A Web for Augmentation, not Automation
June 09, 2004
Earl Mardle comments on Russell Beale's How the Web is Changing: (8QZ)
It fits with my contention that the net is a tool, not a medium. The whole net is an information tool that does a whole lot of interesting things, most importantly, distributing information power; that is the power to produce and disseminate your own information and access, critique and publicise the critique of other people's information, but it also acts as a validator of information and perspective by community, however small or isolated. ps (8R0)
I was going to write this as a comment on his blog and then decided to move it here. (8R1)
In his posting and some others from around the same time, Earl seems to be bouncing around the difference between tools that augment and tools that automate. (8R2)
Tools that augment extend a human's ability while still leaving the human in control. They are often small things (like purple number stuff) that provide a way to grab or manipulate stuff of all sorts. (8R3)
Tools that automate "do it for you", often operate in large swaths, and are based on performing tasks that can be formally described. (8R4)
There's an ethical or world view difference between the two. Some people prefer the latter, some people the former. I'm deep in the augmentation style, I hope. The Semantic Web, as imagined by the W3C, strikes me as in alignment with the latter. (8R5)
Being in one camp or the other doesn't make someone wrong, but it does make conversation across the boundaries of the camps a little confusing and disorienting. (8R6)
I first came upon the augmentation/automation split while writing my Computer as Tool paper: (8R7)
Landauer distinguishes between two phases of computer applications. Phase one applications automate tasks "replacing humans" for the performance of "almost any process that science, engineering, and statistics have captured in their theories". Phase two applications, on the other hand, are applications that assist humans in tasks for which there is no established theory of action. Phase two applications include the very large body of office productivity applications, web browsers, and desktop operating systems; anything where the human uses the computer throughout the process. They are the applications we use to process information in flexible and potentially undefined ways. T (8R8)
Landauer's book: (8R9)
is a good read. (8RB)
Comments
Is there really an ethical or world view difference, or is there more often a difference in applicability, problem domain, and what the computer can actually do a good job of? Consider what Gelertner would like email programs to do versus what they can actually do. Currently, augmentation is the best we can hope for. Ideally, [ideal] automation would be very nice. (8RC)
Can't a fella like Purple Numbers _and_ McFeely?? (8RD)
I guess I wasn't quite clear. Yes, you can be a fan of both. The problems comes in when one world view is attempting to horn in on the domain of the other. (8RH)
For example, in the semantic web world the agent-that-shops-for-you scenario is often used. The agent is supposed to automate the process of the purchase. The automation zealot assumes you can effectively model a purchase decision in a formal way, repeatable way, and that the challenge in the tool is how to most effectively represent all the variables involved in some kind of machine-usable ontology: see RDF, OWL and friends. (8RI)
Someone at the exteme end of the augmentation camp scofs at such notions and thinks a purchasing decision involves far too much that is unknown or at least difficult to model and that the agent style of the semantic web is folly. (8RJ)
A more laid back augmentation person would suggest that it is best to determine which parts of the purchasing process can be formally modelled in a straightforward fashion, automate those, and provide handles to those tools to augment a human making the purchase. (8RK)
McFeely? nailed this particular way of doing things rather nicely, I thought. Those things easy to automate, were automated. (8RL)
And it made life tons better. (8RM)
Okay. With your example as context, I can understand who these automation zealots are, and I pretty much agree with you. I suspect I'm mostly a laid-back augmentation person. You know, I like to automate the automatable. But when you can't automate, you can often augment. That's why I have a personal wiki. (8RO)
To me, augmentation is sort of like making a really elaborate wrench that lets you easily get a handle on the problem. It's tool-building. Good tools make your life better. (8RP)
This reminds me of the Bell phone story where it took some time to figure out that the phones was about what people can do through it, and not what the phone can do for the people. (8RS)
I read it in the "The Hacker Crackdown" by Bruce Sterling: (8RT)
http://purpleslurple.net/ps.php?theurl=http://www.farcaster.com/sterling/part1.htm#purp272 (PurpleSlurple rules, I could never point you to the right paragraph otherwise) (8RU)
The worst case is when applications try to automate peoples; a system made only by rules, wizards and `next' buttons is designed to automate the user to complete a process. (8RV)
What people enjoy is to be able to use the application, to manipulate the application, and not let the application manipulate them; nicely described in the Naked Objects: (8RW)
http://nakedobjects.org/section12.html#empowerment_benefit (8RX)