[20828] in Athena Bugs
Re: clearcerts vs. clear-netscape-password, etc.
daemon@ATHENA.MIT.EDU (t. belton)
Fri Sep 27 15:23:44 2002
Date: Fri, 27 Sep 2002 15:23:41 -0400 (EDT)
From: "t. belton" <tbelton@MIT.EDU>
To: John Hawkinson <jhawk@MIT.EDU>
cc: <bug-infoagents@MIT.EDU>, <web-agents@MIT.EDU>
In-Reply-To: <200209271742.NAA13714@multics.mit.edu>
Message-ID: <Pine.GSO.4.33L.0209271515120.13192-100000@iphigenia.mit.edu>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Yup, all of these points are valid, but there's an opposing view for each
of them.
I got some bad feedback ages ago for having a version of the fix-netscape
scripts which was a little too interactive. In fact I took some of the
clearcerts feedback out because I wasn't sure how much was acceptable, and
I wanted to err on the side of too little rather than too much. If people
can get something like a consensus on how much feedback they want, it is
easy enough to put in.
As for the names, they have been argued before and they'll be argued
again. Greg doesn't like hyphens or underscores and he has made a pretty
compelling case for the dangers of both. On the other hand, you're right
too. This becomes a question of how dumb the user is, as I think with any
'[verb]certificates' name there is some danger that they'll try to make it
two words.
I can make the names parallel but then we have people complaining that we
changed the name yet again! Or we end up with a minefield of symlinks
which - aside from the maintainer headache, which I suppose is
inconsequential to everyone but me - causes problems when one help source
calls the script by one name and others call it by something else and
everyone gets confused.
There is also the matter of how important those Netscape 4 scripts will be
in the future. I would dearly love to get as many people off Netscape 4 as
possible, but I don't have any illusions that it will be deprecated
quickly.
One thing I do NOT want is to call the Mozilla script anything referring
to passwords. Focusing on the cert password, to me, is the wrong emphasis
and misleads the user. I want a name that makes it clear: These Scripts
Kill Certificates Dead.
On Fri, 27 Sep 2002, John Hawkinson wrote:
> [ Feel free to add netscape-release back in if you like. ]
>
> More thoughts on this.
>
> After running both "clearcerts" and "clear-netscape-password"
> together for someone who needed new certs for both Mozilla and Netscape:
>
> It's problematic that they do not have parallel names.
> I was going to run "zap-certificates," but I didn't want to
> have to explain why the command for Mozilla was "clear" and Netscape
> was "zap."
>
> "clearcerts" does its work with no output at all.
> "clear-netscape-passowrd" gives you some reasonable information
> as to what it is doing. At the very least, you could say, "shit!
> I didn't want to remove those, let's get them out of OldFiles.
>
> "clearcerts"'s name is similarly problematic to
> "clear-netscape-password", in that if a user accidently
> adds a space ("clear certs"), they get no error, merely
> a blank screen, which implies that _something_ has happened.
>
> --jhawk
>
