May 11, 2019

Replicating the WIkinews project but how?


The Wikinews project is a well known example for a mediawiki based news website in the internet. After reading carefully through the help sections and take a look into the edits on the page most questions remains open. The idea itself seems attractive, but was it made in Wikinews the right way?
Suppose the project wasn't founded in the year 2004 or it was founded but we want to build our own wiki-based news hub, what is the best practice method in doing so? The technical aspect is easy to answer. The underlying wiki engine is available as open Source. The normal mediawiki works fine, but around 20 other wiki engines are available, the best one was programmed in C++ and is called Wiki++.
but suppose the normal mediawiki system which is based on php is used. The installed software will not work by it's own, what is needed else are some admins and some users who are doing actions in the wiki.
The advantage of creating a wiki news hub by it's own, that all the guidelines can be setup from scratch. But what are the best guidelines? Does the existing policy in the wikinews project sense? And if not, where is the bottleneck?
The main problem which has to be solved by all mediawiki installation is called spam protection. Spam protection means, that on the one side the wiki allows new users to create an account and make some edits, but at the same side the wiki-admin is able to monitor the system ban users. In the mediawiki system a large amount of spam-protection tools are delivered as default. They are from a technical perspective advanced, but they have a large bottleneck. Before the spam protection system can ban a user or revert all the edits from the user, the admin has to flag the edit first. Flagging means, to set a checkbox from false to true, then the edit is flagged. In a normal wiki for example Wikipedia the decision for flagging and edit is a easy one. The admin looks at the edit and then he decides to press the button. Colloquial, it is called the red deletion button, but what the admin is doing in reality is to simple flag an edit as spam and the deletion itself is made by the sql-backend with a cached delay.
The question if and why an edit is recognized as spam depends on the guidelines. Usually the admins have some rules and if not they will discuss which rules they need. In case of a content week the rules are supporting the need of the wiki. But what is the desired state in a news hub? In the original Wikinews project the help section provides some hints what Wikinews is and what not. These hints are not helpful, too many questions remains open.
Let us go a step back and ask what the purpose is of a news hub. A news hub is aggregating information. It is a playlist which tells the reader what the headline of the day is. Basically spoken a newshub is the same what CNN is doing. CNN is searching for the important news around the world and summarizes them into a 5 minute coverage. The input for the CNN show comes from all over the world and it contains more material than it can fit into a 5 minute newscast. So CNN has to leave out most of the content.
The same problem is there for a wiki based news hub. The amount of daily news in the world is huge. It is not possible to put this onto the website. The number of traffic accidents with deadly casualty alone is 3200 each day. That means there is an information overload even without mentioned vandalism on the website itself.
Suppose the users of a wiki news website are reporting correct information to the website, which are happen in reality and are backuped by other media. What the users will doing is to provide to much information. The world contains of 7 billion people and in the life of each of them every day is new. So there is a need to push this information to the news hub. If this happen with a real news wiki the website becomes down. Not because of a DDOS attack or by malicious user which can be banned with the checkuser tool of mediawiki, but because the user are operating inside the desired guidelines but producing too much news for the hub. Which kind of strategy can be used to select important information from less important information? Is a newly released pop song more important than a conflict escalation in the Big Brother UK episode (reality TV)?
Let us take a look into the guidelines of the wikinews project under the assumption that the submitted news story is not vandalism but a normal but less important information. The user logs into the system, creates a new article, puts the URL at the end and press the safe button. Then the article is put to the peer review workflow and gets analyzed by the next stage. Quote:
“All articles must go through independent Review prior to the Publish stage.”, https://en.wikinews.org/wiki/Wikinews:Reviewing_articles
It seems that the review workflow is the standard mechanism to control the information flow. If 100 articles are created in WIkinews and it's not plain vandalism, they will put all in the review queue. That means, after the user has pressed the safe button in the edit menu the article isn't published in the main page but is put into a queue. This sounds familiar, the same procedure is used by Facebook. The moderator of a group has to read through a post before it gets published (pre-publication peer review). The reason why this workflow was established is, because the amount of postings in social networks is extremely high and the postings are creating by a large amount of people in realtime. If all the information would get published without doing a peer review, the filter mechanism of the news hub is broken. That means, the user take over the social network and the owner of the project is out of control.
Can this happen with wiki based news hubs as well?
Let us go a step back and search for the peer review process in the normal wikipedia encyclopedia. The surprising fact is, that there is no such thing like pre-publication peer review. All the edits are submitted direct to the article. What the admins can do is to undo the action later, if the information is wrong. The normal case in Wikipedia is, that the user press “safe changes” and the edit is online.
In contrast the Wikinews project contains of a strict review workflow. All the articles published on the main page have gone through the filter first. Perhaps the idea is, that otherwise million of news posting each day would flood the system?
Let us take a side step to well known CNN like news agency. Do they need a pre-publication review system? Yes they do. A 5 minute live-broadcast is limited to exactly 300 seconds. If it is unclear what the speaker will say in this timeframe the format is broken. Before the show can go on air, the review department has to make a list of the news which should be read in the show. The list has to be shorter then 300 seconds. It is not possible to read out loud a 100 pages long book in 5 minute in front of a camera. To shorten the amount of information, the only option is to establish a review filter. Only information which has passed the filter can become part of the 5 minute show.