Re: do_set check

From: George Greer (
Date: 07/11/01

On Wed, 11 Jul 2001, Daniel A. Koepke wrote:

>On Wed, 11 Jul 2001, George Greer wrote:
>> ..., I consider it rather rude to silently change the command that was
>> asked for.
>Isn't it rather rude to completely ignore the effect of the command, then?

The command worked...for 5 minutes.  Options:

1) Silently change the command to memory.
2) Do the write to file, but not the memory. (current)
3) Same as #2, but warn the person.
4) Write to both file and memory.
5) Refuse to write to file while online.
6) Refuse, with override.

I don't like 1, that's not what they wanted.  Number 3 is better than 2,
but still not exactly preferrable.  Number 5 gets rather annoying (my peeve
that gets petted with Windows).[1]  I'd go for 6.

  > set file daniel gender f
  Daniel is currently online, this will only last until autosave or logout.
  Use "set force file" to really do it.


  > set daniel gender f


  > set force file daniel gender f

It's analogous to the case of "fsck'ing a mounted filesystem."  Our current
situation would be at the "happily screw up the mounted disk" stage.

>Perhaps we're coming at it from different angles, but it seems to me that
>if you type 'set file foo sex male' you intend to change foo's gender to
>male, and for the Mud to silently ignore you or even tell you that the
>change "probably" won't take effect is a bit silly.  Does anyone use 'set
>file' with the intention that the change not be made if the person is

I suppose the argument is need.  Do we need to? Is there ever a reason?  I
subscribe to the "enough rope to hang yourself" philosophy because some day
someone will be more clever than we were.  The mindset is tainted a bit by
writing on a paranoia-based, "everything will go wrong" sort of code but
it's still there.

Sort of like fsck on a file.  Who'd want to fcsk a file anyway? They
probably just messed up typing, right?  Until you have loopback mounts...

>Point being that if we ignore in_file, we're silently changing the
>command, true, but changing it so it has the desired effect, rather than
>have it say it's done something but not actually have any discernable

To really throw a wrench in:

        Why even have the distinction between file/game?  Why not just
        search for them online and set it there?  Otherwise we set it in

>I think it's a judgement call, though.

Moving from bad situation to bad situation isn't an improvement, so let's
pick a good one.  Sort of like our act() replacement could never really be
good enough so we'll just keep it around until we think of something.

George Greer

[1] - 1: "What do you mean I can't delete that file?" 2: (for NT) "What do
  you mean 'Access Denied', I'm the administrator, kill that process!"
  Sadly, the computer doesn't understand "Do it or I'll yank the power!"
  as a threat, but at least it still works.  Some people don't want the
  computer to protect them from stupid mistakes because it's a great
  learning experience if it goes wrong and useful if it goes right.
  Just like _never_ use 'cp' to upgrade your C library. :)

   | FAQ: |
   | Archives: |

This archive was generated by hypermail 2b30 : 12/06/01 PST