RE: Proposal for an Emacs User Survey

[Drew Adams]

> > If we want to learn what users think, we should not limit their
> > responses to a small set of 'valid" possible answers.  The plan
> > I designed for inquiries asks users to answer in their own words.
> 
> But wouldn't that make it needlessly hard to analyse the results,

Ease in handling the results shouldn't be a major
consideration.  Quality of the feedback and quality
of its analysis is more important that speed or ease.

> especially if the question should be numerically quantified?

Quantification of bad-quality information doesn't
help.  Better to get a high-quality understanding,
and if some quantification _then_ helps then do it.

> I get the value of plain text responses, but recognizing that the answer to 
> the
> question "how long have you been using emacs", could result in:
> 
> - "Since 1996"
> - "24 Years"
> - "January 1996"
> - "Around the second time Clinton got sworn into office"
> - "1896" (but actually a typo)

A very specific question like that can be asked in
specific terms: how many years.

But even that is likely to be misleading, if kept
apart from considerations of kind and degree of use.

Getting a good picture of Emacs users and uses is not
easy, nor should it be.  And again, users _reasons_
for actions, preferences, etc. are important.