jack@ztivax.UUCP (Jack Schiff) (02/28/89)
As one input to an overview of knowledge engineering methodologies, I am
trying to collect opinions on potential benefits of such methodologies
and specific areas for their use.
The attached short questionnaire was designed to take not more than
10-15 minutes for completion, since my main concern here is broad-based
orientation. Consequently, the questions below are much less detailed
than would be needed for a formal survey. Still, I hope they're meaningful
enough to convey an idea of what I'm after, and any additional comments
are highly appreciated.
Thanks a lot for your help!
---------------------------------------------------------------------------
1. Background
Experience in little some extensive
- building expert systems ____ ____ ____
- building conventional systems ____ ____ ____
- using software engineering techniques ____ ____ ____
(which?: __________________________ )
- project management ____ ____ ____
Size of system(s) built small medium large
- expert system(s) ____ ____ ____
- conventional system(s) ____ ____ ____
2. Opinion on general issues in building expert systems
Importance
Issue | low medium high
----------------------------------------------------------------------
Effectiveness | ____ ____ ____
(targeting expert systems at the right |
problems) |
Efficiency |
shorter development times | ____ ____ ____
higher reliability of project schedules | ____ ____ ____
better coordination of multiple knowledge | ____ ____ ____
engineers in large projects |
Quality |
higher correctness of expert system results| ____ ____ ____
better user interfaces | ____ ____ ____
better maintainability | ____ ____ ____
Other |
__________________________________________ | ____ ____ ____
__________________________________________ | ____ ____ ____
3. Methods and tools for specific knowledge engineering activities
"Method" here refers to some systematic approach used in a specific
activity, e.g. repertory grid elicitation in knowledge elicitation.
"Tool" here refers to a computer program which supports one or several
methods, e.g. Shaw's & Gaines' KITTEN for repertory grid elicitation
and analysis.
Please indicate your satisfaction with your current practices and your
desire for future methods/tools as l(ow), m(edium) or h(igh), as in the
following example:
| Satisfaction | Desire
Activity | ad-hoc method tool | method tool
------------------------------------------------------------------------
Knowledge elicitation | l l | h m
This would mean that currently some mixture of ad-hoc and methodical
knowledge elicitation is used with low satisfaction, an improved method
is highly desired, and a tool would be welcome.
| Satisfaction | Desire
Activity | ad-hoc method tool | method tool
------------------------------------------------------------------------
Application selection | ____ ____ ____ | ____ ____
Requirements analysis | ____ ____ ____ | ____ ____
Knowledge elicitation from | ____ ____ ____ | ____ ____
experts | |
Knowledge acquisition from | ____ ____ ____ | ____ ____
documents | |
Knowledge acquisition from | ____ ____ ____ | ____ ____
databases with example cases | |
Knowledge structuring | ____ ____ ____ | ____ ____
Selection of knowledge represen- | ____ ____ ____ | ____ ____
tation formalism | ____ ____ ____ | ____ ____
Testing | ____ ____ ____ | ____ ____
Error analysis | ____ ____ ____ | ____ ____
Coupling to other software | ____ ____ ____ | ____ ____
Maintenance, i.e. several of the | ____ ____ ____ | ____ ____
activities above performed for | |
an existing expert system by | |
other developers than original | |
Other: __________________________ | ____ ____ ____ | ____ ____
End of questionnaire
--
Jack Schiff c/o UUCP : jack@ztivax.uucp
Siemens AG, ZFE F2 INF32 or : uunet |mcvax!unido!ztivax!jack
Otto-Hahn-Ring 6 Internet: jack@ztivax.siemens.com
D-8 Munich 83 / W.-Germany --- Standard Disclaimer ---