I am going to start in one place & end up somewhere else. Ready? Here we go.
There's a wonderful trend lately at library conferences of promoting open dialog around failure. I was first acquainted with this at the #drupalfail sessions held by LITA's Drupal Interest Group. Presenters would detail the various ways their projects crashed & burned, or merely did not meet expectations. With Drupal, this is particularly easy: it's a complex
Related news flash: most librarians don't learn Drupal in library school, it's something they learn on the job, so there's a lot of intermediate failures before anyone gets close to something that vaguely resembles success. But Drupal isn't the only example of this trend: I heard good things about Code4Lib's "Fail4Lib" preconference. There have been scattered talks elsewhere discussing the need to create a culture where taking risks & occasionally failing is a welcome. It's certainly a necessary element of innovation.
What stands out about these failure sessions? They're useful. Knowing someone else's mistakes saves you immense amounts of time & often all you have to do is avoid something stupid to gain from it. As a technologist at a small library, I'm constantly bombarded with awesome things I can't use: they require money, or staff, or time, or expertise, or scale that we just don't have. It's cool to hear about Linked Data & Near-Field Communication; it's not something that would be a wise investment on my part. But when I hear someone say that creating a custom theme from scratch in Drupal is a waste of time relative to using a pre-built theme, I instantly am more prepared to do my job. Don't reinvent the wheel with theming, check. Lesson learned, time saved.
When you consider the pedagogical value of failure, some weird issues arise. I had a unique undergraduate career in that I was trained in both the humanities (English) & formal sciences (Mathematics). You know what both of those fields happen to be utterly terrible at? Teaching failure. In math, when a theorem is superseded, it's simply not taught anymore. It might as well have never existed. I never had homework problems phrased "spot the problem with this theorem" or "hey Fermat was a dummy, can you tell why?" Mathematics ignores an entire mode of analysis. You become skilled at deductive reasoning & constructing your own theorem cabins from axiom Lincoln Logs; you never learn how to approach someone else's theoretical edifice other than simply assuming it's true because it's in the textbook.
English is also awful at admitting failure, in its own warped way. We read the classics, but not the failed classics. There are at least two kinds of failed classics: works which were highly regarded in their own time but grew irrelevant & works which were never highly regarded in any time. Either way, why are the canonical works more valued than the telling failures of their contemporaries? While Mathematics education's failure to teach non-deductive modes of logic is troubling, artistic prejudices are even more disturbing.
Everyone has, by now, heard "beauty is in the eye of the beholder." The aesthetic disciplines cling to this maxim as if it somehow places them outside the realm of objective inquiry, paradoxically able to pass judgment without recourse to supporting evidence. If this is true, why do we read Shakespeare?  Wouldn't any arbitrarily chosen text suffice, given that the text itself is irrelevant, it's the Beholder that matters? Of course, what you learn in English is that—to paraphrase George Orwell's Animal Farm—some Beholders are more equal than others. Your professors are Beholders, you as a student are but a Beholder-in-training, & art works are unimpeachable: they do not fail, they either go unmentioned or become canonical via mysterious means. Aesthetics masquerades as subjective judgment while never admitting its own folly or interrogating the social conditions that cause certain works to become canonical while others are summarily discarded.
Doubtless there are objections that I'm conflating disparate fields. Mathematics is axiomatic logic, English is aesthetics, & the specific vein of librarianship I've mentioned is quite practical. These are library projects that failed & perhaps we cannot say a work of art or a theorem fails in any corresponding sense.
But don't give up on me so easily. Librarians are onto something here. We know that art fails. We don't purchase every book, we write harsh Goodreads reviews about books that didn't please our Beholder's eye. My earlier examples were from library technology events, events that skirt around the practice of programming if not engage it directly. & what is a failed program, or a bug in an algorithm, if not a flawed theorem? There are connections & they might even be meaningful. Otherwise I'm just way off base, a deranged squirrel collecting copper washers for the winter. I never was good at aesthetics as theory. I probably should avoid writing about it. But I'm pretty good at failing & perhaps I'll write more about that.
^ The thing about Shakespeare: he's not a very good writer. He has flaws & they're the sort creative writing teachers spell out in red sharpie at the end of student plays: "Heavy handed." "Deus ex machina." "Did you really back yourself into such a corner that the only way out is to kill every single character for which the audience has a shred of empathy left? Please, go back to the drawing board." I'm still bitter about Shakespeare, a sole author, being a required course for my undergrad degree, which utterly ignored the entire 20th century.