[magick-users] Convert option order still confusing

[Damien Dunlop]

Thanks once again, again Fred and Anthony,

At last, I think I have arrived at clarity.

> The real problem then is that -draw only applys positions to the real
> image, and not a virtual canvas positions.  This is a historical thing,
> and probably somthing that will need to be fixed for IM v7.
> 
> See http://www.imagemagick.org/Usage/anim_mods/#draw
> for an example of this problem, with animated images.
> 
> PS:  annotate  positions text relative to virtual canvas!

THAT IS EXACTLY WHAT I MEANT!

It is really difficult to convey ideas clearly in writing! I have 
unintentionally inflamed family fueds that way.

I have looked at http://www.imagemagick.org/Usage/anim_mods/#draw which
does make this distinction. It is the only place amongst the heaps if 
IM material I have read where this is discussed.

http://www.imagemagick.org/Usage/anim_mods/#draw:

"Now while `-annotate' places text relative to the virtual canvas of 
each frame, many other images operations do not. This includes all 
`-draw' operations, which only draw things relative to the actual image, 
and completely ignore any offset it may have on a larger canvas." 

It was particularly confusing for me because in the one command
I tend to have heaps of -draw and heaps of -annotate. The latter
confirmed the way I thought it should work, the former was a 
`puzzlement'. The IM `help' was silent on this issue.

I think it definitely needs attention. Once the sequence inside
parentheses comes outside the parentheses, everything should be
referenced in relation to the coordinates of the new sequence now
in existence.

> |
> | Having to use an extra -flatten in the fourth command
> | below to achieve the result I really want feels `redundant',
> | `spurious', `anomalous' etc.
> |
> | ...
> |
> | convert -size 200x200 xc:white \
> |         -stroke black -strokewidth 5 \
> |      \( -size 150x150 xc:lightblue -repage 0x0+25+25 \) \
> |         -flatten \
> |         -draw 'line 0,100 200,100' \
> |         -flatten D.gif
> |
> The extra -flatten did nothing!  Only one image was present and it had
> no transparency to make the 'background' canvas show through.  Note that
> the color of that canvas was not undefined (using -background), so was
> set to 'white' from the image setting, defined by original generation
> "xc:" )
> 
> As such it doesn't just feel redundant, it is redundant!

But I meant the intermediate -flatten felt redundant, which it
would be if -draw positioning worked in the same was as -annotate
(I have in mind more complex commands where a final -flatten would
be needed).

Regards