Mike Robinson <address@hidden> wrote
on 11/05/2009 16:18:52:
> --- On Mon, 11/5/09, Massimiliano Maini <address@hidden>wrote: > From: Massimiliano Maini <address@hidden>
> Subject: Re: [Bug-gnubg] list of issues
> To: address@hidden
> Cc: "gnubg-list" <address@hidden>, bug-gnubg-
> address@hidden, "Christian Anthon"
> Date: Monday, 11 May, 2009, 9:15 AM >
> address@hidden wrote on
> 10/05/2009 12:45:28:
> - 2 buttons for resign and reject is totally fine according to GUI
> they are 2 distinct actions. If we make only one, how do we label
> "Resing/Reject" ? Or have a label that dynamically changes
between reject and
> resign ? Both bad imho. 2 buttons is fine (one grayed outwhen appropriate
> perfect). > > MJR: Both actions mean that you're quiting the
game - one is done
> during play and the other is only done after being doubled. The
> result though is identical - you quit, Sadly the program responds
> resigning (after being doubled) in a different fashion ie it asks
> you wish to resign a double game - this doesn't make sense in this
> situation. Perhaps you guys as developers can understand a reason
> for having two buttons but to an everyday user (ie me) it's confusing.
You didn't answer my question: if you want a single
button, how would you like it to be labelled ? "Quit" ? The two buttons have different icons and different
(and clear) labels: it's not a GUI quirk if you click on "Resign"
instead of "Decline". BTW, as Christian said, if the "Resign" button is
grayed out when a double is proposed, the problemis solved, you will not be allowed to click
A GUI cannot prevent the user from clicking on the
wrong button, unless you want the silly staple from Microsoft Word :) :) (which,
anyway, doesn't do it neither).
> - what's 'the score' you would like to see along the match ? Your
> FIBS rating ?
> Your Error rate ? Both would be almost pointless, even more if they
> account all your past matches (they would barely move during one match
> you have played, say, 100 matches). > > MJR: The score is the fibs rating and it should
be average out over
> a number of games to give a real indication of progress or even show
> any progress or lack thereof (ie -4 in red). Not too many games
> otherwise no change will be seen and not too few so that the score
> wildly fluctuates.
That's the difficulty of it: not too many, but not
too few. Any value is good ... 4 games ? 10 games ? Or 3 matches ? Or
10 matches ? And in case of matches, equal weight for matches of
different lengths ? Progress in backgammon is slow, there's no point in
checking it after each match (except the curiosity of checking the error
> MJR: I'm sure the current system is fine for you guys but if you
> want feedback from average users then you need a system that is
> usable by average users. The current system is not - I can't even
> find this thread using it! If you base all your development
> decisions on the feedback through a utility like this then your
> decisions will be skewed entirely towards what a developer thinks
> user wants and totally ignore real users as they are excluded.
Why don't you give the mailing list a try ? Many subrscibers
are not developpers. Subscribe (chose digest form or message
form), send an email and wait for the answer. Doesn't look that unfriendly
... it's much simpler than a web-based forum.