Slashdot Mirror
Tool Shows the Arguments Behind Wikipedia Entries
Al writes "A team of researchers at the Palo Alto Research Center have created a tool that shows how much argument has gone into crafting an entry. Ed Chi, a senior research scientist for augmented social cognition at PARC, obtained access to Wikipedia edit data and used it to build a tool that shows whether users have fought over the accuracy of a page by rapidly re-editing each other's changes. Experiments suggest that the method provides a better measure of 'controversy' than simply having Wikipedia editors add a warning to a suspect page. Their software, called Wikidashboard, serves up a Wikipedia entry, but adds an info-graphic revealing who has been editing it and how often it has been reedited. Of course, this doesn't reveal whether a Wikipedia entry is truly accurate, but it might at least highlight an underlying bias or vested interest."
If this technology could be applied to counting and characterizing forum yelling, we could measure how little we really have to say in so many words on other internet venues as well.
http://www.aaronrogier.net
For a minute I thought it was a dupe of this story but it's not (different team, different school, and slightly different goal).
It'd be interesting to compare the two...
--MarkusQ
Articles that I edit on wikipedia get flagged as being arguments because I usually edit them from both my home and work computers. as a reult when I am in a mood to edit there is rapid fire changes from multiple IP addresses. I see warnings when I log in that it looks like I'm in a dispute and I may be banned if further revisions occur.
Some drink at the fountain of knowledge. Others just gargle.
I had a similar idea, but instead colour the text directly by how long (in terms of edit survival) a piece of text has been around (with a little filter to ignore spelling fixes).
The more recently added text can be made lighter, whereas "more reliable" text can be shaded darker.
Also, with more recently added text, if it replaced something that was there for a while, then add a little mark or something that when you mouse over it the old text is shown (or use the alt-text).
I just don't have the time to build it.
I like tools like this much more than I like the various conflicts over what wikipedia's one true behavior ought to be. With so many people, and so many disparate objectives, you cannot have one wikipedia to satisfy them all. However, since the wiki preserves revision and comment data, and it is all available under a liberal licence, it is possible for parties both inside and outside wikipedia to build view into the wiki that are closer to their desired vision, rather than struggling endlessly over what the wiki will look like.
One could, for instance, easily include or exclude comments and revisions based on attributes of the accounts that made them, produce "frozen" versions of pages believed to have gotten to a stable point, treat different pages differently based on input from a tool like the one in TFA, and so on. This is, obviously, more difficult than just using the default; but it seems a shame to treat wikipedia as just a strategy to get a static encyclopedia, when you could take advantage of all the other data that it preserves.
the text should be
but adds an info-graphic revealing who has been
instead of
but adds an info-graphic revealing how has been
Very nice job on the summary! It explained exactly what the topic was about and summarised the key findings and where to go for more information. :)
No actual need to visit the article
As to the tool, the display looks too complex to provide a simple guide as to edit wars/controversy. Presumably more read bars is bad, or is it? That's really just a slightly graphical form of the edit history itself, when whats needed is a simpler, thermometer style presentation.
For every expert, there is an equal and opposite expert. - Arthur C. Clarke
As long as the edits aren't being made by a government or corporate entity and/or their minions, it should be valid.
Who else thought we were talking about THE BAND Tool?
Perl, n. A language spoken by Eskimos.
Lobbying groups for powerful business interests who know they are doing something ethically parasitic to society have a stanard MO of "manufacturing controversy" through thinly disguised think tanks or publications pushing an agenda.
Example: Smoking doesn't actually cause cancer.
Tracking the number of edits only shows whether an interest group is actively trying to revise reality. It does not say which side it is or whether the "controversy" is genuine.
In other words, it's no more dependable than the signs they slap across half of wikipedia because powerful groups of outright looneys shry about it.
VLC FOR MAC IS DYING! IF YOU DEVELOP, PLEASE SAVE IT!!
...and the slashdot effect strikes again...
Intelligent idiots are we. | Evil men do not understand justice.
Lest anyone thing you need to be a well-connected researcher to "obtain access to Wikipedia edit data", it's actually all public. Although you will need 100GB+ of hard drive space, and some well thought out algorithms, to parse the full-history dumps that contain every revision of every page.
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
"Controversy" doesn't even begin to let you know whether something's believable. Various well-funded oil industry whores have been questioning Global Warming and just about anything else that doesn't fit a "Burn More Oil" world view for decades. I think they took lessons from the tobacco industry.
If there's a vested interest, they'll certainly have an army of well-paid, lying pricks standing ready to bury the work of honest researchers under a mountain of bullshit. Wikipedia needs to persuade some people with unimpeachable credentials to evaluate the really questionable entries, and mark the good ones appropriately.
I've calculated my velocity with such exquisite precision that I have no idea where I am.
Unless they have added a new feature, the wikidashboard is old news - as evinced by this Wikipedia signpost article from 2007: http://en.wikipedia.org/wiki/Wikipedia:Wikipedia_Signpost/2007-09-17/In_the_news
The deletionist/inclusionist argument is almost exclusively waged over really, really recent stuff, and most of that related to pop culture. If you're writing about 19th-century history, you have to really try to encounter a deletionist.
It's basically an ongoing process of trying to find a good balance between erroneously/unnecessarily excluding recent and pop-culture stuff that is actually useful in an encyclopedia, and allowing Wikipedia to be used as an advertising platform by everyone with a company, book, academic CV, or piece of software to promote.
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
I wrote a hackish Greasemonkey script to do something similar last year that just did it on the fly. Also tried to get a simple measure of how much was spam just by searching for simple keywords in the edit records.
I'd like to revisit the idea some time, because I'm sure you can also capture more details about the nature of editing (ie, how often are different parts edited, what keywords get added and removed most frequently, how many people are involved or is it just a revert war between two stubborn editors...) within a useful visualisation.
Instead of just showing such stats for the article in total, this tool would become much more powerful by showing statistics for single sections or paragraphs. How hotly was a particular phrase contested?
Especially if it can also show when a particular sentence was introduced, or what was deleted from the article over time.
I spend a lot of my time on Wikipedia restoring lost text, when vandalism was incompletely fixed and then forgotten about. Occasionally I've found entire paragraphs that were deleted more than a year ago. It's pretty annoying to have to explore the history in a kind of binary search - is the problem older than this? Is it newer than this? Older than this revision? Ah, got it.
You might want to check your Internet connection (for average speed as well as variability), as it's fine here.
I've had a few articles deletion-nominated that ended up being kept, which I'd consider not particularly close to being worthy of deletion (and they were even cited!), so the gray area does extend a bit.
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
Cabal-o-meter. That's a better name for the tool.
Most edit wars are caused by cooks who have a specific pet peeve. And when a lot of cooks have the same loony idea (be it creationism, homeopathy, smoking is healthy, pictures of Mohammed shouldn't be allowed, nudity is sin, ...) it looks to this tool as if there is a lot of controversy.
But there isn't. There may be a lot of people with cooky ideas, but that doesn't show that there is genuine controversy, it only shows that people sometimes think things that are bizarre and not evidence supported.
That is not to say that real controversy doesn't exist (where did the amphibians come from, precisely?) but this tool just manufactures controversy where it doesn't really exist.
That guy is a total tool
And it seems "AntiVandalBot" is the most controversial user. Oh wait...
Seriously, in years of casually editing Wikipedia on and off, I've never seen an edit war, but have helped revert vandalism often (in fact, just a moment ago on one of the pages I tested this tool with). Many edits happen on those pages daily.
I've long thought the most useful page isn't the most recent, but the most durable...
The black monolith at the centre of the Wikipedia lie is simply the following:
1) An article on any subject on which there's been written more than 50 newspaper articles to date can be written in any style and imparting any feeling of credence to the reader, from the most positive to the most negative.
2) The rhetorics behind this are highly subjective and debatable and very poorly covered (or even coverAble) in strict written guidelines.
It's simply the case that words have connotations, multiple meanings and implicit assumptions that can lead to a _factually correct_ presentation, but one that imparts a feeling of negativity or disrespect.
A negatively angled article might for example use several instances of the wording "X has accused (..)", followed by a brief description of the nature of the complaint, while a positively angled one might have the wording "X has published a scathing criticism of (..)" followed by a logical chain of X's main arguments.
Why is it that someone making many accusations looks bad, while someone publishing scathing criticisms look good, or at least noteworthy? Beats me.
Selective style application is a different area. For example, if a politician has been described by a newspaper columnist as "disingenious and dishonest", which of the following wordings does the Wikipedia guidelines DICTATE shall ALWAYS be used?
1. "X has been described as disingenious and dishonest[17]".
2. "X has been described by columnist Y at newspaper Z as disingenious and dishonest[17]".
3. Newspaper columnists are not scholars and hence not sourceworthy.
Something tells me that if Barack Obama's article is being edited, approach 1 would be rejected and the editor possibly quarantined (does anyone want to take the challenge?) while approach 3 would most likely be taken. If Jack Thompson's article was edited, wording 1 would very likely be taken.
The initial meme regarding Wikipedia was that this would not be a problem due to the following mechanism: 1. Everyone would be able to contribute, hence 2. The tone of articles would represent
Both of these have over time been subverted, respectively:
1. As Wikipedia articles become more and more developed, "feature creep" such as tables, images, reference styles and others makes editing the code very difficult for casual users.
Simon taught him to curse his parents and the aristocracy and also to blaspheme. He also made Louis-Charles sleep with prostitutes, from whom he caught [[venereal diseases]].{ref name="Dauphin"}{{ro icon}} [http://convorbiri-literare.dntis.ro/PLATONmai4.html "The Reeducation of The Dauphin,"] by Mircea Platon, [http://ro.wikipedia.org/wiki/Convorbiri_Literare Convorbiri Literare], May 2004{/ref} The young Louis XVII was repeatedly threatened with the [[guillotine]], which caused him to faint{ref}[http://books.google.com/books?id=5k5qQ2w_UqAC&pg=PA115&dq=%22louis+charles%22+prostitutes+king&sig=kxfKSLptsOwOd2sTim3wJZXg7L0#PPA115,M1 "The Lost King of France: How DNA Solved the Mystery of the Murdered Son,"] by Deborah Cadbury, St. Martin's Press: 2003, p. 115{/ref}{ref name="Dauphin"/}. He had been told that he had fallen from favor with his parents, who still lived but no longer wanted him.
At this point, creating an article is no less easy than writing HTML, meaning that article polishing, finishing and approval is restricted to the few individuals who know how to edit.
More importantly however is the rise of the Moderator, and the social hierarchy that has arisen to support it. The extensive 'profile pages', with emblems for various sports, martial arts, educational achievements etc. is an interesting artifact, considering that people are effectively publishing blogs on what was initially seen as an editable-by-all encyclopedia.
Having power within the hierarchy gives the opportunity to decide the aforementioned edits of tone. A strong dedication to and strong feelings about a subject are no hindrance - quite the opposite! I've seen th
Maybe come back once you remove the spyware and viruses from your PC...
What part if *GUFFAW!* wasn't clear?
Chas - The one, the only.
THANK GOD!!!