Unix system version clobbering

[KMP@STONY-BROOK.SCRC.Symbolics.COM (Kent M Pitman)]

    Date: Wed, 8 Aug 90 15:23 EDT
    From: P. T. Withington <PTW@JASPER.SCRC.Symbolics.COM>

    ...
	What do I do?  Go back to changing file names after each version is complete?
	Use a lmfs partition (yuch!)?  

    LMFS does have some useful features.  Reap/delete protection is one.
    Presumably you could simulate this by adding more funny characters to
    the file name (just as versions are simulated).  The versioning system
    is designed to be compatible with the semi-standard used by GNUEmacs.
    But you would lose this compatibility if you extended the syntax.
				   
				       Set some global retention count control
	variable that I can't find?  Advise the save file routine not to delete
	protected files?  

    You could use the technique most Unix folks use which is to snapshot
    your sources.  sct:copy-system can be used to make a copy of a system in
    some archive.  You could get really creative and use SCCS or RCS to
    maintain your archival sources.  Check everything out into your .sct
    directory, work a bunch, and when you release a version check everything
    in and out.

This isn't an official Symbolics response.  Just a personal response from
someone who happened to notice this conversation going by.

You could also complain to the people who defined the Unix and NFS
standards.  I know this sounds like passing the buck, but I believe that
at least part of the problem is attributable to the fact that when
people don't like some aspect of Unix, they either try to find a private
hack to work around it or else they just give up on Unix and go
elsewhere.  And the information never gets back to the Unix gurus that
people aren't 100% happy with the product as it came from the `factory'.

The failure of the Unix/NFS maintainers to hear a strong message directly
from their customers which says "we want real file versions", "we want
real GRC support", "we want real delete protection" must be what's keeping
it from happening.  I can't believe that if they were really hearing a
resounding cry for these features, that they would turn a deaf ear.

Of course, we (Symbolics) could call up and try to get them to listen to
some of our ideas, but I think we'd be perceived as a competing force 
(in both operating systems and file systems) just trying to stir up some
muck.  I think the feedback about a need for change has to come from 
`unbiased' users--people who have the freedom to choose, a knowledge of
the choices, and a commitment to providing feedback to their vendors about
what their needs are.