Copyright (C) 1999 by Steve Litt. All rights reserved. Materials from guest authors copyrighted by them and licensed for perpetual use to Troubleshooting Professional Magazine. All rights reserved to the copyright holder, except for items specifically marked otherwise (certain free software source code, GNU/GPL, etc.). All material herein provided "As-Is". User assumes all risk and responsibility for any outcome.
It's important to understand that this article is in no way a criticism of Marc's work. Indeed, if Marc's work were not so absolutely vital for progress, I'd simply leave it as-is and let the reader figure it out (or not). But given Semantic Nets' potential to yield real time improvement to Mental Models, and its potential for yielding quicker, cheaper and better Era 4 Troubleshooting Tools, we dare not let Marc rest. He must refine this material until Joe Average can use it in day to day troubleshooting.
|
|
Likewise, I have no doubt that Marc can use Semantic Nets to create and refine Mental Models. I'll bet he can author great content for Era 4 tools using Semantic Nets. And I find it likely that Marc will produce an Era 4 tool in the near future. But Marc's explanation of Semantic Net construction and use is the tip of the iceberg -- nowhere near sufficient for most of us to reproduce his results.
The hope of the remainder of this article to spur Marc to become aware what he's really doing, thereby allowing him to document it. And hopefully in less than the 13 years it took me to adequately document my Troubleshooting methodology_;-). So I'll be pointing out some missing content -- missing from the article but definitely present in Marc's head.
I tried making a few Semantic Nets for myself. The first thing I noticed
is that the way *I* think, I find it much easier to start with the main
topic (web application), then list all the verbs, then other components
as objects of the sentence. Components I find in this way are then analysed
verbs first. For instance:
|
Probably others find it easier to list components first like Marc does. Some might find it easier to go broad (all headings at each level first), while others might find it easier to go deep (Web Application->accessed through->the Web browser->communicates using->HTTP *before* Web Application->runs on). Some might prefer combinations of the preceding two sentences. The process for creating a Semantic Net should be a cookbook -- probably 100 pages would do it. Once that's done the common man can create Semantic Nets. Given what they can do, it's important enough to make the cookbook.
Marc needs to explain how to translate his Semantic Net derived knowledge into a high quality Era 4 tool script, given Era 4's exploitation of General Maintenance and probabilities. I can give him some help there, and can put him in touch with people who have built and deployed Era 4 tools.
Don't get me wrong -- there have been plenty of incrimental improvements and improvements to documentation, but as far as I know this is a new train of thought in the Troubleshooting World.
As is most new thought, Marc's article is incomplete. It has not yet fulfilled its potential to make Semantic Net usage available to the rank and file Troubleshooter. It has not yet fulfilled its potential to decrease the burdonsome cost of expert system and Era 4 authoring with the use of Semantic Nets. Fulfillment of these potentials will bring a whole new productivity to Troubleshooting.
It's for that reason that I challenge Marc to make himself aware of his Semantic Net actions, and document them in a process or cookbook fashion understood by the masses. And readers, I challenge you to help him.