yak
[Top] [All Lists]

[yak@collab] Re: Personal knowledge mapping

To: yak@xxxxxxxxxxxxxxxxxxx
From: Eric Armstrong <Eric.Armstrong@xxxxxxx>
Date: Tue, 02 Dec 2003 12:54:09 -0800
Message-id: <3FCCFBF1.90903@sun.com>
Ah. Thanks for the thoughtful explanation.    (01)

John Sechrest wrote:    (02)

> Perhaps if you think of it as providing many different ways
> to view relationships between the data.
> 
> Instead of providing you with specific data and specific
> relationships, I provide you with a means to query and
> show an abitrarily filtered set of relationships.
> 
> For example, I am trying to buy glass for glass blowing.
> Right now, I can crawl thru a bunch of online catalogs. 
> And it takes me a long time to find who is doing what.
> Mostly because I am still learning about the relationships 
> between the different types of glass.
> 
> What I really want to know is a cross vendor/sales set of 
> information.
> 
> If each of the catalogs provided access to the data that they
> already have, in a way that I could easily manipulate,
> then I could have a higher level program query all
> of the vendors, showing all the types of glass available.
> And come up with a cross vendor comparision.
> 
> Instead, each one provides a different view of the same
> types of data in subtly different ways, such that I can not
> find what I am looking for without crawling thru many 
> different pages each time I am trying to buy glass.
> 
> This happens because:
> 
>      1) There is an assumed view of the data
>      2) There are assumed relationships between the data they have
>       which are not presented
>      3) There is assumed knowledge about the data that is not
>      available in the data.
> 
> (IE, what types of glass are compatable with what other types of glass)
> 
> So this article is talking about abstracting out the access to 
> the underlaying relationships. Which then allows you to slice
> different views of the data. 
> 
> This is the exact same question we were talking about when we
> talked about graphical views of knowledge. How do I filter,
> organize and present a sub set of information to illustrate
> a set of relationships between bits of data.
> 
> So they are saying instead of giving  you one view of the data, 
> provide more raw access to the data, so you can make many different types
> of queries across it.
> 
> this is a step back towards allowing programming to be available
> to the end user as a way to explore information.
> 
> If the data is available, I can then write a query/program to 
> organize it based on my goals.
> 
> 
> 
> Eric Armstrong <Eric.Armstrong@Sun.COM> writes:
> 
>  % > OLDaily November 28, 2003
>  % > by Stephen Downes
>  % > 
>  % > - -----------------------------------------------------------
>  % > Personal Knowledge Mapping And The Concept Of Data
>  % > Emergence
>  % > 
>  % > ..."Content providers should not be
>  % > trying to guess how I want to interact with their
>  % > information. They should just be providing the information.
>  % > I will customize my experience as I see fit."
>  %  >
>  % Makes no sense whatever, to me.
>  % 
>  % "Knowledge" cannot be abstracted from the notion of "sequence".
>  % A series of data points only becomes "knowledge" when they
>  % are grouped, categorized, and, for any complex system, sequenced
>  % in presentation so as to *communicate* the knowledge.
>  % 
>  % An unnconnected collection of data feeds in which I make all
>  % the connections is a lot like looking at the catalog of classes
>  % in a system -- without a class hierarchy, and with no interaction
>  % diagrams. It's a lot like reading source code, actually -- the
>  % mystery of how things connect is only discovered after an
>  % arduous period of search and discovery.
>  % 
>  % A "content" provider that isn't providing *some* kind of
>  % organized view of the data has no value add. Certainly,
>  % there many ways to view the same things -- so the idea of
>  % creating and sharing alternative views makes a lot of sense.
>  % 
>  % But the idea of publishing the information in the first place
>  % without using such a tool (presuming it exists) to create at
>  % least one organizational framework makes absolutely zero sense.
>  % 
>  % 
>  % 
>  % -- 
>  % This message is archived at:
>  % 
>  % 
>http://collab.blueoxen.net/forums/cgi-bin/mesg.cgi?a=yak&i=3FCC0A25.1060106@sun.com
> 
> -----
> John Sechrest          .         Helping people use
>                         .           computers and the Internet
>                           .            more effectively
>                              .                      
>                                  .       Internet: sechrest@peak.org
>                                       .   
>                                               . http://www.peak.org/~sechrest
>     (03)


-- 
This message is archived at:    (04)

http://collab.blueoxen.net/forums/cgi-bin/mesg.cgi?a=yak&i=3FCCFBF1.90903@sun.com    (05)
<Prev in Thread] Current Thread [Next in Thread>