Re: Off-Topic: Communicating with users

FROM : Scott Ellsworth
DATE : Fri Jun 23 23:45:12 2006

On Jun 23, 2006, at 1:24 PM, Bobby B wrote:

> It all takes place on a public forum, so all responses are visible.  I
> was curious how you, as developers, would respond to users'
> suggestions when they don't fit in with the scope of the program, just
> aren't good, or are just plain horrible.


When designing a program, I try to be clear to users that I make the 
final call about features, price, etc.  That said, I also try to be 
clear that I want to know how they are using it, what they are using 
instead, and what would make it more useful to them, at relatively 
low cost to me.  If a suggestion does not fit my vision of the 
product, I try to figure out what need drove that suggestion, and 
then will suggest product niches where the suggestion will fit.

For example:  I want to put together a symbolic and computational 
math package, much like Mathematica or Maple, but targeted more for 
people working outside academia, as well as high school and college 
students.  Both Mathematica and Matlab cost over $3k a license, and 
thus only academics and the wealthy get the benefits.  I believe that 
just about anyone could get benefits out of having a symbolic math 
calculator.

If someone suggested tensor analysis as an appropriate first place to 
start, it would not fit my intended target audience.  It is not a bad 
suggestion, and not a bad idea for the n_th revision, but it is a 
terrible idea as a starting place for something usable by a typical 
knowledge worker who wants to do a bit of automated crank work.

Thus, that suggestion gets a clear description of the intended 
audience, and a suggestion that they write up a few more details for 
me to think about later.

Someone who suggested audio and video file processing, on the other 
hand, is doing something completely out of the scope of my idea.  A 
bad suggestion, in other words.  Again, I would want to probe a bit 
to see if they had a higher level purpose - 'I want to understand 
what a convolution kernel does in core graphics'.  Ok, suddenly not 
such a bad idea, even if still not something I want to do.

The same person saying 'I want a photoshop clone', though, is asking 
for something I do not want to deliver.

I am still not going to say 'dumb idea' but I will suggest Core Image 
Fun House, or real photoshop.

In summary, try to figure out the overall task they want to perform. 
If it fits your goal, even if it is not on the to-do list just yet, 
then file it away.  If it is completely alien to your goal, then 
direct them to a better package, after explaining why it feels 
foreign to your goal.

Scott