Re: Reserve Demobilization System Built Around Re

[jcardow@afit.af.mil (James E. Cardow)]

esink@turia.dit.upm.es (Eric Wayne Sink) writes:


>What irks me about some of this is the rapidly graying distinction
>between 'reuse' and 'use'.  From a research perspective, most people
>will agree with the idea that reusable software is hard to make.
>Some would even say that USABLE software is hard to make.  But where
>are we going to draw the line between these two ideas ?  If we make the
>idea of 'reuse' too wide, it ceases to be quite so interesting.  For
>example, I'm using a terminal program called Telex to access a Sun
>to write this message.  Come to think of it, I used Telex yesterday too.
>Actually, I've used Telex EVERY day for months !  However, I consider
>this to be software USE, certainly not reuse.  On the other end,
>a linkable library of routines offers me reuse.  A collection of well
>written routines with source offers me reuse.  It's the middle ground
>that bugs me.  Is there a fine line between use/reuse when talking about
>a spreadsheet ?  Is Emacs REUSABLE just because it has a programming
>language built in, or is it just a whole lot more USABLE than vi ?

>I'll admit I don't know how to put this distinction into a well worded
>definition.  Perhaps USE is associated with a USER ?  Therefore, REUSE
>occurs when software is used by a non-USER ? (ugly definition)
>Perhaps, REUSE occurs when software is employed to develop more
>software (but what do you call a compiler then ?)

>I honestly don't know, but I haven't seen a definition I like yet.
>That's why I ask questions. :-)

I agree that there is considerable confusion over the boundary lines.  In fact,
I've used that in classes to make the students think about their preconceived
notions of what constitutes reuse.  In looking for a good definition of
reuse I came across the following by Spencer Peterson:

  Software Reuse  (1)  The process of using pre-existing software during the 
  	development of implementing new software systems and components.  (2)
	the results of the process in (1).

I usually fall back to something I believe comes from Tracz, that nothing
is reusable until it has been used successfully three times.  Trouble is, 
that doesn't account for packages (i.e. a wordprocessor).  Peterson's definition
takes care of that by limiting the use of reuse to construction activities.

Anyone else have a good definition or distinction?


	Jim Cardow, Capt, USAF
	Air Force Institute of Technology
	Instructor in Software Engineering
	Professional Continuing Education Program
		E-mail:  jcardow@blackbird.afit.af.mil