On Thu, 2008-09-18 at 15:42 +0100, Barney wrote: > > I think that would be quite a bad idea, as it just leads to a false > sense of anonymity. How do you figure? > If the pixels are small enough LOL. Yeah, well. I thought everyone would assume that I meant a pixelation with ideally, a selectable pixel size or at least a size well chosen relative to the area being pixelated. > that you can see > that there's a face underneeth then you might be able to recognise the > person, especially with further processing. Oh, like that CSI crime lab stuff where they zoom in 10s of times on a photo and invent clarity in pictures where it simply doesn't exist? > On that basis, I think a better feature would be one to fill a > selection with a solid colour, Sure, that's an alternative for sure. > but perhaps its best left for an actual > image editing app. I see this argument over and over again. For simple operations such a solid fill (as you suggest) or a pixelization, why have to pay the startup and runtime cost and the context switch of another heavy application (i.e. the gimp). Based on your argument, I guess f-spot should get rid of the Crop, Red-eye Reduction, Soft Focus, and whatever else "simple" editing functions are already in it. I'm all for it being an extension that people can choose (not) to load if they wish. But on the "false sense of security" front, it is worth mentioning that I have heard of at least one situation where there was a thumbnail in the EXIF data, so care would have be take care to either replace that with the post-pixelated/color-filled photo. Do any of the editing functions take care to make sure any EXIF containing thumbnails are an accurate representation of the post-edited picture? Or remove (or not include) a thumbnail in the edited photo's EXIF data? b.
Attachment:
signature.asc
Description: This is a digitally signed message part