From: Eric Armstrong <eric.armstrong@eng.sun.com>
Yeah!! What you said.
In particular, the need for <== and <=> is duly noted.
The "router" (or agent) idea to generate new documents also
makes excellent sense.
The only trick is to distinguish nodes, so the agent can find the
right ones. For example, it needs to distinguish:
* A link to a reference node
* A design note implied by a feature
* A secondary design note implied by a design note
* A test case implied by a design note
etc.
We're probably going to have to label the nodes -- or possibly
the links, in some way. In a "fixed" domain like software
development, where we know the knowledge products (documents)
we want to produce, we can probably get away with forcing the
node-originator to specify the type of node they are creating.
But in a more fluid domain, in which most of the early explorations
will be unlabled, there will be a need to add labels later, and probably
the need to recategorize entries later on.
The part I still don't have a good handle on, though, is how IBIS
style OR's of alternatives co-mingle with document-style AND's
of features, design notes, or what have you.
Jeff Miller wrote:
> From: Jeff Miller <jeffm@dynamite.com.au>
>
> Interesting idea. sounds like a reasoning tool. I'd imagine that you'd
>
> have a routing algorithm that would trace from a selected node to it's
>
> logical conclusion(s). two other operators you'd need to add though
> are
> <== and <==> as a design idea could imply a feature (bug?) by it's
> presents and a feature may imply a design idea that implies that
> feature.
>
> For example,
>
> F1 ==> D1
>
> <==>
>
> F2 ==> D2
>
>
> To include feature F1 you imply D1 which implies the presents of
> feature
> F2 and design idea D2. but having feature F2 while it imples design
> idea
> D1 and D2 does not imply the presents of F1.
>
> Back to the routing; You could conceviably select a starting point
> which
> would then be traced to one or more logical conclusions. These
> conclusions
> may not be whole as that reasoning in the automated system breaks down
>
> when there are no explicit connection between nodes. At this stage the
>
> human user steps back in to carry out the fuzzy logic jump that is
> needed
> to find a new starting point for the router to take over again. In
> this
> manner you could connect islands of design patterns to design a system
>
> without actually having to implement it. Think of it as a CAD/CAM
> simulator for software engineering.
>
> Jeff.
> -----------------------------------------------------------------------
> [ONElist Sponsor]
> Please click above to support our sponsor
> -----------------------------------------------------------------------
> Community email addresses:
> Post message: unrev-II@onelist.com
> Subscribe: unrev-II-subscribe@onelist.com
> Unsubscribe: unrev-II-unsubscribe@onelist.com
> List owner: unrev-II-owner@onelist.com
>
> Shortcut URL to this page:
> http://www.onelist.com/community/unrev-II
--------------------------- ONElist Sponsor ----------------------------
GET A NEXTCARD VISA, in 30 seconds. Get rates as low as 2.9 percent
Intro or 9.9 percent Fixed APR and no hidden fees. Apply NOW.
<a href=" http://clickme.onelist.com/ad/NextcardCreative4CI ">Click Here</a>
------------------------------------------------------------------------
Community email addresses:
Post message: unrev-II@onelist.com
Subscribe: unrev-II-subscribe@onelist.com
Unsubscribe: unrev-II-unsubscribe@onelist.com
List owner: unrev-II-owner@onelist.com
Shortcut URL to this page:
http://www.onelist.com/community/unrev-II
This archive was generated by hypermail 2.0.0 : Tue Aug 21 2001 - 18:56:46 PDT