Hacker News Reader

Comment by andrewla | Hacker News Reader

As an inclusionist who has basically given up on editing wikipedia because of deletionist forces, I read this article and see nothing but points I agree with.

What I would like to see, however, is a good defense of deletionism. I have a lot of trouble even understanding the point of view.

If there's some higher-level motivation about storage costs or something, then I can buy it, but I don't think anyone is making that argument. Does it hurt wikipedia to have low-quality articles? I think only insofar as the articles are about subjects that are important, so the effect seems to be self-balancing. Can wikipedia be used by people to glorify themselves with personal wikipedia pages? Sure, but who cares -- unless someone does care in a particular instance (i.e. a person named Albert Einstein tries to take over the main article -- people will fight it because they want the article to point to the right person).

View on HN

I think that the main argument that isn't simply a matter of storage costs (which probably does kick in at some point) is simply a matter of brand-management. If the average Wikipedia article is a discussion an incomplete and poorly sourced article about Pokemon then Wikipedia gets a reputation as poor source of information. It actually had the reputation pretty strongly early on, but it's reputation for reliabity has actually improved over the years.

I suppose that there is also the inverse problem where poor articles can use Wikipedia's relatively good reputation to give visibility to some pseudoscience or conspiracy theory. If it's really obscure, the only people who will be writing articles about are will be proponents who will then link to the article from their closed Facebook groups. It won't necessarily be obvious to non-experts what's going on so deleting certain kinds of articles could be a protection against that as well.

(I'm not any kind of Wikipedian btw; just speculating off of the top of my head.)

If that's the case, I'd prefer a two-tiered Wikipedia over a deletionist Wikipedia. In such a Wikipedia, each page could be clearly-labeled as either an encyclopedia-quality entry or as a "wannabe" encyclopedia entry. (A better term would be needed for the lower tier.) Such labeling should be very noticeable and visible when you open an entry, at least for the lower-tier entries.

They used to use the term "Stub" for brief pages that weren't full articles yet and they do put warnings on articles that don't meet certain quality standards, though I think that's different that what you're talking about here.

There's some overlap, though not full equivalency. Stubs presumably fall in line with Wikipedia's editorial vision, even if they haven't been fleshed out into full entries. The entries I'm referring to are mostly ones which would run afoul of Wikipedia's notability criterion.

Going back to the brand management issue, I'd be OK with deleted articles and their history being moved to another wiki with a different URL, with a name that is not "Wikipedia". Cookies or some other mechanism could be used to manage user/session preferences so that users who want to to be redirected to this other, questionable-page wiki can be redirected when they try to access pages that don't fall in line with Wikipedia's editorial vision. I'd be fine with having huge in-your face warnings about the likely lack of quality or veracity of such entries... just as long as I can access that content.

But this is contradictory. If you don’t care about credibility, why are you posting the content on wikipedia to begin with? Anyone can start a wiki, or a blog, or whatever.

What about additional information on encyclopedia quality pages that would normally have been excluded?

It seems like anyone could make an extension site to wikipedia where articles that aren't allowed at wikipedia could find a home, and the search there could also search wikipedia. The issue seems that nobody wants to put the hard work into making this happen, marketing it, and managing it. They just want wikipedia to do it for them, using the limited resources wikipedia has at their disposal, for free.

The argument for deletionism is straightforward: the scarcest resource on Wikipedia --- and this article agrees! --- is volunteer time. Every article on WP is a drag on volunteer time, because every article needs to be policed for spam and vandalism. Wikipedia is an encyclopedia, not a Hitchhiker's Guide, so scoping it to encyclopedic topics reserves volunteer time for meaningful tasks.

later

If you want to come at this from the vantage point of principles: one of the oldest rules on Wikipedia is "No Original Research" (an encyclopedia is a tertiary source). Articles on non-notable subjects are almost by definition original research.

This is one of the few responses I've seen that actually articulates a defense of deletionism rather than just a summary of deletionist theory, so let me engage here.

What exactly does policing for spam and vandalism consist of? There are a number of automated tools already deployed on wikipedia that get improved over time for detecting particular types of spam and vandalism. The use of anonymous accounts, in particular, make it hard to do any interesting spam at scale. It seems to me that the larger wikipedia is, paradoxically, the easier it will be to detect patterns of spam, especially out on the tail.

As for "No Original Research", that's probably the principle of wikipedia that I'm most skeptical of; and really, it isn't respected, as wikipedia has become such an important secondary source. I think it's useful as a guiding principle, in that you don't want people using wikipedia as a platform to advance their alternative theory of gravity, but for the most part the harm is small to allow us to push the envelope a little bit.

Policing spam and vandalism entails reading all the changes to pages you're policing. That's why WP editors have watchlists, because no one person can keep context for all the pages on the site in their head. A lot of spam detection is automated, but a lot of spam isn't automated: it's people with a fairly decent understanding of a topic surreptitiously editing its page to give it a preferred spin (or promotion). The whole premise of Wikipedia is that not being allowed to stand.

As regards "no OR isn't a problem": if your feeling is that the project is rife with OR, you should have no problem generating links to 5 pages that demonstrate the phenomenon. Let's discuss specifics.

I wonder if this measurement takes I to account the voulenteer time that could be gained by letting new members get a foothold.

> Does it hurt wikipedia to have low-quality articles? I think only insofar as the articles are about subjects that are important, so the effect seems to be self-balancing.

Here is where I disagree. I think Wikipedia is hurt by any low-quality articles. If I see some information on Wikipedia, there is a moderate positive signal that that information is correct. (Not as strong as a signal as I'd like, but oh well). Reducing the strength of that signal would make Wikipedia less useful.

And you dont see how that's an incredibly slippery slope, where the relative quality is rated by the worst article ala stack ranking?

I see two problems. One, as the qualifications loosen, Wikipedia will tend to have not just more articles, but more lesser quality articles. At a certain point, Wikipedia would become mostly poor quality.

Two, managing the articles has a cost for the editors involved. As the number of articles increases, their ability to do a good job will decrease.

Reply on news.ycombinator.com