azurelunatic: "My user interface is pastede on (yay)": scenes from an Access database that is not working so well.  (ui)
Azure Jane Lunatic (Azz) 🌺 ([personal profile] azurelunatic) wrote2006-12-15 02:29 am

I trust the LJ development team. Do you?

I trust LiveJournal's development team to have LiveJournal's best interests at heart, to lead LiveJournal in a good direction, and to listen to constructive feedback.

I will do my best to make the feedback I leave for LiveJournal developers constructive in nature.



If you wouldn't do it in fandom, don't do it to the devs. I know that this is the choir section here that I'm ranting at, but I've spent the past couple hours in a room with some rather irritated engineers who are really code people, not people-people. They've been busting their asses for months to track down random crap that goes wrong. This site is so bloody huge and robust that Bantown could not take it down for long, even though they tried. Slashdot fails to have the Slashdot Effect on LJ. LJ is thriving and functional thanks to the developers who put it together and the engineer-types who keep it running day-to-day and the people who keep the money coming in to feed the monster bandwidth and all the rest of it, and the people who make sure that other people know how to use it, and the people who stay here and hang out and talk with friends. The developers work hard to keep things working and keep the site evolving so it doesn't become a great big code dinosaur. Lately it's been seeming that the harder they work to fix things that are broken and update things that are out of code (building code metaphor, not computer code; work with me here), the more they get screamed at for trying to ruin LJ.

In every [livejournal.com profile] news feature-type post where something new and bell/whistle is announced, there is the inevitable complaint that things like virtual gifts are a waste of developer time that would be better spent on problem X, Y, or Z. And when LJ has been having a couple weeks where there are problems, and the problems stay there even though people are complaining about them, and the problems are still there, and still there, and still there -- yes, it does seem illogical that developers would go and do something like make it possible to put a flaming bag of poo on your least favorite serial adder's profile page. But sometimes you have to step away from a problem to get it back in perspective. I'm not in LJ Central, so I'm not there watching them bang their heads into a stubborn problem until headaches ensue, but I trust that they are allocating their time reasonably.

You know what I think the number one biggest waste of developer time is?

Dealing with unaccountably rude and hostile users.

LJ as a culture has the hugest sense of fandom entitlement ever.

LJ users want the same thing they've always had from LJ, namely, a place to put their journals and communicate and be with friends, and a geek-friendly, open, caring, open-source, user-supported, small-town environment.

LJ geeks want pretty much that same thing. Really. Truly.

Somewhere along the line, LJ users as-a-collective got the idea that if the development team did something that they didn't like, the best way of solving this was not to give constructively critical feedback and debate it with vigor and the knowledge that the developers had the good of the site in mind, but to jump on any available surface and flame away.

Imagine the utter fucking joy that the LJ developers must be having, wading through gods know how many hundred comments of flame to find the legitimate kernels of actual problems in between the complaints. Go through one of those posts announcing changes to LJ some time, and pretend that the changes to LJ are a fic that's already been beta-read, and the comments to those posts are comments in response to the fic. Read those comments with an eye to constructive criticism. The analogy doesn't stretch particularly far, because the core site pages of LJ are not a piece of fanfiction, but the principle of effective communication holds true.

Dear users, the way to get the development team to listen to your concerns is not to scream abuse at them and then expect them to abandon their ideas of what is right for the site and adopt yours. The louder you scream, the louder they're going to hit the delete key and say "Na na na can't hear you na na na." I don't actually think they're doing that now, but the temptation is very much there and very much real. LJ is a maverick site in that it has such open forums for user feedback and discussion. Plenty of services do not have anything resembling that. Do you really want to convince the developers and volunteers that an open forum will only collect whining and flames?
Hint: Bantown tried forcing the issue by attacking LJ. We all know how that turned out. Pwned, craxx0rbitches, pwned. In a similar case, visible nipple is still not allowed in the default userpic, and the flaming tantrums thrown at LJ's support staff by assorted self-proclaimed "boob nazis" have assured that visible nipple will never be allowed, on the principle that it's bad precedent to cave when the toddler has a meltdown because they didn't get their little way. Even though there are many people who do love the boob.

Tell them what you like about the shiny new stuff. Let them know what they did right. Sit on your hands for a few hours until you try using it a few times before you flame off at them. If you have to say something immediately, remember what you learned in those sensitivity training sessions and use your "I" statements. "I'm frustrated with this new user interface, and I'd really prefer something with the look and feel of the older version" comes over a whole lot better than "What the fuck did you do to my user interface, you morons? I liked it the way it was! Put it back!"

LJ, even current LJ under 6A management, is capable of recognizing if something goes really badly. The developers actively ask for reports of broken or unusable behavior. Things may not be fixed immediately, but there are little things coming out every here and there to make things better, things that you may not be aware of unless you're watching [livejournal.com profile] lj_releases or [livejournal.com profile] changelog.

LJ really is a group effort. I do not have Super-Secret Inside Information that no one else has. I'm a relatively average occasional Support volunteer. (Very occasional, since Life Attacks.) I put time and effort into making LJ a better place, and I see the results of that effort. Things may not always go my way when LJ policy and I disagree with each other (I wouldn't mind seeing nipples in any boobtacular default user pictures, for example), but at least my technical suggestions are often dead-on, and my social suggestions are at least listened to respectfully.

I really do think it all boils down to three or four questions:
  1. Do you trust the people who are running LJ, including Six Apart core and the developers?
  2. If you do not trust the people running LJ, what can they reasonably do to demonstrate that they're worthy of your trust?
  3. If there is nothing the people running LJ can do to gain your trust, why are you still here?



And you know? I find that I'm never short on database handles after this update. How about you?

[identity profile] selenite.livejournal.com 2006-12-15 07:55 pm (UTC)(link)
I trust the developers. But I'm seeing them make the same mistake I've seen in other projects, one which I'm prone to myself: focusing on the programmers' priorities instead of the customers' priorities. The new update screen isn't bad, it only took me a few seconds to find the stuff that moved, but that's not what I'm objecting to. The old one worked. People were content with it, so any change, no matter how much of an improvement, would get objected to. And there's not much improvement between the new and old one, it's just fine-tuning.

Fine tuning dialogues that work well is a classic programmer behavior. "This must be PERFECT!" is the thought, so they're going to fiddle until it meets their aesthetic ideal. Meanwhile there's other tasks--less fun ones--which are lying undone even though they'd contribute more to user satisfaction. Or there should be. If rearranging the text fields in the update screen is actually the most important thing on the development team's to-do list they need to step back, proclaim LJ finished, and start something totally new. Something whose worth will justify the value of the programmer hours put into it.

On a more practical note, is there a PO Box for the dev team? I'm tempted to send them copies of Peopleware and The Inmates are Running the Asylum as references.

[identity profile] cmwinters.livejournal.com 2006-12-16 11:10 am (UTC)(link)
My biggest problem is that they add things that people don't seem to want, and ignore the things people do want, and then seem quite put out when people aren't delighted with the new stuff they didn't want in absence of the useful stuff they did want. (I.e. - "no you can't have full-text searching but LOOK! New hideous bright green and fuschia styles!!! That makes the WHOLE SITE better!") *shrug* It's why I don't even bother posting to [livejournal.com profile] suggestions anymore - it's a great idea but it's the black hole of nothingness at LJ, and I don't need false hope because it's frustrating and demoralising.

Neither, however, do I bother ranting and screaming. Not any more. :P

[identity profile] selenite.livejournal.com 2006-12-19 03:50 pm (UTC)(link)
My problem with [livejournal.com profile] suggestions is that it's damn hard to find if something's already been suggested. My one suggestion was shot down with "that's already been discussed, check the memories" which was very annoying since I'd spent 3-4 hours digging through the memories looking for anything similar.

(no subject)

[identity profile] selenite.livejournal.com - 2006-12-20 19:51 (UTC) - Expand

(no subject)

[identity profile] museumfreak.livejournal.com - 2006-12-24 04:37 (UTC) - Expand

[identity profile] vvalkyri.livejournal.com 2006-12-20 10:18 pm (UTC)(link)
Worse, things getting shot down as already suggested when it's not really the same thing, or, as with my request to make it so I could view my own journal by security level, "tags will take care of this." Okay, tags didn't, but then I got rejected as already suggested.

And it took a heck of a lot of effort to get a suggestion through about the security hole in the notifications system, -- since one way to fix it would be to NOT automatically unscreen comments I was gettign rejected as already suggested. Feh.

[identity profile] kunzite1.livejournal.com 2006-12-24 02:28 am (UTC)(link)
i'd like a message more like "already suggested, check the memories under this category or the tags under this tag to find it. if you think it's been rejected in error, please email [email protected] with your thoughts on the matter."

(no subject)

[identity profile] selenite.livejournal.com - 2006-12-24 04:24 (UTC) - Expand

[identity profile] kunzite1.livejournal.com 2006-12-24 02:27 am (UTC)(link)
i posted an entry (http://kunzite1.livejournal.com/744657.html) to my journal about it earlier. i wasn't able to figure out how to exactly draft a [livejournal.com profile] suggestions post.

if you'd like to, go right ahead.

and after reading the entry, you already commented on it! XD

(no subject)

[identity profile] kunzite1.livejournal.com - 2006-12-24 02:35 (UTC) - Expand

(no subject)

[identity profile] kunzite1.livejournal.com - 2006-12-24 03:10 (UTC) - Expand

(no subject)

[identity profile] kunzite1.livejournal.com - 2006-12-24 03:43 (UTC) - Expand

[identity profile] rahaeli.livejournal.com 2006-12-17 02:36 pm (UTC)(link)
Yeah, but what you identify as a problem isn't the same as what we identify as a problem, honestly. The thing is? Our usability studies scare us. Badly. Our task-flow metrics terrify us. We have a higher task-flow abandonment rate than any other site anyone on the team has ever worked for. The percentage of people who, say, go to post to a community, can't figure it out, and give up and leave LJ (or just drift off, because they can't make the site do what they want it to do) rather than try to figure it out is terrifying.

Sure, you can address that with documentation. You can put little pop-up inline help bubbles everywhere -- I'm staring at two of them right now writing this comment. You can write tutorials. You can do a whole lot of things. But there comes a point where you have to just bite the bullet and redesign your site to stop hiding critical paths and frustrating your users.

It doesn't seem like a priority to existing LJ users, especially the power users, because they've already figured it out and it's second nature to them; to them, it looks like rearranging the deck chairs on the Titanic. But we have hard, fast evidence that indicates that with the update page redesign, our taskflow abandonment rate dropped from something crazy like 80% to something far more sane like 30%. (I'm pulling these numbers out of my ass, since I don't feel like logging into the stats server right now, but it was a HUGE drop like that.)

So, because we can't release those stats and those figures -- to do so would be like inviting our competition to come wandering in behind the scenes and shaft us -- it looks like we're just arbitrarily deciding to change shit. We're really not. I swear to you. There's no way we'd have put priority on something like that if it were developer whim. We're way, way better managed than that. Everything aesthetically unpopular over the past six months -- the navstrip, contextual hover menus, redesigning the update page, changing the site scheme, the profile redesign, etc, etc -- has been designed for one purpose and one purpose only: to fix some statistic or another that fucking scares us, because it implies bad things for the long-term health, stability, and viability of the site.

Also, you've gotta remember -- we have two sets of engineers, front-end and backend. Having one of our frontend engineers working on redesigning the update page doesn't take time away from the guy who's working on, say, journal search -- the two positions have very, very few overlapping skillsets, and if the guy working on the update page redesign wasn't working on that, he'd be working on some other form of user interface.

[identity profile] ataniell93.livejournal.com 2006-12-18 12:31 am (UTC)(link)
The one thing that LJ could do that I personally think would shut down a lot (not all, some people will always be unreasonable) of the anger is to stop making some of the new bells and whistles opt-out and make them opt-in. There was a lot of anger about the way the navigation strip was handled, particularly the option which allows people to force others to look at it even if they've said they don't want to. (I still don't get why that was even permitted.) People who use LJ to RP a lot especially hated it because they'd have to go into 20-100 journals and change it. Defaulting to hide birthday was another one that I had to personally change in over 100 journals and I still haven't got everyone in my 200+ journal RPG to fix theirs.

The CProds? The fact that one of the LJDesign people got up in one of the anti-ad comms and said that we should not be allowed to turn them off hacked a lot of people off.

I understand what you are saying about task abandonment rates. But I'm not seeing the reason for making some of these optional changes opt-out. If it's optional, it should be opt-in. If it's not optional, don't pretend that it's optional?

Does that make sense?

Also, with regards to the virtual gifts? I really don't like the idea that people who don't like me can pay you to put things on my profile that I didn't ask for and don't want and can do so without giving their names. I'm someone who has a lot of fans and a lot of foes online, and someone actually DID pay to put a flaming bag of poo on my profile anonymously. I think the option for anonymous gifting should at the very least be restricted to the unambiguously NICE gifts, and yes, I'm annoyed that even though I set my journal up so that no-one can comment anonymously or do anything else anonymously, they can still put a flaming bag of poo in my userinfo anonymously. I'll say that here, because Azz is a friend, but I won't say that in public, because the person who did it might read it and have the satisfaction of knowing that it pissed me off.

[identity profile] rahaeli.livejournal.com 2006-12-18 12:37 am (UTC)(link)
Without getting into the rest of it because I'm rushed for time right now (and have been up for like, ugh, 24 hours) -- the problem with opt-in is that there are millions (literally) of LJ users who would have no clue that we've just fixed something that might be their primary peeve with LJ. [livejournal.com profile] news readership is awful.

(no subject)

[identity profile] torrilin.livejournal.com - 2006-12-18 01:33 (UTC) - Expand

(no subject)

[identity profile] rahaeli.livejournal.com - 2006-12-18 01:39 (UTC) - Expand

(no subject)

[identity profile] dzurlady.livejournal.com - 2008-11-24 01:23 (UTC) - Expand

(no subject)

[identity profile] torrilin.livejournal.com - 2006-12-18 04:55 (UTC) - Expand

(no subject)

[personal profile] archangelbeth - 2006-12-22 20:55 (UTC) - Expand

(no subject)

[identity profile] 7rin.livejournal.com - 2008-11-29 05:25 (UTC) - Expand

(no subject)

[personal profile] archangelbeth - 2008-12-01 21:05 (UTC) - Expand

Plug ins'n'outs

[identity profile] 7rin.livejournal.com - 2008-11-29 05:19 (UTC) - Expand

(no subject)

[identity profile] ataniell93.livejournal.com - 2006-12-18 08:45 (UTC) - Expand

(no subject)

[identity profile] rahaeli.livejournal.com - 2006-12-18 19:38 (UTC) - Expand

(no subject)

[identity profile] ataniell93.livejournal.com - 2006-12-20 04:39 (UTC) - Expand

(no subject)

[identity profile] rahaeli.livejournal.com - 2006-12-18 19:39 (UTC) - Expand

(no subject)

[identity profile] ataniell93.livejournal.com - 2006-12-20 04:47 (UTC) - Expand

(no subject)

[identity profile] rahaeli.livejournal.com - 2006-12-18 19:39 (UTC) - Expand

(no subject)

[personal profile] archangelbeth - 2006-12-22 21:12 (UTC) - Expand

(no subject)

[identity profile] rahaeli.livejournal.com - 2006-12-23 03:52 (UTC) - Expand

(no subject)

[identity profile] 7rin.livejournal.com - 2008-11-29 06:33 (UTC) - Expand

(no subject)

[identity profile] rahaeli.livejournal.com - 2006-12-18 19:40 (UTC) - Expand

(no subject)

[identity profile] rahaeli.livejournal.com - 2006-12-19 04:54 (UTC) - Expand

(no subject)

[identity profile] ataniell93.livejournal.com - 2006-12-18 08:40 (UTC) - Expand

[identity profile] kunzite1.livejournal.com 2006-12-24 02:32 am (UTC)(link)
The one thing that LJ could do that I personally think would shut down a lot (not all, some people will always be unreasonable) of the anger is to

... post to something like [livejournal.com profile] lj_dev and say "hey, here's an idea that we have that will solve this-and-this-and-this, what do y'all think?"

[identity profile] ex-shattered767.livejournal.com 2006-12-24 07:40 am (UTC)(link)
Burr argued with me over that one in [livejournal.com profile] no_lj_ads. I ended up throwing back in his face that we should be able to, because for some of us the thing is just plain useless, citing that it had told me to make a phone post. His response? "Um...you've already made phone posts? You shouldn't have seen that."

So now I have it disabled via Greasemonkey.

[identity profile] ex-shattered767.livejournal.com 2006-12-24 07:48 am (UTC)(link)
First off, apologies to Azure for spamming with a second comment.

I have to second the Vgift issue as well. I have gotten quite a few anonymous gifts (poo bags, coal lumps) with harassing messages attached. They sit on my profile page all day until I get home from work and can remove them. And I have no way to stop them. The last one I received was on - you guessed it - the day the latest newspost was made announcing new Vgifts, before I had even had a chance to see the newspost.

When Vgifts was first introduced, people said that it could and would be abused. Staff said that being able to delete gifts after the fact should be good enough, but it's not.

[identity profile] selenite.livejournal.com 2006-12-18 04:26 pm (UTC)(link)
Okay, sorry, y'all are doing better than other folks I've worked with. I really want to take a look at your data and try to break out the categories of problem triggers, but even if you'd give a copy to J Random User my day job isn't leaving me the time to do a decent job on it. Sigh.

So instead let me strongly suggest getting a copy of Alan Cooper's book on interaction design. (http://www.amazon.com/Inmates-Are-Running-Asylum-Products/dp/0672326140/sr=1-1/qid=1166458314/ref=sr_1_1/103-2025606-3515834?ie=UTF8&s=books) Despite the flame-bait title it's a very good overview of how to design software for usability at the architecture level. Tinkering at the lower level can't fix problems caused by the top level design. The best way to use the book may be to get two programmers from each team to read it and then get together for a brainstorm of "how should this work?"

[identity profile] rahaeli.livejournal.com 2006-12-18 08:08 pm (UTC)(link)
Trust me, we do an amazing amount of brainstorming, and we pay a lot of attention to how things are being used and etc -- but the problem is, LJ is so versatile that just about everything is an edge case in some way, shape, or form. People use LJ in a way that nobody ever consciously intended, and so there's no one answer to "how should this work?" and no one "intuitive" usability design.

(no subject)

[identity profile] selenite.livejournal.com - 2006-12-18 20:34 (UTC) - Expand

(no subject)

[identity profile] rahaeli.livejournal.com - 2006-12-18 21:02 (UTC) - Expand

(no subject)

[identity profile] selenite.livejournal.com - 2006-12-18 23:10 (UTC) - Expand

(no subject)

[identity profile] rahaeli.livejournal.com - 2006-12-18 23:15 (UTC) - Expand

(no subject)

[identity profile] selenite.livejournal.com - 2006-12-19 04:56 (UTC) - Expand

[identity profile] vvalkyri.livejournal.com 2006-12-20 10:23 pm (UTC)(link)
BTW, I didn't have any trouble with the old interface, but I like what you did with the new one. And the taskflow metrics are compelling.

ext_5856: (And a monkey.)

[identity profile] flickgc.livejournal.com 2007-01-31 02:59 pm (UTC)(link)
It's interesting: I find Horizon a lot less usable than XColibur, and as far as I can tell it still doesn't fix things like syndication management (although I may have missed it), but I chose to use it, because I wanted tag management, and as such I can live with my friends' page being in 'the wrong place'. (At least the mobile posting bug has *finally* been fixed, though!)

The new update page, though, confuses me every time I post to a community: I get to the end of the post, go "argh, where's the 'change journal' option?", remember, scroll back up to the top of the page, realise I wanted to add a tag, scroll back down.... Plus, the huge font is horrible, and keeps scaring me into thinking I'm somehow stuck in bloody rich text mode.

Yes, the power users already figured it out. Yes, the power users just decided that it was easier to add /poll/create.bml? to the url. And, yes, the power users are more likely to figure out the way to get around the new changes with the least hassle: just because they aren't showing up on site stats as reaching a dead end and giving up doesn't mean that they aren't finding it really bloody hard work to deal with. The non-power users, who've gradually figured out how LJ works over a period of time, are even more stuffed by it, because they're not only having to re-learn stuff but they're also not so able to figure out the new optimal path on the fly.

I appreciate that "people who've been around and would leave if it weren't for the fact that they've been around for so long" aren't a significant percentage of your users, and that (as many of them are perm members) they also aren't a significant percentage of your income stream (and must be really horrible, now, from a cashflow persepective), but they do exist. Beta access to New Stuff is nice. Consideration would be nicer.

Anyway: sorry, got here via a link. Sorry to [livejournal.com profile] azurelunatic for mini-hijacking.

[identity profile] selenite.livejournal.com 2006-12-18 04:31 pm (UTC)(link)
I'd been contemplating making that same comment in the official post, but couldn't figure out a way to stand out from the noise. I'm glad [livejournal.com profile] rahaeli responded here.

[I sometimes think nothing useful ever happens through official channels, they're just there to provide context for the backchannels]

[identity profile] nakeisha.livejournal.com 2006-12-24 04:23 pm (UTC)(link)
I trust the developers. But I'm seeing them make the same mistake I've seen in other projects, one which I'm prone to myself: focusing on the programmers' priorities instead of the customers' priorities. The new update screen isn't bad, it only took me a few seconds to find the stuff that moved, but that's not what I'm objecting to. The old one worked. People were content with it, so any change, no matter how much of an improvement, would get objected to. And there's not much improvement between the new and old one, it's just fine-tuning.

This sums it up perfectly, it really does. I cannot understand why the LJ team cannot see this for themselves, I really can't.

'If it ain't broke, don't fix it', might be something of a cliche and joke, but it's also darn true.