[78] in Release_7.7_team
Re: another last minute item
daemon@ATHENA.MIT.EDU (cfields@MIT.EDU)
Wed Apr 27 11:15:33 1994
From: cfields@MIT.EDU
Date: Wed, 27 Apr 94 11:14:09 EDT
To: release-77@MIT.EDU
It's not clear to me from the mail to release-77 that this has been
well thought-out. I don't mean to imply that it hasn't been, but the
information is not in this mail message. My problem at the moment is
that I don't see what the planned implementation of this migration is
to be, and I think it's important to know this when deciding what to
do in the release about it.
Just to be clear, my understanding is that the goal here is to be
prepared to easily replace the current help system, olh, with a new
www based system.
Currently, olh may be started either by the user's typing "help" or
"olh." In the future, when www clients are supported for help, what
will the user type? Not "olh," I hope... I associate that with the olh
system, so using the same name would be confusing. What I would prefer
to see is "olh" returning information that olh is no longer supported,
and have "help" start up whatever new system exists.
In proper preparation for a migration such as I have described (and if
you don't like my idea, then don't worry about half of this
paragraph), the proposed change to the olh script is neither necessary
(though it might be preferable) nor sufficient. "help," currently a
symlink to olh, should be replaced with a script checking for the
existence of the new help system, and run either the new system or olh
accordingly. The "olh" script need not be replaced - it is only
necessary to change the database underneath olh to indicate to users
starting it that they should quit olh and run help instead. However,
given the unreasonably long startup time of olh, it may be desirable
to have the script change anyway.
Given that this migration implies that the www clients must eventually
be supported by development (it does imply this, no?), has development
been consulted on this?
Craig
