Re: automated screenshots (was Re: more doc work, Re: completed gataxx doc)

[Ryan Paul <gaerdin gmail com>]

> Couple of points about it.  I know its only a proof-of-concept:
> 1. Without the instant-apply enabled, how do changes happen?  I presume
> when you save it.  Better to have instant-apply always on, I think.

When instant-apply is disabled, the changes happen when you click the
update button. Instant-apply is off by default because the image
processing operation is relatively processor intensive and I didn't
think it would work well on most computers. Dragging the slider with
instant-apply enabled is a bit sluggish even on my excessively powerful
AMD64 X2 computer.

In the long run, it doesn't particularly matter because the finished
solution would just use default values that we select and the size of
the fading wont be customizable. 

> 2. When the border is set to 0, the image gets whited-out.

Yeah, I'm not entirely sure why that happens. I thought about coding it
so that it would automatically disable the effect (by deactivating the
edge shadow check box) when the slider gets pulled to 0, but I didn't
think it was worth bothering with for a simple proof-of-concept utility.

> >From looking at the sample image you attached: I'm not sure about the
> fading at the edges.  To me, it feels like it looses something (esp.
> with the drop-shadow).  The other problem is that information may become
> faded-out.  E.g. (again from the sample), the status bar at the bottom
> and the title of the window are faded, and difficult to read.

That's a good point, and I don't really have a solution. One of the
other possible approaches that has been discussed to the screenshot
recognition issue is scaling the image down to 70% of normal size, which
I think detrimentally affects readability even more. If I had to choose
between that and edge fading, I would choose edge fading. If someone can
come up with a better approach than edge fading, please share it.

-- 
Ryan Paul <gaerdin gmail com>