that you claim you want to reconstruct.
So I recognize I have a strong bias toward copyleft licenses as general
solution. But as I already stated in this thread, I am also for
promoting solutions with less legal constraints depending on the context
of production and fixed goals. And this nothing new, I surely might be
able to provide links or get some testimony that here and there I do
promote and myself use solutions with less legal constraints.
For this project, believe it or not, I had no pre-established agenda to
criticise and change Wikidata in a predetermined fashion as point of
departure. Of course before starting this project I had an opinion, and
yes CC0 for Wikidata didn't look appealing to me. But a strong
motivation behind this project was to give me a chance to change my mind
with a broader view of this choice of CC0 as unique license. Its origin,
its impact, and opinion of the Wikimedia community regarding this topic.
And I stay in this open minded dynamic.
Now while doing my research with this goal, I found strong hints of
potential conflict of interest, which was absolutely not what I was
looking for. Now strong hints and potential conflict of interest are not
proof of conflict of interest. If there was no such a thing, then it's
great and I'll document that in this way.
Finally note that while I'm taking part of the debate right now won't
change the fact that I didn't at the moment that the decision was took.
That is, I don't have the power to change the past, and I am aiming at
documenting past events on the topic using verifiable available sources.
I don't expect anyone to blindly trust me. Don't blindly trust me.
Everybody should really interested in the subject should check sources
on which claims are done and possibly draw a different conclusion and be
bold and make evolve the project or at least provide feedback.
From a research perspective, any material you gather in this way comes
with a big question mark. You are not doing us much of a favour
either, because by forcing us to refute accusations, you are placing
our memories of the past events in a doubtful, heavily biased context.
Well, I'm sorry for that. But it's not nothing new that our community is
full of freaks obsessed with transparency, "respect the license" and
"reference needed", is it? So how possibly it wasn't envisioned that one
day it would be embarrassing to not have a documented information about
how exactly was done this license choice and by who? My guess is that
the simple answer is that human make errors. I do errors. A lot of it.
Many reply in this thread surely can attest that, doesn't it? But may be
it would be good to recognize that you too can make errors, rather than
trying to put all the shame on me for asking information about such an
important topic so many time after the decision occurred.
Your overall approach of considering a theory to be true (or at least
equally likely to be true) unless you are given "proofs that this
claim is completely wrong" is not scientific.
Claiming that some approach is the one I'm following, discrediting this
approach and conclude that anything I say is then wrong is not fair either.
Contemporary scientific method mostly agree that you have to come with a
falsifiable theory, as exposed in by Thomas Kuhn in /The Structure of
Scientific Revolutions/. So this is a condition to have any chance to
have some scientific value. But of course this is not a guarantee that
the theory is true. At best it makes the theory not proven wrong by any
This is not how research works. For a start, Occam's Razor should make
you disregard overly complex theories for things that have much
simpler explanations (in our case: CC0 is a respected license chosen
by many other projects for good reasons, so it is entirely plausible
that the founders of Wikidata also just picked it for the usual
reasons, without any secret conspiracy).
Occam's Razor states that you should always prefer the theory which
requires the smallest set of entities/rules couple available to explain
a phenomena in regard of empirical data. It's completely different from
opting for the simplest explanation. The possibility of conflict of
interest require no hidden conspiracy, no additional entity, and simply
consider the possibility of occurrence of a phenomena which is widely
documented in social science fields.
Maybe at this point it might also be interesting to explicitly state
that knowing that there was no conflict of interest intervening in this
decision is interesting for the sake of governance transparency. But
going with this hypothesis don't really have much importance with the
rather independent question of whether using CC0 as unique license for
Wikidata is the best choice for reaching the goal of the Wikimedia
movement in a sustainable manner.
And once you have an interesting theory formed, you need to gather
evidence for or against it in a way that is not affected by the theory
(i.e., in particular, don't start calls for information with an
emotional discussion of whether or not you would personally like the
theory to turn out true).
I totally recognize that on this point I've misbehaved in this post, I
should have refrain of adding so much emotional emphaze in my message.
What you are doing here is completely unscientific and I hope that
your supervisor (?) will also point this out to you at some point.
Moreover, I am afraid that you cannot really get back to the position
of an objective observer from where you are now. Better leave this
research to others who are not in publicly documented disagreement
with the main historic witnesses.
This research don't have a supervisor. This is a Wikiversity research
project. Anyone can join and improve it.
So you should understand that I don't feel compelled to give you a
detailed account of every Wikidata-related discussion I had as if I
were on some trial here. As a "researcher", it is you who has to prove
your theories, not the rest of the world who has to disprove them. I
already told you that your main guesses as far as they concern things
I have witnessed are not true, and that's all from me for now.
The question is not whether you want to give me that kind of details. Me
and the feelings I might inspire doesn't matter here. The question is
whether you are willing to comply with the exigence of transparency that
the Wikimedia movement is attached to, on a topic which directly impact
its governance and future on a large scale.
On 01.12.2017 03:43, mathieu stumpf guntz wrote:
First rest assured that any feedback provided will be integrated in
the research project on the topic with proper references, including
this email. It might not come before beginning of next week however,
as I'm already more than fully booked until then. But once again it's
on a wiki, be bold.
Le 01/12/2017 à 01:18, Markus Krötzsch a écrit :
Your post demands my response since I was there when CC0 was first
chosen (i.e., in the April meeting). I won't discuss your other
claims here -- the discussions on the Wikidata list are already
doing this, and I agree with Lydia that no shouting is necessary here.
Nevertheless, I must at least testify to what John wrote in his
earlier message (quote included below this email for reference): it
was not Denny's decision to go for CC0, but the outcome of a
discussion among several people who had worked with open data for
some time before Wikidata was born. I have personally supported this
choice and still do. I have never received any money directly or
indirectly from Google, though -- full disclosure -- I got several
T-shirts for supervising in Summer of Code projects.
Maybe I wasn't clear enough on that too, but to my mind the problem
is not money but governance. Anyone with too much cash can throw it
wherever wanted, and if some fall into Wikimedia pocket, that's fine.
But the moment a decision that impact so deeply Wikimedia governance
and future happen, then maximum transparency must be present,
communication must be extensive, and taking into account community
feedback is extremely preferable. No one is perfect, myself included,
so its all the more important to listen to external feedback. I said
earlier that I found the knowledge engine was a good idea, but for
what I red it seems that transparency didn't reach expectation of the
So, I was wrong my inferences around Denny, good news. Of course I
would prefer to have other archived sources to confirm that. No
mistrust intended, I think most of us are accustomed to put claims in
perspective with sources and think critically.
For completeness, was this discussion online or – to bring bag the
earlier stated testimony – around a pizza? If possible, could you
provide a list of involved people? Did a single person took the final
decision, or was it a show of hands, or some consensus emerged from
discussion? Or maybe the community was consulted with a vote, and if
yes, where can I find the archive?
Also archives show that lawyers were consulted on the topic, could we
have a copy of their report?
At no time did Google or any other company take part in our
discussions in the zeroth hour of Wikidata. And why should they?
From what I can see on their web page, Google has no problem with
all kinds of different license terms in the data they display.
Because they are more and more moving to a business model of
providing themselves what people are looking for to keep users in
their sphere of tracking and influence, probably with the sole idea
of generating more revenue I guess.
Also, I can tell you that we would have reacted in a very allergic
way to such attempts, so if any company had approached us, this
would quite likely have backfired. But, believe it or not, when we
started it was all but clear that this would become a relevant
project at all, and no major company even cared to lobby us. It was
still mostly a few hackers getting together in varying locations in
Berlin. There was a lot of fun, optimism, and excitement in this
early phase of Wikidata (well, I guess we are still in this phase).
Please situate that in time so we can place that in a timeline. In
March 2012 Wikimedia DE announced the initial funding of 1.3 million
Euros by Google, Paul Allen's Institute for Artificial Intelligence
and Gordon and Betty Moore Foundation.
But that's all the problem here, no one should have to carry the pain
of trying to reconstruct what happened through such a research.
Process of this kind of decision should have been documented and
should be easily be found in archives. If you have suggestion in
methods, please provide them. Just denigrating the work don't help in
any way to improve it. If there are additional sources that I missed,
please provide them. If there are methodologies that would help
improve the work, references are welcome.
So please do not start emails with made-up stories around past
events that you have not even been close to (calling something
"research" is no substitute for methodology and rigour).
Putting unsourced personal attacks against community members before
all other arguments is a reckless way of maximising effect, and such
rhetoric can damage our movement beyond this thread or topic.
All this is built on references. If the analyze is wrong, for example
because it missed crucial undocumented information this must be
corrected with additional sources. Wikidata team, as far as I can
tell, was perfectly aware of this project for weeks. So if there was
some sources that the team considered that it merited my attention to
complete my thoughts on the topic, there was plenty of time to
provide them before I posted this message.
Our main strength is not our content but our community, and I am
glad to see that many have already responded to you in such a
measured and polite way.
We completely agree on that. This is a wonderful community. And
that's concerns for future of this very community which fueled this
I only can reiterate all apologies to anyone that might have felt
personally attacked. I can go back to reformulate my message.
I hope you will help me to improve the research, or call it as you
like, with more relevant feedback and references.
On 30.11.2017 09:55, John Erling Blad wrote:
> Licensing was discussed in the start of the project, as in start of
> developing code for the project, and as I recall it the arguments for
> CC0 was valid and sound. That was long before Danny started
> As I recall it was mention during first week of the project (first
> of april), and the duscussion reemerged during first week of
> development. That must have been week 4 or 5 (first week of may),
> delivery of the laptoppen was delayed. I was against CC0 as I
> problems with reuse og external data. The arguments for CC0
> And yes, Denny argued for CC0 AS did Daniel and I believe Jeroen and
> Jens did too.
Wikimedia-l mailing list, guidelines at:
New messages to: Wikimediaemail@example.com