Comment 30 for bug 882274

Em 30-10-2011 10:44, Mark Shuttleworth escreveu:
> On 29/10/11 15:40, Marco Biscaro wrote:
>> I can't! :) I can't understand the costs of adding an option in these
>> cases. Could you (or someone else) tell me how the users that don't
>> use this feature would be affected?
> Say the option is expressed in a dialog box. That dialog box is now
> longer and more complicated than it needs to be. When a dialog box is
> more complicated, it feels harder to use, for everyone.
> So, putting that option in a visible UI makes the system feel harder for
> everyone. You know the feeling - you just want to get something simple
> done and you have to look through lots of complicated dialog boxes. Now,
> if you're having fun exploring all the options in the system, that's not
> work, that's fun. But for most people, they are happiest if the system
> just works; having to go find an option to change its behaviour is bad
> for them, and having lots of options is worse.
> That's the cost to end-users.
> So, why not make it an invisible option, say a dconf key that could be
> turned on with a command line tool or power user config tool?
> The cost there is in the codebase and in the design.
> Say the option allows for both vertical and horizontal layout. Then we
> need to think about all the behaviours and mechanics both horizontally
> and vertically. And if it's horizontal, we have to think about
> right-to-left languages as well, like Arabic. So the number of codepaths
> just got much bigger, the number of quality assurance tests just got
> bigger. We only have time for a limited amount of testing in any given
> cycle, so that option means we won't get something else.
> All these costs add up.
> Mark
Thanks for these clarifications. It would be much more simple if they
were posted in the original bug report, without the need of such
discussion. I'll quote this explanation in the original bug report.