@@ -1790,3 +1790,22 @@ A lot of filters are disabled/deleted bc:
** "in addition to filter 148, let's see what we get - Cen" (https://en.wikipedia.org/wiki/Special:AbuseFilter/188) // this illustrates the point that edit filter managers do introduce stuff they feel like introducing just to see if it catches something
and 197: "amerikanisch -> US-amerikanisch ([[WP:RS\#Korrektoren]])"
"Korrektoren sind besonders gebeten, sich an die hier vereinbarten Regeln zu halten. In Fällen, in denen verschiedene Schreibweisen zulässig sind, werden Korrektoren um taktvolle Zurückhaltung gebeten: Es ist kein guter Stil, in einer schlüssig formulierten Passage eine zulässige in eine andere zulässige Schreibweise zu ändern." from \url{https://de.wikipedia.org/wiki/Wikipedia:Rechtschreibung#Korrektoren}
on software is political; the software that Wikipedia runs on is political; who writes it? what values do they embed in it? (cmp also Code)
"Wikipedia’s biggest problems have come when it’s strayed from this path, when it’s given some people official titles and specified tasks. Whenever that happens, real work slows down and squabbling speeds up. But it’s an easy mistake to make, so it gets made again and again.
Of course, that’s not the only reason this mistake is made, it’s just the most polite. The more frightening problem is that people love to get power and hate to give it up. Especially with a project as big and important as Wikipedia, with the constant swarm of praise and attention, it takes tremendous strength to turn down the opportunity to be its official X, to say instead “it’s a community project, I’m just another community member”."
The present thesis just lay the ground work for future edit filters research.
I gave an initial overview and summarised/showed/pointed out interesting paths/framework for future research (syn!).
The present thesis conducted an initial inquiry into an important quality control mechanism on Wikipedia previously unexplored by the scientific community—the edit filters.
The role of edit filters in Wikipedia's quality control ecosystem, the tasks the filters take care of, as well as some historical trends in filters' usage were studied.
It was further discussed why such an old-school rule-based technology is still actively used today when more advanced machine learning approaches exist.
Additionally, interesting paths for future research were suggested.
% TODO Refer back to title! Who is allowed to publish? Who decides?
%Outlook
According to the Wikipedian community people adding references to Brazilian aardvarks or <inser-another-hoax here> preferably shall not publish at all.
Edit filters are not the ideal mechanism to deal with this type of disruption:
they can warn editors adding information that their contribution does not contain any references (or outright disallow such contributions), but that was pretty much it. %TODO look into all filters tagges as "hoaxing"
However, what edit filters can do more effectively is prevent someone from moving XXX pages to titles containing ``ON WHEELS'', thus sparing users the need to track down and undo these changes, allowing them to use their time more productively by for example fact checking unverified edits and thus (syn) reducing the number of fake aardvarks and increasing the overall credibility of the project.
It is impressive how in under 20 years ``a bunch of nobodies created the world's greatest encyclopedia'' to quote Anrew Lih~\cite{}. %TODO verify how big is Wikipedia and whether it's the biggest collection ever.
Taking a step back,
according to the Wikipedian community people adding references to Brazilian aardvarks or <inser-another-hoax here> shall preferably not publish at all.
If we are to handle this type of disruption with edit filters, two approaches seem feasible:
Warn editors adding the information that their contribution does not contain any references, or outright disallow such edits
(which does not solve the problem of freely invented sources)
, but that was pretty much it. %TODO look into all filters tagges as "hoaxing"
Albeit edit filters may not be the ideal mechanism to deal with hoaxes, what they can do more effectively is prevent someone from moving XXX pages to titles containing ``ON WHEELS'', thus sparing vandal fighters the need to track down and undo these changes, allowing them to use their time more productively by for example fact checking unverified claims and hence reducing the number of fake aardvarks and increasing the overall credibility of the project.
%Outlook: centralisation, censorship
It is impressive how in under 20 years ``a bunch of nobodies created the world's greatest encyclopedia'' to quote Anrew Lih~\cite{Lih2009}.
This was possible, among other things, because there was one Wikipedia to which everybody contributed.
As the project and its needs for quality control grew, a lot of processes became more centralised. %TODO verify
As the project and its needs for quality control grew, a lot of processes became more centralised~\cite{HalGeiMorRied2013}.
It is, at the end, easier to maintain power and control in a centralised infrastructure.
However, centralisation facilitates not only the contribution of everyone (tm) towards a common goal—creating the world's biggest knowledge database, but also control.
It is not an accident that at the very introduction of the edit filters extension, critical voices expressed concern that a large scale censorship infrastructure is being installed. %TODO cite!
If there were multiple comparable projects, you had to censor all of them in order to silence people.
The debate is not trivial/petty/insignificant: whose knowledge is entered and who decides what is knowledge worth preserving.
It is more relevant than ever: the European Parliament basically voted the introduction of upload filters on the Internet just couple of months ago. %TODO give more details on Copyright directive
\begin{comment}
On centralisation
\cite{HalGeiMorRied2013}
"policy calcification and increasing centralization of policy"
"Wikipedia has changed from “the encyclopedia that anyone can edit” to “the encyclopedia that
anyone who understands the norms, socializes him or herself, dodges the impersonal wall of
semi-automated rejection and still wants to voluntarily contribute his or her time and energy can
edit”"
\end{comment}
%TODO also refer to power concentration in the hands of edit filter managers: small group, even smaller on catalan wikipedia (4 users) for instance
However, centralisation facilitates not only the contribution of everyone towards a common goal—creating the world's biggest knowledge database, but also control.
It is not an accident that at the very introduction of the AbuseFilter extension, critical voices expressed the concern that a really powerful secret tool was created to which very few people were to have access and thereby a large-scale censorship infrastructure was being installed~\cite{Wikipedia:EditFilterTalkArchive1}.
If there were multiple comparable projects, all of them had to be censored in order to silence people.
With Wikipedia being the first go-to source of information for a vast quantity of people all over the world today, the debate whose knowledge is included and who decides what is knowledge worth preserving is essential.
In the present moment, it is more relevant than ever:
The European Parliament basically voted the introduction of upload filters on the Internet just couple of months ago. %TODO give more details on Copyright directive
Since Wikipedia is distinctly relevant for the shaping of public opinion, despite its ``neutral point of view'' policy~\cite{Wikipedia:NeutralPointOfView} it is inherently political.
At the beginnings of this research, I heard the rumour that there was an edit filter on the German Wikipedia targeting gendering.
``To gender'' is a linguistic praxis whereby words referring to people are explicitely marked to designate more genders than the standardly used generic masculine.
It is a political praxis aiming to uncover under-represented groups and their experiences through the conscious use of language.
Even though no linguistic norm has established gendering to date, conscious decisions for or against the praxis are political, and so are technologies implementing these decisions.
As it turned out, no such filter existed on the German Wikipedia
\footnote{Although, as I have heard from women active in the German Wikipedia community, there is a strong general backlash against gendering. The community is also extremely men dominated.}.
This illustrates a point though:
Artefacts do have politics and as Lawrence Lessig puts it, it is up to us to decide what values we embed in the systems we create~\cite{Lessig2006}. %TODO Do Artefacts have politics?
Despite the famous neutral point of view and tralala, Wikipedia is the first go-to source of information for a vast quantity of people all over the world today.
As such it is distinctly relevant for the shaping of public opinion, and inherently political.
Artefacts do have politics and it is up to us to decide what values we embed in the systems we create. %TODO refine
% Refer to the rumour about gendering filter on German Wikipedia
\begin{comment}
\cite{Lessig2006}
"When we know what values we
want to preserve, we need only be creative about how to preserve them." (p.165)
%TODO put this somewhere, fun fact
This year the filters have a 10 year anniversary^^
Special attention: following edit filters from DE Wikipedia:
and 197: "amerikanisch -> US-amerikanisch ([[WP:RS\#Korrektoren]])"
"Korrektoren sind besonders gebeten, sich an die hier vereinbarten Regeln zu halten. In Fällen, in denen verschiedene Schreibweisen zulässig sind, werden Korrektoren um taktvolle Zurückhaltung gebeten: Es ist kein guter Stil, in einer schlüssig formulierten Passage eine zulässige in eine andere zulässige Schreibweise zu ändern." from \url{https://de.wikipedia.org/wiki/Wikipedia:Rechtschreibung#Korrektoren}
Both are log only filters;
and it's a political fight
%In a way, not taking a side is positioning in itself.
% Values, Lessig! --> check copyright blogpost
% think about what values we embed in what systems and how; --> Lessig (and also Do Artifacts Have Politics)
%"Most people try to help the project, not hurt it. If this were untrue, a project like Wikipedia would be doomed from the beginning. "
%(comes from assume good faith?)
% The merits and perils of centralisation: the project can get this big because there is *one* Wikipedia everyone can edit (and a lot of people do), but it also centralises power and the possibility to silence people --> censorship is possible/much easier in a centralised setting
\url{http://www.aaronsw.com/weblog/wikicodeislaw}
on software is political; the software that Wikipedia runs on is political; who writes it? what values do they embed in it? (cmp also Code)
"Wikipedia’s biggest problems have come when it’s strayed from this path, when it’s given some people official titles and specified tasks. Whenever that happens, real work slows down and squabbling speeds up. But it’s an easy mistake to make, so it gets made again and again.
Of course, that’s not the only reason this mistake is made, it’s just the most polite. The more frightening problem is that people love to get power and hate to give it up. Especially with a project as big and important as Wikipedia, with the constant swarm of praise and attention, it takes tremendous strength to turn down the opportunity to be its official X, to say instead “it’s a community project, I’m just another community member”."
\cite{GeiRib2010}
"We should not fall into the trap of speaking of bots and
assisted editing tools as constraining the moral agency of
editors"
* Think about: what's the computer science take on the field? How can we design a "better"/more efficient/more user friendly system? A system that reflects particular values (vgl Code 2.0, Chapter 3, p.34)?
Alternative approaches to community management:
compare with Surviving the Eternal September paper~\cite{KieMonHill2016}
"importance of strong
systems of norm enforcement made possible by leadership,
community engagement, and technology."
"emphasizing decentralized moderation" //all community members help enforce the norms
"ensuring enough leadership capacity is available
when an influx of newcomers is anticipated."
"Designers may
benefit by focusing on tools to let existing leaders bring others
on board and help them clearly communicate norms."
"designers should support an ecosystem of accessible and ap-
Interesting fact: there are edit filters that try to precisely identify the upload of media violating copyrights
%TODO refer to Lessig, Chapter 10 when making the upload filter commentary
% think about what values are embedded how in what systems (Lessig)
From talk archive:
"Automatic censorship won't work on a wiki. " // so, people already perceive this as censorship; user goes on to basically provide all the reasons why upload filters are bad idea (Interlanguage problems, no recognition of irony, impossibility to discuss controversial issues); they also have a problem with being blocked by a technology vs a real person
...
...
@@ -149,6 +103,7 @@ Freedom of speech concerns
" Do we think that automatons have the judgement to apply prior restraint to speech? Do we think they should be allowed to do so even if they can be imbued with excellent judgement? We don't allow the government to apply prior restrain to speech, why would we build robots to do it? Laziness?
