Re: Reserve Demobilization System Built Around Re

[johnson@cs.uiuc.EDU (Ralph Johnson)]

In article <1991Jun14.231725.1@east.pima.edu>, rharwood@east.pima.edu writes:
|> 
|> COTS (commercial off-the-shelf) software could indeed be defined as "software
|> reuse".  And if that's the case, then everyone who's ever purchased
|> WordPerfect (<=insert the name of whatever word processor you like) instead of
|> WRITING THEIR OWN word processor has indeed REUSED software.
|> 
...

|> (1) Commercial Off-the-shelf (COTS) software:  executable machine instructions
|> written, compiled, and distributed by an organization as a primary market item. 
|> Includes word processors, spreadsheets, desktop publishing, and run-time-only
|> DBMS licenses, and I suppose also includes operating systems.

There is a world of difference between a word processor or a desktop
publisher and a spreadsheet or a DBMS or an operating system.  The
first two are usually systems that are not programmable and can not
be used with other software.  Most DBMSs can be linked to application
programs, and let those programs reuse all the file access and search
routines.  Operating systems let application programs reuse I/O and
file system routines.  A spreadsheet (when treated as reusable software,
and not just a fancy calculator) provides a user interface, a constraint
solver, and a library of math and database routines.

EMACS is more reusable than vi because it comes with an extension
language and can be completely reprogrammed.  Lots of people use it
as a user interface manager.  Similarly, DBMS's, spreadsheets, and OS's
differ in reusability.  Nevertheless, they can all be considered
reusable software.

Almost nobody who gets published equates reuse with making the source
code available.  In fact, most people consider it a liability to have
to distribute source code.  I guess I do to, though I consider it a
necessary liability, because the systems I've seen that offered the
most reuse all came with source.  It is bad to require people to look
at source, but the ability to change the source adds an extra degree
of reuse.

While I am at it, some people think that reuse occurs ONLY when you do
NOT change the source.  I disagree with this opinion.  It is best when
you do not have to change the source because that makes it easiest to
see what is reused and what is new.  However, if you can reuse 1,000,000
lines of code and only have to change 100 lines then it seems to me to
be a great success.  It is impossible for the designers of reusable
software to predict every use to which it might be put.  Access to source
lets the reusers take paths that the designers did not forsee.

Ralph Johnson -- U. of Illinois at Urbana-Champaign, Dept. of C.S.