@@ -149,14 +149,20 @@ Claudia: * A focus on the Good faith policies/guidelines is a historical develop
\section{Limitations}
This work presents a first attempt at analysing Wikipedia's edit filter system.
It has several limitations (we could think of).
First, it focuses on English Wikipedia only.
We are convinced that there are valuable lessons to be learnt (about the communities, usefulness of filters, ..) from comparing edit filter use across different language versions.
Second, unfortunately, including an ethnographic analysis was not possible.
Several limitations of this study come to mind.
Firstly, it focuses on English Wikipedia only.
This presents (syn!) an excellent starting point for analysis of the edit filter system, since this was also the first language version to which the mechanism was introduced.
However, valuable lessons can be learnt (about the communities, models of governance, usefulness of filters, etc.) from comparing edit filter use across different language versions.
Just recall, how for instance the role of edit filter managers doesn't exist in certain language versions (comapare chapter~\ref{}) and instead it is administrators who have an \emph{abusefilter-modify} permission next to their other rights.
Secondly, unfortunately, including an ethnographic analysis was not possible.
It would have been particularly insightful to talk to edit filter managers (above all such who are simultaneously also bot operators) and developers of the extension, as well as regular editors who have tripped a filter.
This is partially due to the fact that we employ a computer science perspective on the question and partially due to limited time.
Third, the manual filter classification was undertaken by one person only, so biases of this person have certainly shaped the labels.
Fourth, edit filter history table was not available, no hollistic quantitative analysis of the filters' development over time
Thirdly, the manual filter classification was undertaken by one person only, so biases of this person have certainly shaped the labels.
Fourth, edit filter history table was not available, so no hollistic quantitative analysis of the filters' development over time was possible.
Fifth, no access to the details of hidden filters, so no insights into the areas they target (although couple of educated guesses: bunch of persistent long term vandal, who often employ sockpuppets; harassment/personal attack cases hidden to protect the affected persons)
