Hacker News new | past | comments | ask | show | jobs | submit login
Introducing Infinithree: Against Wikipedia Deletionism (infinithree.org)
68 points by wslh on May 29, 2011 | hide | past | favorite | 38 comments



Hopefully this takes off.

After years of complaints and issues with rampant deletionism and abusive editors, Wikimedia has finally started to collect statistics on how this has negatively effected the quality, depth and breadth of the project.

I can't find the Wikimedia Foundation report on this (it coincided with the great programming deletion incident), but it was almost universally negative. Some of the more recent data collection tells a similar story, it shows that new users, while attempting to make good faith contributions, are quickly drummed out of the place by established wikipedians via negative comments and rampant deletionism.

Despite this, and years of evidence that this is a serious problem (to the point of being blindingly obvious) with the project, the foundation hasn't taken any particular move to rectify it. I fear they will never get around to doing anything, or the response will be tepid and pointless.

With bits being basically free these days, there's just no real reason to limit Wikipedia to just being a duplicate of a Paper-and-Ink encyclopedia.


> With bits being basically free these days

The work required to maintain quality is not free though. The larger Wikipedia becomes, the more resources are needed to ensure a high quality.

There is already a system completely free of "deletionism" with limitless scope which anyone can publish anything on. Its called the web. Why do people want their favorite topic published on Wikipedia rather than on their own page? Because they want the additional authority which comes from being on Wikipedia. But this authority comes from quality control.

I think some of the complaints against deletionism could be resolved with independent domain-specific wikipedias with different rules. For example a programming language-pedia could allow original content written by the language designers, something that is not appropriate for wikipedia.


Why do people want their favorite topic published on Wikipedia rather than on their own page? Because they want the additional authority which comes from being on Wikipedia. But this authority comes from quality control.

I actually find wikipedia's aggregation and summarization of information just as, if not more, valuable than the authority angle.

There actually are some quite interesting domain localized wikis around, on top of the random wiki for a game or fan wiki (or encyclopedia dramatica or conservapedia for that matter)

http://www.wikia.com/Wikia

http://www.wikispaces.com/


I see the most promising line as actually being something considerably different from Wikipedia, rather than an opposition to it, which despite the top-line rhetoric does seem to be what the proposal is hinting at.

Wikipedia aims to be a summary of existing sources (books, journal articles, newspapers, etc.), with citation to where it got its material, but originating nothing; basically, the ultimate tertiary source. I think that's a useful thing to have, but it's not the only possible thing you could imagine building.

"True and useful" is a rather different set of criteria, and there is a whole space of websites that could pursue that, ranging from some that already exist (music and literary fansites, some academic wikis, knowyourmeme.com, Everything2) to many that don't yet.


For one thing, it implies the ability to work on primary research right there in the site. It would need sufficient justification to be accepted as "true" and it would need a "use" to somebody even if only in the same field, but that covers basically all well conducted academic research.


My longstanding beef with deletionism is that it goes against the Wiki Way.

If Ward's wiki had operated in the same way, it wouldn't have been anywhere near as influential and useful.


But Ward's wiki is topic-focused, which is itself a good layer of defense. It's full of important (to us) but abstruse subjects like Liskov-substitutability, which simply aren't going to hold the interest of the people who want to maintain exhaustive lists of Pokémon characters.


Hmm, that would suggest a decentralized network of domain-specific wikis would be a more robust model than Wikipedia's centralized one.

Back in the day, Lion Kimbro was heavily advocating for something like that using the notion of WikiNodes to interconnect the wikis: http://meatballwiki.org/wiki/action=browse&id=WikiNode&#...


As a "deletionist" (at least in theory if not in practice) I welcome this project.

I'm glad that somebody is finally doing this, instead of just sitting around complaining that wikipedia isn't what they want it to be. Complaining that wikipedia isn't an indiscriminate collection of information is like complaining that Target isn't a fine lobster restaurant -- it was never intended to be and it's fine the way it is. There's enough "inclusionists" out there that I'm sure they can build a parallel encyclopaedia with a slightly different philosophy, and if (as I suspect is true) it turns out to be swamped with useless and unverifiable information (not to mention a huge number of vanity pages) then at least we'll know.

However:

[it] will discard ‘notability’ and other ‘encyclopedic’ standards in favor of ‘true and useful’

I'm not sure how "useful" differs from "notable" and I suspect they're just going to wind up having the same arguments about different words. For instance, is a ten thousand word biography of my cat both true and useful? It's certainly true, and it might be useful to someone at some point, so why delete it?


"Complaining that wikipedia isn't an indiscriminate collection of information"

Nobody's making that complaint, so that's a classic straw man. The complaint is the legitimate information is being kept off the project by people who don't think it's important enough for wikipedia. I recently argued against the proposed deletion of an article on a charter school at its AfD. Despite the school being mentioned in many articles, people were arguing that elementary schools simply aren't worthy of mention on Wikipedia. It is this attitude that we oppose, not that "that wikipedia isn't an indiscriminate collection of information".


Nobody's making that complaint, so that's a classic straw man.

Well, I've met at least one person who advocated the idea that wikipedia should contain any information that was true. When I asked him if I then should be able to create a many-terabyte article entitled "List of random numbers that hugh3's computer generated at [such and such a date and time]" then he said it should. So there are certainly some folks taking up that extreme position.

Otherwise you're just arguing at the margin. Maybe the notability criteria for schools are too strict. Maybe they're not. I can't think of anything about which I give less of a damn. (For what it's worth I checked, and my high school is certainly listed while one of my two primary schools is and the other [no less intrinsically notable] redirects to the particular suburb in which it is located].)


The person you cite is in no way representative of inclusionists. The analogous position would be to say that deletionists believe Wikipedia should only have articles on subjects that Britannia has articles on. You might be able to find a person who holds such a view, and it might be right to call them a deletionist. It would be disingenuous, however, to say that deletionists hold those views.


The sad part is how often the type of pattern you identified is used, especially on the 24 hour "news" networks by pundits of all sorts. Take any group and find the extreme or hangers on then use those people to define the group as a whole.


On notability, if wikipedia only contains things that are generally known, what's the point of putting that information in a reference resource? The point of a resource like WP is so people can find things that they may not know.

Absurd examples like people's cats and random numbers are not really the problem and I suspect you know that. The problem is not the notability guidelines per se. It's the abusive use of the guidelines (evoked as gospel) by deletionists to keep n00bs out of their playpen.


On notability, if wikipedia only contains things that are generally known, what's the point of putting that information in a reference resource? The point of a resource like WP is so people can find things that they may not know.

Now there's a strawman. All the facts on wikipedia are widely known to some group of people, but nobody knows 'em all. Though actually, you've just given me another excuse to link to my new blog in which I'm (perhaps slightly facetiously) attempting to become an expert on everything by reading every one of wikipedia's 3.5 million articles: http://projectomniscience.blogspot.com/ ... I only started it two days ago but I've already done both aardvarks and zebras.

Absurd examples like people's cats and random numbers are not really the problem and I suspect you know that.

They would be, if you removed the notability guidelines. Well, the cats would be.

It's the abusive use of the guidelines (evoked as gospel) by deletionists to keep n00bs out of their playpen.

Possibly. I've never actually tried editing wikipedia, only reading it. On the other hand, I have been through Articles for Deletion a few times and the vast majority of stuff that gets deleted clearly is crap -- vanity articles and the like.


I don't think removing notability guidelines is the right call either. I think treating the guidelines as guidelines is the right, simple and obvious call.

Unfortunately, notability guidelines are treated like gospel by the deletionists. In cases where stuff is wrongfully deleted, the entire discussion will center around notability, usually with source material references being dismissed out of hand by an overzealous admin.

I'm sure there should be some kind of penalty of some sort or another for editor types who habitually flag AfD and then have them overturned, and that would remedy some of it. But to be honest, speculative discussions about how the foundation could and should deal with the issue have gone nowhere for the better part of half a decade.

Possibly. I've never actually tried editing wikipedia, only reading it.

You really should sometime. I've had things deleted so fast that I didn't even have time to save my edit, then move down to the references section to paste in my references, I swear there are just bots now that monitor new edits and immediately revoke any changes.

It's a pity really.

Good luck reading all the articles! I usually get lost of an extended tour of interesting links whenever I go to the WP.


Not bots, but users of Huggle and the like. With a thousand or so "vandal patrollers" getting fed the recent changes queue in real time, most edits get pretty heavily scrutinized. Unfortunately, many vandal fighters get overzealous and overstep their bounds.


It's at the point where directly editing anything is a waste of time for new users. At most I'll put a comment on a talk page and hope the cabal likes a proposed edit and does it themself.


Here's just one example. There are thousands of companies and people on Crunchbase that are not notable enough for Wikipedia. Having that information on Crunchbase is extremely useful.


Quite ambitious. I like it.

This post is a few months old, though, and no public developments can be seen. I hope it goes on.


Isn't Google the answer to deletionism?

Why should and and all content be stored in a special place called "wikipedia" or some other name, when you can write anything you want about any subject on your blog and have it indexed by a central search engine?

Wikipedia offers two things that lone blog posts don't: collaborative editing and "findability" (SEO).

Findability should be the problem of search engines; the fact that we need to go to Wikipedia (or IMDB...) to find specific information is a testament to Google's imperfection.

If you want to publish "a 10,000 words biography of [your] cat", you should do so on your blog, and that should be easy to find when searching for "hugh3's cat" (but not when searching for "cats", unless you're some modern-day Montaigne who can teach us eternal truths about cats just by speaking about his own).

I would even argue that Wikipedia hurts findability to some extent, since it relies on anonymous editing, and reputation builds on identity. Knowing who wrote what is a very useful piece of information (something Quora is addressing).

What about collaborative editing? Successful collaborative editing is made of two parts: software and a community. The software problem is largely solved; a working community of people who know and care, is the hard part. But it's also in the process of being solved by what StackExchange is doing: building sites around existing communities.

So here's what I think will eventually happen:

- human-machine interface / findability engine: Google

- clusters of (constantly updated) information: StackExchange sites (and others)

- niche content: your blog

In this future I'm not sure there's much room for Wikipedia; Wikipedia was built around the idea that there is such a thing as "static" knowledge or "eternal truth". There isn't: knowledge is a process and a conversation.

(Wikipedia is in fact a kind of cache for the current state of knowledge, but that problem could probably be solved differently and more efficiently.)


Wikipedia is "the mainstream consensus", a compilation of knowledge that is generally agreed upon by reasonably sane and well-informed people, in a format accessible by the layman. I think that will continue to be valuable.

If you are a layman who wants to know "what is a computer" or "what is programming", then Wikipedia is a much better starting point than Stackoverflow or Don Knuths personal blog.


Hey! This is my project, and I almost missed this thread entirely! @$&*!# offline time!

The ∞³ blog has been thin on updates but I'm still aiming for a launch of core functionality soon and I'm very interested in feedback/collaborators.


My biggest complaint with Wikipedia isn't deletionism, but epistemology. Stating that the sky was blue in London on Monday May 30, 2011 demands less rigorous verification than claiming that the rulers of some state did or did not commit genocide. One is superficially inconsequential while the other is likely to lead to a lot of strife.

To give a concrete example, a prominent US geneticist states in the biography on his personal website that his biological father was a highly decorated American pilot. This is a fairly uncontentious claim, yet fails to meet the Wikipedia standard of verifiability.


This is the difference between saying "the US geneticist's father was a decorated pilot" and " the US geneticist claims his father was a decorated pilot". The latter would be acceptable in Wikipedia, wouldn't it?


There are some major problems with Wikipedia but I don't think deletionism is big enough to sustain a new effort.

My biggest gripes are that many articles are just too long. I'd like to see a word limit or something.

Also, I really don't like the emphasis on citations. It leads to articles comprised no more of disorganized sets of citable sentences. Somewhat counter-intuitively, the citation emphasis leads to less popular but citable statements dominating in many places.


Jason Scott proposes a different set-up[1], where he compares an evolution of Wikipedia to how newsgroups got set up, with different servers choosing to retain articles based on whatever rules they are run by. He, of course, envisions himself subscribing to one that would retain everything, notability be damned.

[1] http://ascii.textfiles.com/archives/2646


Doesn't the article say that this (forking wikipedia) has been tried several times and never worked?


I was ranting about this issue for long time, really like that someone is working on that.


How is this different than everything2?


It'd be great if everything2 was converted into a wiki instead of the individual submission model it follows now.


It depends - for purely factual content, it would work well. However, it would also discourage the random articles, essays and writings which give E2 its particularly personal voice. Wikipedia is very much about stating the facts about each subject in an encyclopedic way; E2 often ends up being more about several peoples' (possibly differing) opinions on a subject. They're both very valid approaches, and I'm glad both can exist at the same time.


Even E2 will delete content: http://everything2.com/title/Writeup+Deletion

I haven't written anything on there for a very long time, but I do hold the dubious honour of having a writeup (a short treatise on the difference between terminology for bread products) deleted for being pretty much plain wrong :)


I outlined some of the intended differences when the same question was asked at the Infinithree blog:

http://infinithree.org/post/2616968223/how-does-compare-to-e...


Smells like vaporware. I never really trust it when people first make a blog and a domain and a fancy headline and a fancy logo, and write prose with lots of "will be's" and "will have's", all without showing anything substantial.


I'm delighted that two unicode characters smushed together ('∞³') strikes you as a fancy logo. If you're patient enough, you'll probably like the other design choices that are coming, too.


Don't take his criticism too hard. It's just part of the startup culture: Get something out there and in front of real users, as early as possible, so you can begin the virtuous cycle of success.


No worries; this isn't my first spin around the roller rink.

I've struggled with the question of whether it's better to talk in advance of something to show, or just wait until launch and let the site speak for itself.

I've decided to talk about the vision and general themes before tangible launch for two reasons:

(1) With any such community-built project, who shows up in the early days will have a giant influence on how it develops... and I'm hoping a little bit of advance discussion will mean more of Infinithree's natural userbase can find it and participate as soon as it's ready.

(2) We'll need plenty of help. So if the general outline strongly resonates with potential deep (even eventually full-time) collaborators, I'd love to start working with them even before public launch.

So I'm taking the risk of both the immediate 'vaporware' and eventual 'underwhelmed' reactions with full awareness of how these things can go.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: