It's always annoyed me that WRITE permissions were not handled better.
Even under R83, one could set up a file with a user able to read, but
not write to it (file definition attributes 5 and 6, I think?).  If a
WRITE was encountered and the permission was not correct, the program
would just crash horribly.  As an application programmer who tried to
put out solid, error-free code, this really ticked me off.  Any program
I wrote that updated a file could be made to crash by a change to an
external definition.  GAAHHH!!!  This is probably the reason why Pick's
read/write security mechanism never seemed to be used, even though it
was very flexible.

>From my point of view, I want one of the following (although I wouldn't
turn my nose up at both) -

1.  The write fails nicely and I can catch it in code.

2.  I have a REALLY easy way of finding out if the write will fail
before I even try the write.

Fiddling about with the STATUS and then trying to figure out if this
user is owner, or in the group, or whatever is VERY trying.  Especially
since a user can be in more than one group (On this UNIX box, anyway).

It seems to me this should have been addressed years ago - wasn't the
Spectrum Manfacturers' thing a waste of space?  IBM, holding 2 big cards
at the multivalue game, ought to be taking a lead on this, but they'll
probably leave it to Martin

Regards, Keith
-------
u2-users mailing list
[email protected]
To unsubscribe please visit http://listserver.u2ug.org/

Reply via email to