On Fri, 21 Sep 2001, Juergen Spitzmueller wrote:

> Allan Rae wrote:
> > Just because a user can't see the changes wouldn't make it
> > misleading. Changing to a different class in my experience has always
> > involved saying "yes" to the question -- when you take a paper for a
> > conference and want to turn it into a journal article you have to
> > comply with what the journal wants (so you have to use their
> > formatting).
>
> Not always. E.g. in Germany Separation is "Indent" by Default. If I
> write a doc and want to have it "Skip" instead for some reason, I
> differ from default settings. If I decide to use scrarticle instead of
> article and confirm the question (or use autoreset), Separation is
> "Indent" again. That's of course not what I want! Same goes for line
> spacing (this indeed caught me some times in the past before I noticed
> what this question leeds to). I think it's good to see *what* will be
> automatically changed before closing the dialog. And the new settings
> will only be taken if you push "OK" or "Apply", so nothing will go in
> without user confirmation. IMHO it's good to have this visual control
> before applying a change.  Moreover, with the new "Reset" button you
> have the additional ability to reset to defaults without changing the
> class. That was not possible before.

So the user might notice something changed by a flash of a text input
changing or a radio button toggling and wonder: "Did I really see
something change?  If so, which one was it?"

With a warning message or some other feedback they'd know they will get
changed entries.  As it is we have surprises.  Sure, you can argue that
the auto reset button is off by default and that it should be clear that
something will be automatically reset but think about a newbie and how
they will react and how they use software.  They/we/I usually don't read
manuals until all else fails.  They/we/I push buttons blindly hoping that
one of them will do what they/we/I want it to do.

Even if we can't tell them which entries have changed (or which are about
to change) we can at least warn them (or confirm for them) that something
did/will change.

> > I think the autoreset stuff is just making the interface more
> > complicated (not quite the word I want... maybe confusing, because
> > things are happening automatically -- so a warning message would be
> > useful but then we get back to the annoying little dialog again...)
>
> I disagree here too. IMHO "Class Options: Auto Reset" is as clear as it
> can be and certainly as clear as a message. People might correct me if
> they think it is confusing. And finally, we will switch autoreset off
> by default, so nothing happens until the user decides to.
>
> Of course, this is just my personal opinion. If people think that this
> is not clear or confusing, I will certainly try to improve the
> behaviour.

I think I'd be inclinded to just get rid of the auto-reset and let the
user push the "class options: reset" button if they really want them
reset.

Allan. (ARRae)

Reply via email to