The article was clearly talking about repairing bot edits; since bots are almost* universally restricted to editor capabilities, all their actions are undoable. I have myself occasionally reverted bot edits - but very rarely, since generally bots are very good at what they do, and as the article says, they have made a huge difference in vandalism fighting.
(On the other hand, the article omits one of the major instances of bot abuse: the use of scripts/bots by a few editors to ram through removal of all the {{spoiler}} templates despite the absence of consensus to deprecate their use.)
* I remember a few proposals for admin-bots, but I don't remember whether they passed the Bot people or whether they were to do irreversible actions like deletion.
Wikipedia has never been about collecting sum of human knowledge, as it claims. It's about collecting sum of human knowledge that it can verify.
Wikipedia is designed from the beginning to be a tertiary source, and it's not an accident. This is why if you try to put any original content in it, it will usually get reverted, not because the content is wrong or even unhelpful, but the fact that nobody can verify it. (I'm not saying whether this is a good or a bad thing, it's just the way Wikipedia works) If you want to contribute original knowledge, don't use Wikipedia as a vehicle for doing so, you are better off writing to a blog/magazine/academic journal.
Then it must absolutely enthrall you to think about how much human knowledge is made available by wikipedia. You can be just as persistent as a "delitionist", theres nothing stopping you. It doesn't mean the system is broken just because some people are passionate about it.
I thought that line of reasoning was completed years ago. Yeah, the people who are more willing to waste their life on Wikipedia policy pages will win. Now, what are the bad effects of that?
For what it is, wikipedia is incredibly effective. If you don't care about wikipedia policy, how can you care about the exact content of wikipedia?
Then someone had the idea of putting a template with the "importance" of the page in the talk page. So now every page has a talk page with text, and I never check them.
It used to be I would ask questions in the talk page and get an answer within hours, now I'm lucky to get an answer that year! (Not exaggerating.)
The amount of content that has been lost by vandals changing text to garbage, and then someone comes and removes the garbage - but doesn't revert the edit, is staggering. I think it's time to auto-lock virtually all old pages, require a second opinion on every edit.
And yes, it is staggering, I am not exaggerating given that I've personally restored quite a number of pages - some of them have sat completely gutted of content for a year.
This is rather a broad brush, and a gross oversimplification of a complex debate that's been going on for a long time.
A curious thing with Wikipedia is that lots of people think they know the obviously right thing it should do, but many of these "obviously right" things are very different from each other. For example, co-founder Larry Sanger split and founded Citizendium because he thought Wikipedia was far too permissive in letting "unencyclopedic" crap into the encyclopedia, which is the exact opposite complaint of the people who are worried about "deletionism". That is also probably the most common complaint about Wikipedia from academics and in the mainstream media. The mainstream media gets particularly inflamed if something incorrect is found in Wikipedia, like the "Seigenthaler incident", and demands that Wikipedia should institute stricter edit controls.
One area of particular interest to me is where to draw the line on science articles. With no policies on inclusion at all, there would by now be thousands of physics articles on concepts that are not recognized by the physics literature, created by the same fringe-science people who post prolifically on Usenet, and cited to their own websites as a source. They tried to do so on Wikipedia in 2003-04, and that was the impetus for some of the policies such as "no original research" and "must be verifiable in reliable sources". (Fringe physics theories that are well-known and where third-party documentation exists, such as Time Cube, can of course still be covered.)
I wrote a bit more on that verifiability/notability history last year, in case anyone is interested: http://www.kmjn.org/notes/wikipedia_notability_verifiability...
But only as fringe physics, which is where another great big bunch of ill-will towards Wikipedia is from: Even if your pet theory gets an article, it isn't going to be treated the same way as Quantum Electrodynamics. Most of the citations will be to the reliable sources, which are by the people who think you're a crackpot. This causes a lot of pain to the people who insist 'NPOV' means 'Treat my nuttery like it was a real Grown-Up Person science'.
Since Wikipedia closely tracks edits, however, mistakes can be repaired almost as quickly as they happened, administrators say.
I think fairly consistent commentary over the years demonstrates that this is patently false. Deletionists are capricious and arbitrary, reversals are beauracratic, complex and lengthy (if even possible) It dissapoints me to think about how much legitimate human knowledge has been wiped from WP, the authors discouraged and WP made poorer for it.