I have grown weary of responding to Paul Gilmartin's moral outrage of
the day, but today's does require some comment.  He writes:

<begin extract>
If I _own_ a thing, you shouldn't be able to do anything that prevents
my doing what I want with it.  RACF is badly oblivious to this.
<end extract>

This is a bad design notion.  (It is also bad law; the state can
invoke eminent domain to take my property away from me for purposes
that I may disapprove.)  MVS and its immediate antecedents have,
properly in my view, always judged that many operations must be
virtualizable.

The functioning of LOAD and DELETE provide a convenient example of
what this implies    If I LOAD a reentrant program that is not already
present in storage, it is loaded and its usage/invocation count is
increased from zero to one.  If instead it is already present in
storage, having been loaded by someone else, its usage count is
increased by one.  If I then DELETE that program, its usage count is
reduced by one; and only if that count is thus reduced to zero, i..e.,
if no one else is currently using it, is the storage occupied by this
program recovered for reuse.

Paul's notion that the absolutist property-rights doctrines of
19th-century robber-baron capitalism provide an appropriate
system-design paradigm has a fatal flaw.  It prevents sharing those
resources.

Like many of his other comments, this one makes sense for a literally
'personal' computer, one used by a single person.  It is radically
inappropriate to a mainframe, the resources of which are shared by
many people.

John Gilmore, Ashland, MA 01721 - USA

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to