User:Jorge Stolfi/DoW/Deletionism is evil

[This is an expanded version of a reply that I wrote for the 2010-01 RfC on a new policy for deleting unsourced BLPs]

Longer version

This page is one megabyte long, and I am still unable to see why it was started in the first place. What I have learned from what I could read of it is that "unsourced BLPs" are not a real problem, and the "damage" that they are causing to Wikipedia is purely imaginary.

Some years ago, a group of editors decided that "getting rid of non-notable articles" was to be a top-priority goal of Wikipedia. "Getting rid of unsorced BLPs" is basically a consequence of that decision --- rationalized over the mere possibility that the subjects of those BLPs may be non-notable, and the decision that the burden of proof about notability falls on the creator and contributors of an article, not on the editors who wish to delete it. The editors who took that decision never bothered to seek the opinion of the general Wikipedia community, much less its consensus. Now those editors are realizing that even the mock-trials on the AfD cucking stool require too much work, and at the current pace they will never reach their goal. Unfortunately, the other editors are not helping: they do not nominate or vote articles for deletion, they keep creating more and more "improper" articles, and they generally complain when their articles get deleted. Well, to me, this only shows that the goal is everything but a consensus. Indeed, the so-called "deletionists" seem to be a small minority of all editors. Unfortunately, they have been able to impose their goal over the other editors, because deleting an article is so much easier than creating one; and because deletions, unlike other edits, cannot be reerted by ordinary editors.

In spite of what the deletionists and their guidelines say, the "notability rule" --- the notion that an *article* may be deleted solely because its subject is "non-notable" --- never was a majority opinion, much less a consensus. The length of this page shows that. The dozens (hundreds?) of talk pages out there with unending debates on this topic show that. The backlog of unsourced BLPs most clearly shows that. Indeed, each one of those 50,000 unsourced BLPs is a clear and emphatic vote against any policy of deleting unsourced BLPs. Moreover, every editor who created or improved one of those articles invested far more work and thought into his "vote" than what it takes to delete the article. So, if anything, each of those "sweat votes" should carry more weight than the typical vote for deletion --- which is often cast without even reading the article. But even without any weighting the outcome of this unintended and implicit poll should be pretty clear --- if we consider that Wkipedia now has much less than 50,000 active editors.

The root of most of our problems (including the unsourced BLP "problem"), is that the number of articles keeps growing, while the pool of active editors is shrinking. Until 2005, editor recruitment for the English Wikipedia was growing exponentially, in step with the article count; since 2006, the article count continued growing at almost constant rate, but recruitment of new editors is practically nil, and old editors are steadily leaving. The only way to solve this root problem is to make whatever radical changes in WP policies are needed to restore the pre-2005 recruitment pattern, so that the pool of editors may again grow proportionally to the number of articles.

In my view, scrapping the "notability rule" should be part of those necessary changes. I believe that most would-be editors only become "hooked" after they create one or two articles on subjects that are dear to them, and those generally tend to be on "non-notable" topics - like their elementary school, local church, company, favorite book, etc.. The aggressive deletion of well-meaning newbie *articles* only because they are "non-notable" may not be the main reason why we are not recruiting, but it certainly does not help. A non-notable article would then be a very small price that we need to pay for a new regular editor, who may eventually contribute hundreds of man-hours to Wikipedia --- *including* the cleanup of hundreds of other articles.

Here is anecdotal evidence: in jan/2010 I casually ran into a dozen articles about schools in Winnipeg. From the general style an other clues, it seemed that they had all been created in 2008 or earlier, more or less together, by a single person, probably a staff worker of a particular school disctrict in Winnipeg. While the articles had many style problems, every one of them cited a reliable source for the basic info, that could be easily checked with one click. It was obvious that their author and other later editors had put many hours of bona-fide work into the lot. Unfortunately, that person apparently stopped caring for the articles soon after, when he/she was still halfway through the list of schools in that district. That apparently happened after 2--3 of those article were deleted for being about K-9 schools (which are "non notable" by defintion, according to the deletionists). I can well imagine how that person must have felt, because I myself spent several hours checking and cleaning the surviving articles --- only to see another half a dozen of them being summarily deleted (by a single admin who did not even bother to notify me first, or list them in the AfD).

There must be hundreds of thousands of elementary and junior-high schools in the English-speaking world. If we could convince only one teacher or principal from each school to become a WP editor, even at the level of half an hour per week, it would make a *huge* difference to Wikipedia. That will hardly happen if, on their first attempt to create an article, they get rewarded with rude tags or summary deletions.

In conclusion, the "notability rule" (including its corollary that every BLP must include references that prove notability) is not a solution to any real problem: it is itself a major problem. So, please, let's scrap it, and let's make its opposite into a truly official policy. Namely, it should be a rule that an article can be deleted *only* if it cannot possibly have *any* valid contents at all (e.g. Jimmy Hoffa's last words, Credit cards of Hollywood actors, What Britney Spears had for lunch on 2010-01-25) or if its *mere existence* would do *actual harm* to Wikipedia (e.g. timmy jenkins is an idiot or Sexual misbehavior of Joe Somebody) Specifically, no article should be deleted simply because the subject is not "notable enough", or because there is a chance in 1,000 that it might be vandalized without anyone noticing. In particular, let's stop bothering about non-notable articles and unsourced BLPs; and let's restore all those articles that were deleted for the sake of that rule alone, with an e-mailed apology to their authors and contributors.

Consequences

Critics have commented that this proposal would allow "a BLP on every casual editor and his school chums". Not necessarily. My proposal still allows those bois to be deleted, by the rule (uncontroversial and current) that authorizes deletion of an article that cannot possibly have valid content (such as last words of Jimmy Hoffa or Credit cards of Hollywood celebrities); and by the rule (also uncontroversial and current) that an article should only contain verifiable factual information that, by editor's consensus, is considered relevant and "encyclopedic" (in the context of the article).

The last rule authorizes editors to challenge (by simple deletion, with a short note in the comment line) any statement that they deem suspicious or which they believe is unverfiable, irrelevant, OR, etc. If the the article becomes empty as a result, and none of the deletions are successfully contested within (say) one year, then the article can be deleted. Note that an empty article will not show up in searches (other than searches for its name) and cannot possibly lead to complaints, so there is no harm on keeping it around for a long time. On the other hand, the history and talk page will remain accessible, so any editors can examine the deleted maerial, and restore it easily (if such is the consensus of all watchers).

Now most ninth-graders will find it difficult to prove their own existence, not to mention any other factual statement about themselves; and they will have to put up with cumbersome article names. At best, he will be able to create an article called Tomas Jenkins (student at Glenview School, Peoria, IL) that contains just one line saying "As of September 2009, Thomas Jenkins was a registered student at Glenview School, Peoria, IL." Even if we pretend that these are relevant and "wikiclopedic" pieces of information (just to be nice to the kid), they will survive cleanup only if someone can find a supporting source that is trustworthy and accessible, such as the online student roster at the school's *official* website (which of course will hardly exist). On the other hand, if the kid wins a prize or a competition *that is reported on a trusted and available medium*, that information can be added too.

In the same way, a company employee whose job has no public visibility will probably be unable to provide proof of his present or past employment, or even of his existence. A teacher at an elementary school will be able to create an article William Smith (teacher at Glenview School, Peoria, IL) with the same basic info plus perhaps data on diploma and prior teaching jobs, all easily verifiable in official websites. An MD will be able to prove his diploma, the manager of a supermarket will be able to prove its street address and floor area, a rock band can prove its

So in practice my proposal will not open the doors to a hundred million articles on K-9 students. Even if we get 50,000 such articles, they will not make us waste as much work as defeding articles on the AfD or debating with the deletionists. The main effect of my proposal will be to eliminate the rude welcome "I am a Mighty Administrator of Wikipedia, you please prove that this person is notable or else I will delete the article". Instead the dispute will be limited to the article *contents*, with the creator and critics on the same level. In fact, this "delete the content" approach will probably demand less work by everybody than the current AfD mechanism (which requires a dozen editors to look at the article).