Tip: Using Parse command to dynamically build and execute statements

Last edition, we included a survey asking you, our customers, about our support, our products, and where you would like us to go next.  Unfortunately, survey participation was low, and the results cannot be considered statistically significant.  However, we did get some good responses back and wish to share them with you.

 

As far as our support, in general, people were satisified.  Responses rated support as good to excellent and our help files as good.  The preferred support method is email, which is good since that is the method we prefer :-)  The newsletters were rated as good.  Of course, since people had to read the email to see and answer the survey, this might not have been a good question :-)  From the survey, we received the message that we are in general doing a good job.  Our help files could be improved, but unfortunately no concrete suggestions were given.  If you have ideas for improving the help files, please email us.  Also, noone had any requests for tips and tricks for the IECS, so we must have designed the IECS to be really easy to understand ;-)

 

As far as new development, we were encouraged to develop new components.  Most people seemed to want to see Neural Network and then Genetic Algorithm components.  The least popular new component ideas were Finite State Machines, Path Planning, and Flocking (perhaps because we offer 2 of the above free already).  As far as compatibility with other development environments (Kylix, .NET), the results were mixed.  For some people, Kylix support was very important and .NET support was not important at all.  For others, it was exactly the opposite.  C++Builder support was least important to all respondents. 

 

Finally, we were definitely encouraged to keep developing the IECS.  The fuzzy logic and Expert System Wizard in the IECS got good marks while the DB support was rated fair.  Interestingly, though, improved DB support was listed as not important.  As far as new features, you told us that the most important new features should be an Interactive Expert System Debugger and, following that, defmodule support.  Other important new features should be an XML format (though that was a divisive feature, some wanted it very much, some were lukewarm to the idea), improved CLIPS support, and an Expert System Application Wizard.  The least important new features should be more functions, improved DB support and Editor frames.

 

Again, the desire for Kylix/C++Builder/.NET compatibility was mixed, with C++Builder support a distant third.  Happily, all respondents said they would recommend the IECS suite.  Though, since the number of respondents was not statistically significant we will have to keep trying harder.

 

So what does the survey mean to us?  Unfortunately, due to the limited number of respondents, there was no clear overwhelming message.  Also, this survey doesn't give us information about what people who are not customers want.  We want to balance what our customers want, what can attract new customers, and what represents the best return on development effort investment.  We are still working out our strategy and no decisions are final at this time.  As we investigate the effort further, the balance will be changing.

 

There were two messages we did receive: develop new AI components and improve the IECS.  New AI components that we should concentrate on are neural networks and genetic algorithms.  The others have a lot less priority unless they can be added with little effort.  For IECS v3.0, we should definitely not bother with editor frames, more functions (except by accretion :-) ) and more DB support.  Favorite features were for a debugger and then defmodule support.  As far as environment compatibility, we are going to take a wait and see attitude.  We'll investigate the difficulties of porting with Kylix and .NET.  .NET support is particularly problematical considering it is not out yet and all the unsafe warnings the IECS gets about pointers in Delphi 7.

 

We hope this article was useful to you.  We would dearly love to hear any comments or suggestions.  Please send any email to support@RiverSoftAVG.com