On 14/08/06, David Gowers <00ai99@xxxxxxxxx> wrote:
While the fill/stroke dialog could stand some improvement, I think that
being able to duplicate vector layers properly is more important for
usability; currently any duplicates behave like ordinary pixel-based layers,
therefore do not keep up to date with changes to their base vector.
I am not sure what behaviour would be most useful here, a shallow copy
or a deep copy. A shallow copy would let you easily create multiple
vector layers based on the same path, which would easily allow for
neat effects if different stroke widths / styles were used, but a deep
copy might also be useful for making sure the layers could be changed
seperately.
Come to think of it, in my mind the shallow copy would seem to be the
most intuitive. However, that could just be me. =)
I suggest delaying updating the vector layers until the user releases the
mouse button as they edit the base path; Editing paths used by vector layers
is currently rather sluggish.
That is something that I am considering. I have not personally had
experience using this on a slower system.
While presently vector layers are useful as a drawing tool, they would
approach the usefulness of normal layers if they could be moved. As far as I
understand, information about any necessary transformations can be stored in
the form of a single transformation matrix (as a parasite?)
The way I am currently thinking of implementing transformations is by
just applying the transformations to the original path instead. It
seems to me that this would be most intuitive in that it makes sure
that the vector layer always corresponds to the position of the path
it is based on. More complicated things, such as having multiple
vector layers based on the same path but in different places, may be
considered in the future.
Thank you very much for the input,
Hendrik Boom
_______________________________________________
Gimp-developer mailing list
Gimp-developer@xxxxxxxxxxxxxxxxxxxxxx
https://lists.XCF.Berkeley.EDU/mailman/listinfo/gimp-developer
