I have no idea where a lot of the stuff on my computer actually is (a little like the condition of my office or shop). If I want it, I remember some part of the name of the file, or what kind of file it was (pdf or jpeg or whatever) or maybe some of the words in it, and I tell my computer to go look for it. While the computer is looking, I do something else. Maybe when I find what I'm looking for I put it in a place that seems better suited, but that doesn't mean I'll remember that location either. Why should I?
The same thing is true for information on the web: why bookmark a page when you can just remember the search terms that will get you there (or maybe somewhere just as good). It's not about people, places or things any more, it's about the actions you can order machines to perform that will make those people, places or things accessible to you.
Of course, this is really nothing new in principle -- people have had secretaries and other servants to go get things for them for centuries -- but as far as computers go this is a change. (Or maybe it's just a sign of how messy my desk and 15 years worth of assorted useless information stored on disk have become.)
One of the things that makes this work for me is what I'm going to call reliable lag. I know it's going to take half a minute to a minute for my machine to finish any given search -- roughly the same amount of time to download the first half-dozen likely candidates in a search -- so I can go off and do something else (finishin up some email, looking at another web page, exchanging kitten pictures with Julie) secure in the knowledge that when I come back something useful will be waiting for me. Running the search in the background doesn't perceptibly slow whatever I'm doing in the foreground, nor do my foreground actions appear to hinder the machine from carrying out its background tasks.
There's a long literature on human-computer interaction that talks about the importance of stable, sub-second responses to typing or mouse clicks. That instant feedback is crucial for developing a sense of rhythm in using a computer -- just try using a text editor at the other end of a jerky network connnection. But I think there's another class of response times, on the scale of 30 seconds to a minute or more, that are equally useful in making comfortable interactions with a computer. Once people have come to trust that they don't have to stare at a particular window on their screen waiting for the next thing to happen (just as most of us have come to trust that we don't have to keep pressing the elevator or the "walk" button, and that those conveyances will let us know when they're ready) there's a huge range of additional work computers can do behind the scenes. Imagine web searches that really took your preferences seriously, or email readers that were willing to spend an extra five minutes really filtering out your spam -- or whatever else I can't even imagine on the spur of the moment. The algorithms are there; it's the working styles that have to change.
Posted by wallich at March 30, 2004 08:01 PMThis is why I am a huge fan of distributed computing. The human mind precieves lag as soon as 300ms, and for a computer, this is a HUGE gap of time. It is not supprising that I am spoiled by my home computer, with all it's customizations and configurations for how I do things. When I use my mom's computer, I am constantly tapping my fingers, waiting for it to finish what I know my machine could do so much faster.
I also think that it is interesing as to how you have your computer set up. I have an extreemly hirearichal setup. I can find anything I need by following my own logical path. I don't even have to search for things on my own computer because I always know where everything is. The same thing goes on in my room though, so maybe it is just me.
Posted by: Tim at March 31, 2004 01:08 PM