Subject: Re: Support as insurance
From: Ian Lance Taylor <>
Date: 1 Dec 1999 17:16:07 -0500

   From: "Stephen J. Turnbull" <>
   Date: Wed, 1 Dec 1999 18:17:10 +0900 (JST)

       Ian> I disagree.  You have to learn to think that way, but, once
       Ian> you do, it's pretty easy.  Benefits are the things you want.
       Ian> Features are the things which bring you the things you want.

   If you say so.  But you personally still have trouble, it seems:

I don't think I have trouble, although I apparently misunderstood your

   Sure, you can respond "I declare benefits `volatile'," but as
   Stroustrup points out, most programmers don't notice that declaration.
   If it works for _you_, great!  But my experience in teaching (personal
   and watching my colleagues) is that it is too easy to reify student
   "benefits" into something static, then associate them, more or less
   permanently, with "features" that we put into lectures.  Fortunately
   for us, we have tenure, so that doesn't put us out of business.

You have to be careful with anything.

You have to be just as careful with a focus on what you can do to make
customers happy.  You may wind up solving the particular problems your
existing customers have, without noticing that they are highly
customer specific, and that you aren't getting any new customers.

You would presumably say that you have to avoid that mistake.  I say
that I have to avoid the mistake you pointed out.

None of this is theory, or perhaps I should say that what theory there
is is quite obvious once you see it.  It's all practice.  That's why
so many people do it badly--certainly including myself, but not
including, say, Bob Young, or Proctor & Gamble.

       Ian> The second question tells you how to find new customers--it
       Ian> tells you what market segments to focus on.

   Again, you are presuming your product as currently defined by its
   features.  In the short run, you must do that to survive---but in the
   long run, you want to find out where the customers are going, and be
   there when they arrive.[1]  There may not be very many customers in
   shouting distance of your current product.  What then?

I agree that features vs. benefits is a short run issue.  You have to
solve the long run, but you also have to solve the short run.  This is
business: if you don't solve the short run, there is no long run.

To succeed, you have to make the short run serve the long run.  They
are not independent.