Log in

No account? Create an account

Wed, Jan. 31st, 2007, 12:36 pm
Ça plane pour moi ...

I think I'm fairly unusual here at LJ towers in that I wasn't plucked from the ranks of habitual LJ volunteers or long term users to come work full time. So in a way I have a sort of unique perspective on the interaction between us and our users - sometimes I kind of feel like the internet's Jane Goodall.

Watching the responses to my first news post was a nigh-visceral experience. Friends of mine who've written books have said that they had a similar feelings watching the sales rank and comments on Amazon. It was kind of shocking to see how fast people responded. burr86 reminded me to turn off email notification and I said I'd do it when I came back from making some tea. "Now", he urged, "You need to do it now."

By the time I came back from tea there was already 3 pages of comments.

People CARE.

Because most of my coworkers came from the community I don't think they make the distinction between them and us that I think the community at large does. They get affected. Some of the comments users make hurt them. Deeply. Because I'm often on a different timezone to everyone else I've sat on the end of IM with people who can't sleep out of distress.

This is the flipside of The Cluetrain Manifesto that nobody talked about.

Part of the problem, I think, is that, no matter how hard you try, you can always communicate better. There's always going to be a filter between us and the users. And there's always going to be people who fear the worst or who willfully interpret every action as being done out of malice or a very vocal few who take their own personal views and ascribe it to everyone.

And it's not just LJ. It's everywhere. I have a several friends who work at Flickr - all long term Flickr users to a (wo)man. Today they're dealing with this thread about the deadline to finally move to a combined Yahoo! ID. Like us they've been dealing with accusations of selling out to a larger company (ignoring the obvious that, without the larger company, both of us would have probably ended up shutting down due to overwhelmed resources - technical and personal) and like us with, say, editing comments there are features that users are clamouring for (sets-of-sets is one, of the top of my head, for them) which the users can't understand why the team has not given them.

On the one hand I'm baffled by some of the attitudes - logically we must either have a good reason, be lazy, be incompetent or be maliciously withholding a feature because we don't want the users to have it. For some unknown and inscrutable purpose.

And some users clearly don't think that it's the first and I think our track record (hopefully) shows that it's not the second or third so they must think it's the fourth. Which just seems ... odd. Do we really come across as malicious?

On the other hand I've been on the other side and so I can understand. I can really understand. I'm a geek and a user and a customer and I'm passionate about the things that matter to me. And, amonst many other thing, LJ matters to me. I use it everyday. We use it for work. I feel a burning urge to make it completely awesome. I get defensive about it with other people. When I'm back in England and I'm talking to my (largely LJ using, nay OBSESSED) friends my eyes shine when I talk about it. I really want to finish search now because I've got a really, REALLY cool idea I want to prototype and get signed off which I hope will completely rock everyone's world.

I have no conclusion apart from I wish I had a magic way to make the users trust 'us'.

Not that I think it needs saying but, naturally, these are my views, not necessarily the views of LiveJournal or SixApart.

Thu, Feb. 1st, 2007 11:30 am (UTC)

Sadly I don't think there is a way to make all your users trust you. When you've got the size and variety of userbase that LJ has, whatever you do is going to do piss someone off. All you can do (I mean you as the LJ hivemind) is give people fair warning that something is going to happen and pay attention when they complain.

From my own perspective, the thing that currently annoys me most about LJ is that very little apparent effort is going into making the basic functionality work properly. The proportion of comment posts that go awry, or poll answer submissions that don't work is getting ridiculous, and nothing seems to be being done about it. All we ever hear about via news is exciting new feature X, and when X = getting a bunch of flowers on yr userinfo, it rankles all the more when my next comment disappears into the ether. Just telling us what you're doing to fix stuff would make an awful lot of us a lot happier.

Oh dear. This comment seems to have turned into an instance of what you were talking about. I'll shut up now.

Thu, Feb. 1st, 2007 06:47 pm (UTC)

So there's two points there - the giving people a warning and whether to work on new features (trivial or otherwise) or work on scaleablity.

So the warning thing. Fair cop. Our only defense is that announcing new features before they've been finalised can lead to overpromising but really, yeah, we can do better.

So, the second thing - the scaleability problems. It's unfair to say we're not working on it. We work on it every day. We've had to invent new technologies (mogilefs, Perlbal, memcached) just to keep up which we then release for free and are used by other large sites. We had to invent our own mail solution to deal with the amount of outbound messages we deal with. It's not so much that we're not working on it it's more that every time we give ourselves some breathing space we get more users and more traffic and ...

You say basic functionality, we say a huge cluster of DB and Web machines, massive amounts of bandwidth and working on the cutting edge of what some of the technologies we use can do. If we weren't working hard on scaling then LJ would have fallen over a long time ago.

You've also got to remember that we're small. From the outside it may seem like we're some huge conglomerate with battery pens of programmers slavishly working away. But if you look at the holiday post picture then that's it. That's LJ. And two of those people have moved ointo bigger and better things. And technically I don't even work for LJ any more - I work for the internal RnD group. So that leaves, err, 5 engineers.

Things like vgifts - they're a bit of fun and, in the grand scheme of things they don't take very long and they let the team do something a bit silly which stops them going insane. Flickr got the same thing with the whole "Ho Ho Ho Hat" malarkey over Christmas.

One complaint we get is over LJTalk. What people forget is that brad and crucially wrote DJabberd in their spare time. Integrating it into LJ was really trivial (basically an authentication and roster plugin which used existing LJ apis) and, of course, for the people who don't want to use it then they never even have to see it. So for us it rankles a little when people complain about it.

However, despite all that, you're right (in a way at least) in that you have every right to get knarked when you post a comment and it just vanishes. We can only promise to try (even) harder (THE BEATINGS WILL STOP WHEN MORALE IMPROVES!).

Thu, Feb. 1st, 2007 10:09 pm (UTC)

Ah, man, reading my comment back, I didn't emphasise that 'apparent' enough. I really wasn't trying to accuse you guys of not working on scalability (tho it does read like that, sorry) - my main point was that this work goes unannounced and so is invisible to most users. And that this lack of visibility then leads to people assuming that all the dev time is being spent on Rad! New! Features! rather than improving the services that they paid for. Which is an untrue and unfair assumption, but given the relative lack of info, an understandable one.

Thu, Feb. 1st, 2007 11:57 pm (UTC)

Ah, yes. Re-reading I get what you're saying.

Actually, that's kind of a tricky question - should we post to news and say "So, we squeezed another 3% speedup from the DBs this month. Not that you'll notice but, hey"?

That seems like it'd get real old, real fast but on the other hand then at least there'd be a sense we were doing something.

I shall have a ponder.

Fri, Feb. 2nd, 2007 03:34 pm (UTC)

FWIW I think this would be a Good Thing. A monthly roundup of the dull-&-boring (but useful & practical!) changes would I think make it clearer that stuff *is* going on behind the scenes, and would increase tolerance. Maybe also an update on just how *difficult* the whole thing is ("the servers are 3% faster but we've also had a 5% increase in post volume" or whatever), as I don't think people realise that either. I certainly wasn't aware (although I guess it would have occurred to me if I'd thought about it for any length of time) that LJ is running so close to the technological edge. Making that more obviously public might generate a bit more goodwill as well.

It seems to me a bit like train delays: if one is stuck just outside of Middlesborough or whatever then obviously that's going to suck whatever. But it does IME suck a lot *less* when you have the driver explaining what's going on (even when they're saying "I'm not quite sure what the problem is, but I'll keep you updated" or "There's a lorry stuck under a bridge and we're waiting for the engineer to haul ass over here & I'm sorry but I have no idea how long it'll take - but I'll let you know when I hear more!" (why yes, that *did* happen to me) ) than when there's just radio silence.

Mon, Feb. 5th, 2007 04:19 pm (UTC)

Maybe this kind of dull stuff wouldn't be any use to the average LJ user. But to me? I'd be interested. As would all of the other people who spend all their time working on scaling sites would to.

Wed, Feb. 14th, 2007 11:56 pm (UTC)

Actually, yeah, being a bit more open (current buzzeword: transparent) tells people where the effort is going. It educates us about the problem and gives us a feeling for how difficult these giganto-scale issues are. Context is powerful!

Thu, Feb. 15th, 2007 03:32 am (UTC)

What people forget is that [info]brad and [info]crucially wrote DJabberd in their spare time

The thing is? I follow a lot of news and meta LJ discussion comms. That's the first I heard it was a spare time project. If that had been said, then it would probably have come across so much better, because Brad's known for doing that sort of thing.

Ah well, water under the bridge.

Thu, Feb. 15th, 2007 07:41 pm (UTC)

It didn't go in the official LJ news and meta discussions because, well, it wasn't an official LJ thing ;)

Most of the info was on brad's journal - the initial announcement was here I think and you can follow the development using the 'djabberd' tag

But yeah, when we announced it we probably should have mentioned that it was something they'd done in their spare time.