On Tue 17 Oct 2006 at 03:45PM, Joseph Kowalski wrote:
>
> > From: Dan Price <dp at eng.sun.com>
> ...
> > > So, insted we send these beginners off believing that gmacs is the way
> > > it is. I don't think this is better.
> >
> > To me it sounds like you are saying that beginners must be forced
> > through a steep learning curve because it's "good for them."
>
> Not at all.
>
> I'm saying its bad to mislead them into thinking gmacs is "THE WAY".
[from another email]
> You failed to list one choice: "none". As per my next mail, I think this
> is the best choice. I don't like pushing stylistic choices as defaults.
I feel strongly enough that I want to discuss this further. It's this
last sentence which upsets me. The use of "stylistic choices" as a phrase
seems to conjure the idea that we're interior decorators here, and I think
that's misleading.
I'll let others decide if it is in scope for ARC to give guidance, I
have no opinion on that point. As priniciples: Yes, give customers
choice where appropriate. Yes, give them defaults which unlock a
reasonable set of functionality. Make the system approachable to new
users. If possible, give them an experience that delights.
> Oh right, that's your preferred **style**. 8^)
>
> Of course, the *wonderfulness* of gmacs isn't the issue. Its the
> pushing of a stylistic default.
I *never* said I prefer emacs. I only applauded Roland's project team for
making a choice which they believe will be best for the customers they
have studied, and pursuing that. If it was vi mode and they made a case
that vi mode is truly wonderful, then great.
My thought is that we should revamp all of the interactive shell defaults
to have consistent (across the shells) and excellent default interactive
settings, with useful prompts and default behaviors whereever possible.
And yes, we should do so judiciously, with all due deliberation. Would
you defend:
bash-3.00$ (bash)
or
> (tcsh)
as a reasonable default prompt for a shell suitable for an interactive
user? "bash-3.00$" is the kind of crapola we wind up with when we follow
the policy of "none" (bash and tcsh shown, respectively).
-dp
--
Daniel Price - Solaris Kernel Engineering - dp at eng.sun.com - blogs.sun.com/dp
