On Fri, 2005-03-11 at 08:31 -0500, Steve McClure wrote: > On Fri, 2005-03-11 at 07:08, Christian Reis wrote: > > On Sun, Feb 27, 2005 at 11:52:41PM +0000, Gustavo J. A. M. Carneiro wrote: > > > On Sat, 2005-02-26 at 08:43 -0300, Christian Reis wrote: > > > >On Fri, Feb 25, 2005 at 11:08:00AM +0000, Gustavo J. A. M. Carneiro > > > >wrote: > > > >> My solution, although also not perfect, was to place a small > > > >> gtk.STOCK_DIALOG_WARNING icon inside the entry, and toggle between > > > >> making it insensitive (=value OK) or sensitive (=bad value). > > > > > > > >Interesting! Can we see some screenshots of this? > > > > > > Attached. Code in [1], but you won't be able to run it without > > > modifications (or installing gnumexp), I'm afraid. > > > > Would it be worse to make the icon invisible instead of inactive? > > > > Second question is if it would be possible to right-align the icon > > (instead of left-aligning it, where it makes reading a bit more > > difficult). > > > > > > I had never thought of > > > >this approach and it sounds way cool (though it _does_ take up some of > > > >the entry's real estate). Dream scenario might be a window which > > > >overlays the screen when a validation error occurs. > > > > > > Now that I don't understand. Can you explain better? "window which > > > overlays the screen"? > > > > Basically, balloon help for input errors. > >
> > .---------------------------.
> > | sin(x)+ .---------------------------------------.
> > '---------| /!\ This should be a valid expression |
> > '---------------------------------------'
I would find this a bit too annoying. Validation should be much more
dissimulated, IMHO.
> >
> > It could be that this appeared only when a certain time had passed
> > (though the input would still be considered invalid) or if focus left
> > the input area.
>
> Jef Raskin mentions this in The Humane Interface. In his view, the
> message window should be semi-transparent and not interfere with work.
> That is, you don't have to click on it to acknowledge it, just keep
> typing/clicking and it then will eventually fade away. He likes that
> concept in place of dialogs that popup and require you to read the
> message and click OK to continue. It seems appropriate here as well,
> they just wouldn't fade away until the error is cleared up.
Well, that would be interesting idea, but unfeasible in X11 until we
have Composite X.Org extension widely deployed and stable.
> >
> > But I have come to generally think that multiple mechanisms for
> > different types of validation errors are the way to go; that implies
> > having a [small] set of forms of feedback implemented in a library.
Agreed.
Regards.
--
Gustavo J. A. M. Carneiro
<[EMAIL PROTECTED]> <[EMAIL PROTECTED]>
The universe is always one step beyond logic.
smime.p7s
Description: S/MIME cryptographic signature
_______________________________________________ pygtk mailing list pygtk@daa.com.au http://www.daa.com.au/mailman/listinfo/pygtk Read the PyGTK FAQ: http://www.async.com.br/faq/pygtk/
