Sorry, I think I'm not being clear here (and no, I haven't read that book, but I've read a whole lot of the software-design-usability-book-of-the-month type things and have worked in software design and usability research before, so though I have a pretty good grounding in the field, I might not be making the exact arguments you're using) -- basically, what I'm saying is, 'your' (generic you) LJ experience/taskflow/what-have-you is, statistically speaking, most likely identical or highly similar to the experience of your friends list out to about the second degree (more-or-less, and yes, I am using weasel words here but that's because it's an inexact science), and more or less diametrically opposed to the experience of Random User N taken from outside your cloud. The core underlying technology (the LJ site itself) is the same, but the purposes for which you use it and the methods you've used to adapt that technology are entirely different.
Which means that when we change one thing to make that core technology "more intuitive" or "more usable" for Alice, Betty looks at it and says "wait, what? That breaks everything I use LJ for!" That's what I mean by everything being an edge case; every problem that's a core problem for one subset of LJ users is an edge case for another. The subset of problems that everyone can agree is a problem is very, very small (and have mostly been nailed by now, random Acts of Technology and operations issues notwithstanding). The stuff we're trying to fix now are usability problems evidenced by business statistics and metrics, which won't be visible to the average user; they're not core infrastructure problems, they're human adaptability and usability problems, and the subset of the overlaying social environment most of the vocal people are exposed to don't see those problems because they either don't use LJ in that fashion or because they've trained themselves into not finding it a problem.
We actually have something like five major problems we've identified as being serious, critical flaws in the entire LJ core technology, and some of them can get quick fixes and some of them need systemic re-visioining. I can't publicly identify them, but -- without knowing a thing about you or how you use LJ -- I would bet you $29 and an alligator purse that you would probably agree with one or two of them, think one or two were "eh, whatever", and think at least one of them wasn't a problem at all. Statistically speaking, that's how it shakes down. I don't know which problems you'd ping on and the like, but just from watching which subsets of LJ freak out about which fixes for which of the Core Problems because they think it's not necessary, I am very, very comfortable saying that nobody who doesn't have the big picture can identify all five of them.
(And I'm not saying that we have all of them identified, either! We're far from perfect. But from where we sit, there's some really interesting stuff implicit in the fact that LJ is simultaneously both a technological product and a community product, and fixes for problems in one often cause more problems in the other...)
no subject
Which means that when we change one thing to make that core technology "more intuitive" or "more usable" for Alice, Betty looks at it and says "wait, what? That breaks everything I use LJ for!" That's what I mean by everything being an edge case; every problem that's a core problem for one subset of LJ users is an edge case for another. The subset of problems that everyone can agree is a problem is very, very small (and have mostly been nailed by now, random Acts of Technology and operations issues notwithstanding). The stuff we're trying to fix now are usability problems evidenced by business statistics and metrics, which won't be visible to the average user; they're not core infrastructure problems, they're human adaptability and usability problems, and the subset of the overlaying social environment most of the vocal people are exposed to don't see those problems because they either don't use LJ in that fashion or because they've trained themselves into not finding it a problem.
We actually have something like five major problems we've identified as being serious, critical flaws in the entire LJ core technology, and some of them can get quick fixes and some of them need systemic re-visioining. I can't publicly identify them, but -- without knowing a thing about you or how you use LJ -- I would bet you $29 and an alligator purse that you would probably agree with one or two of them, think one or two were "eh, whatever", and think at least one of them wasn't a problem at all. Statistically speaking, that's how it shakes down. I don't know which problems you'd ping on and the like, but just from watching which subsets of LJ freak out about which fixes for which of the Core Problems because they think it's not necessary, I am very, very comfortable saying that nobody who doesn't have the big picture can identify all five of them.
(And I'm not saying that we have all of them identified, either! We're far from perfect. But from where we sit, there's some really interesting stuff implicit in the fact that LJ is simultaneously both a technological product and a community product, and fixes for problems in one often cause more problems in the other...)