The KnownSpace Datamanager
"Halfway To Anywhere"


The Personal Computer Revolution Isn't Over

The personal computer revolution is not over. You can now buy a computer that is customized for you: a store will assemble a computer with the processor you want, the memory you want, and the peripherals you want. But you cannot buy software that is customized for you. Almost all software, commercial or otherwise, is written to address the largest possible market. Software products are churned out like soap bars or lawn mowers---every copy is the same: my copy, your copy, it doesn't matter. My software does not adapt to me the way that my easy chair adapts to me, or as my shoes adapt to me. Personal computers cannot be truly personal until that happens. We'll know the revolution is over when you can't use my computer and I can't use yours.

Take most information retrieval software: it pays no attention to who you are. It will treat your queries exactly the same way that it treats mine. Which would be fine if information retrieval was being done in a library by many patrons with a skilled librarian at your elbow helping you through the tricky bits when you get stuck. If the users are constantly changing, and they all want different things, then it wouldn't make sense for the program to change to suit what the last user wanted since the new user will likely want something else. The librarian too, who is supposed to help the hapless users deal with the difficult software, would be in trouble if the software modified itself from instant to instant.

Unfortunately, this model of reality is obsolete. Most modern information retrieval is no longer being done in libraries. It's being done on the web and on the desktop. It's not being handled by skilled full-time librarians, it's being handled by innocent users aided only by browsers, search engines, and cataloging companies.

To make information retrieval software personal we must exploit the massive amounts of information that users provide about themselves simply by using the system normally. My information retrieval software should change to fit me the way my comfortable old slippers have changed to fit me. Every step I take wearing my slippers softens the places where I put pressure, and makes them mold better to my feet. Similarly, everything I do with my data should cause small changes to my data retrieval software so that my computer suits me better.

Further, I don't just want a program to handle data retrieval, then have to get another completely separate one to handle my email. First, the two tasks are very similar, and second, the two tasks often share a lot of information. Someone sends me an email that triggers an information retrieval request using my browser. I shouldn't have to manually link those two events myself.

Web browsing and email handling are only two of many such data management tasks that most of us have to deal with every day. The major software corporations give us software to handle some of these tasks, but in piecemeal fashion. There is little uniformity and interoperability. And even when there is, the software itself is set in concrete. It's not possible to get the source code, never mind modify it to suit our needs. This is ridiculous. Until our computers adapt to us they cannot be truly personal.

Today everyone is Odysseus---which is kinda fun actually, since I like to go exploring every now again. The only problem is it's... tiring. Often, I just want to sit in my comfy easy chair and have things brought to me. I want to be Penelope.