Wednesday, April 4, 2007
I didn't get the whole web 2.0 thing in time to be cool about it. Luckily I'm not a guru who led my consultancy down the tubes as a result. I only really got it around the time I first saw this movie summarizing it rather cutely. I still find it easier to describe most things in the negative, what they are not, than what they are. Call it a character flaw, but it works for me. In that vein, IMDb is by far the least web 2.0 site I have used lately. Sure, lots of others are worse in absolute terms, but it suffers so much due to what it is. A huge relational data store (its got "Db" in the address!). The one, best one in the world for its data type. Traditionally, I have hated it for not being mobile enabled, for being bloated, for generally poor UI. I especially like the lack of deep-linking after signon. Oh, and the periodic intrusive advertising. And, none of this improved after their recent purportedly well researched redesign. But the big ah-ha moment came the other day. I add stuff to IMDb occasionally. Its not an encyclopedia with enough permanent staff to update it all, but more of a wiki, so needs user input. Of course, woe be unto you who think it should /act/ like a wiki. Lets walk thru it. I finish Alphaville and am wandering the page looking at interesting bits. A friend recalls a reference to it, which is not listed, and I remember one that I am sure about, so want to add. I remember in this one Homicide (life on the street!) episode there was a villain who used the screen name of Lemmy Caution. Bayliss even mentioned that it was from Alphaville. Great. Good reference that should be in there. So, lets update the Alphaville page. Um... how? Seriously, this took me five minutes to find. You cannot call users dumb, but as a designer, I'll admit I am dumb, so maybe its my fault. Anyway, look for yourself. Its the "update" button at the bottom of the page. Okay, so you click it. And... well, its not worth giving them crap about every bit of the interface. Its not clear right away (the most obvious part to me is the "forgot something" function, which I STILL do not understand) but eventually you see the pulldown. Its also weird, but you get there. And its reasonably contextual. At least its about the "movie links" page as opposed to having you drill thru everything. So, you get to the ability to add items by line. Its a bad sign that it needs to have four paragraphs of text in front of it. Probably hard to use. And way too many categories, but my issue is around the "other title" field. So, I want to add a link to Homicide. I go to the correct page for it, and add the title straight out of their search results ("Homicide: Life on the Street" (1993)) and get this: Hmm. Okay. Its not obvious there even /are/ episode pages to me before this. For example, I went to another site to confirm the episode name, because its so unclear. So, we go to find that. Which is hard. A search for the episode title leads nowhere. Eventually I see the "seasons" links at the top of the series page leads me to a huge list of all episodes in all seasons. I find the right one, and it even has a link. Click it and what do ya know, there's a whole page about that episode. Not super-clear that there's a superordinate series or anything, but there it is. So, I grab that title removing the carriage return, so it looks like ""Homicide: Life on the Street" Homicide.com (1999)" and try it. I even remember to remove the episode info from the description. Not even close: In fact, we're getting further away. Now it doesn't even know what that means! I sigh. I watch more TV. I eventually try again. Maybe its web-centric. Try the URI "http://www.imdb.com/title/tt0604379/" Nope. Even worse (I won't show any more screenshots). Okay, maybe its all using codes. How about the value for the title "tt0604379." Again, not a finger. So I think, maybe they are as bad as some of the developers I have to work with, and stare blankly when you say "parse out irrelevant characters." Because its a title field, maybe it never occurred to them to accept anything but numbers. Yup, that's it. It works. I submit it. But by the time you are entering secret code numbers, not available anywhere ON the page, we might as well be using command line interfaces. Now that I know, its not that bad at all (like command line interfaces, really). But who would be able to get thru this process otherwise? I only finished out of a sense of self-sacrifice as an interface designer. I feel the post needs to end in some great, or at least witty, big conclusion. But there's not much to say. Its bad, in a way that's so bad it points to a cultural issue. And its easy to blame IMDb but my company has dozens of customer-facing web applications at least this obtuse. I suspect I'll write about it more, but clearly its beyond the scope of just designers. There needs to be a cultural change for developers, and business stakeholders. I'm just not sure when that might filter down to this level.