This is actually an essay I wrote (under my old pseudonymous account) back in 2008. It's located at Wikipedia:Delete the junk, and this is the version as modified by various editors over the years. It shows its 2008 roots in its rather aggressively-sarcastic tone, and the specific way voters are portrayed – which is more relevant to how people acted here in 2008 than today, I hope – and the references to <ref> tags, which were just coming in at the time, and did, certainly help.

Sometimes, an article comes up for AfD (“Article for Deletion”), which, though its subject may be notable, has no redeeming qualities whatsoever. Perhaps its only source is a promotional, questionable website. Perhaps its material seems to be completely made up from thin air. In such cases, just delete it. Wikipedia lacks articles on a lot of things, and, if the people who found 87 blog and chatpage sources using the University of Google really cared about the subject, they'd find reliable sources to remake the article.

In the end, Wikipedia can only maintain articles at sufficient quality if there are people interested in improving them according to Wikipedia policy. Where large walled gardens exist, it may be necessary to cut them down to a few, manageable articles, so that they can be brought up to sufficient quality. This means going through the huge swaths of bad articles and picking out the worst and least notable for deletion. Likewise, fixing a very bad article on a small aspect of a larger subject may waste resources better spent fixing the articles on the larger subject.

On Wikipedia, we are all unpaid volunteers. Very often, "keep" votes on these sort of articles will be combined with an insistence that... other people rewrite the article from scratch, whereas the person saying this has no intention of editing the article at all. If you're insisting other people do work creating an article on your behalf, and claiming you have the right to do this, you need to rethink your position: If you are not willing to take responsibility for improving the articles you gaily vote to keep, then you are making the jobs of the people genuinely trying to improve Wikipedia by upmerging content, reducing walled gardens to a manageable number of articles, and trying to use limited resources effectively much, much harder.

It is worse to have an article on a notable subject than not to have it, if it contains information that is misleading, or could be slanted, due to a lack of sources to verify the text is still accurate. Some articles have been hacked or slanted with incorrect text, for weeks or months, because the text was not compared to reliable sources and corrected. That problem is being reduced by use of ref-tag footnotes ("<ref>...</ref>") that pinpoint each statement to a particular source, for rapid verification. (NB: That's a pretty 2008 thing to say, isn't it?) The goal is a balance: to make articles tamper-resistant but also allow for improvements, with updates for later research or news reports, by anyone in the world.

This sort of attempt at misleading the reader can often be identified at Articles for Deletion. Horrifyingly, though, some people don't care, and instead insist the article should be kept, even when the entire article is demonstrably full of such attempts to mislead, and thus cannot be trusted, in the idea that other people should, once again, fix the problems they don't want to do the work to fix. This is wrong. Neutral Point of View is a core policy, and if the article has no redeeming merits, then the mere theoretical idea that a (completely different) article could be written on the subject which would be acceptable under Wikipedia policy is not an argument to keep.

Imagine you wanted to build a house, but the sewer main has just burst, spreading sewage across the area where it's to be built. You'd fix the sewage main and clean away the sewage first, leaving yourself with a clean, pristine area on which to build your new house. And yet, on Wikipedia, we can sometimes insist the sewage remains until the house is finished.

A badly written, poorly structured, and, especially, a POV-ridden article can be a nightmare to edit, and can intimidate editors away from it. It gives the perception of a monumental task, which has to be done all at once. And if there are any problems with claimed ownership of articles, any attempts at improvement can be halted before they even start.

However, a clean slate offers the chance to do things right. A new editor can come in, think about how best to structure the article, and create a much more useful framework for further work. It also gives permission for the article to be fairly short, but with the potential for expansion. It's just much more pleasant to work on a clean slate, than in a cesspool of sewage.

Of course, sometimes an article isn't entirely junk. Perhaps it could be partially salvaged?

• Check previous versions: Check the article's history to see if a good version exists that can be reverted to.
• Stubbify the article: If the subject's notable and not something that can easily be redirected, but the existing content is unusable, then cutting everything that's poorly sourced or problematic might salvage a small part of the work.
• Redirect the article to a relevant one of better quality. This isn't quite deletion, as the history can still be seen, and anything relevant can be picked out. If there's actually usable content, a merge can be used instead. For obscure subjects, especially ones prone to fringe or pseudoscientific beliefs or pushing a point of view, we can much easier maintain them if there aren't a lot of articles largely duplicating the same content. Also good for stubs that have no real way to expand.

The point is that it's better to have nothing rather than something that's actively misleading, unreadable, or, for more fringe subjects, part of an unmaintainable mess of interconnected articles. Lacking an article encourages people to create one. And they'd surely do a better job at it than whatever terrible mess got someone linking you this essay.