Date: Thu, 22 Jun 2000 20:33:52 -0700 (PDT)
| From: | Eugene Eric Kim |
|
eekim@eekim.com Reply-To: unrev-II@egroups.com |
| To: | unrev-ii@egroups.com |
| Subject: | Augment + categories = OHS v0.1 |
Based on the discussions we had at our meeting today and some of the discussions afterwards with Eric, Howard and others, I came up with the following conclusions.
Fundamentally, our system consists of two things: nodes and links. The difference between our system and the previous two revisions of Doug's work is that our system allows us to categorize nodes. Categorizing nodes is crucial, because it provides an additional semantic layer that is crucial for knowledge management. Both the users of the system and the system itself can take advantage of this additional layer.
For example, in order to adapt our system to accomodate IBIS-style discussions, we add categories such as "question," "answer," and "alternatives." With this information, we can search for all the questions pertaining to a particular topic of discussion. Or, we can write a system module that automatically compiles and presents an IBIS view based on the current question and answer nodes.
Categorizing nodes allows us to provide some structure to discussion, even if the discussion itself is unstructured. For example, if we're trying to come up with a Use Cases document for a piece of software, I may propose five different Use Cases in three different e-mails. However, if these nodes are properly categorized in each node, then I can easily create a Use Cases view that shows all of the Use Cases in one view, regardless of when or where these were proposed.
I think there's still one open question that needs to be resolved for version 0.1 of the OHS. That question is, are links categorizable? This relates to some of Eric's previous questions regarding relationships. In my opinion, a link is the way you represent a relationship. The question is, should you be able to specify categories for a link.
My suspicion is yes. Even if this doesn't immediately affect system behavior, I think it is a useful attribute to have in the data structure. In Gil's Knoware, which he demonstrated last week, you can label relationships, but those labels have no meaning. However, it should be fairly easy to give those labels meaning in the system in the future, which is one of the things Gil said he planned on doing. I see no reason why we shouldn't take the same approach.
-Eugene
Eugene Eric Kim
eekim@eekim.com
http://www.eekim.com/
"Writer's block is a fancy term made up by whiners so they
can have an excuse to drink alcohol." --Steve Martin