Re: Enable accessibility by default in development releases?

[Elijah Newren <newren gmail com>]

On 11/11/05, Claudio Saavedra <csaavedra alumnos utalca cl> wrote:
> On Fri, 2005-11-11 at 12:00 -0700, Elijah Newren wrote:
> [...]
> > FWIW, it effectively disables the reduced resources mode in metacity
> > (except that the minimization animation remains off) which would
> > result in part of Metacity being untested.  It may also drive users
> > mad from the constant slow keys and sticky keys dialogs that appear.
> > (I know it shouldn't seem like common operations to press shift 5
> > times in a row or hold it down for over 8 seconds but I seem to
> > periodically do one or the other without thinking about it,
> > particularly the latter fairly often)
>
> Considering that there are pros and cons, why force all developers,
> contributors and testers to take part in this? It would be good for the
> debugging of a specific part of the desktop, sure, but at a cost that
> probably not everyone wants to pay.

That sounds like "since there are tradeoffs, let's ignore the proposal
(and implicitly all the tradeoffs) and keep things as they are."  Why
not weigh the tradeoffs and determine which is the more sane default? 
In particular, it's not hard to manually turn accessibility off. 
There's even a little dialog for it.

If you don't like it on by default, please explain why the change
overall does more harm than good.

> Just like with the suggestion that Vicent made a few days ago regarding
> making GNOME crash. Instead of making this changes a *must* for all the
> people involved with the HEAD code, why not better making more noise
> (mailing lists, p.g.o, etc.) so everyone know that exists some areas of
> the desktop development that need some love?

Most developers may not trigger the critical warnings, or may not be
aware of them since they don't watch the terminal (or, even if they
do, they don't watch closely enough and the critical warnings get
drowned out by all the normal warnings).  Granted, work should be done
to get rid of any common ones before it's turned on, but I think
Vincent's proposal makes sense as a default at some point.  (However,
if it turns out that crashes aren't rare after turning it on, then of
course we'd need to turn it back off until we could fix stuff to the
point that they are more rare.)

Just my $0.02,
Elijah