LibTiff Mailing List
TIFF and LibTiff Mailing List Archive
August 2005
Previous Thread
Next Thread
Previous by Thread
Next by Thread
Previous by Date
Next by Date
Contact
The TIFF Mailing List Homepage
This list is run by Frank Warmerdam
Archive maintained by AWare Systems
Thread
2005.08.09 06:33 "Re: Overlay Hyperlinks On TIFF", by Rob Tillaart
Joris wrote:
> The region type value can be either of
> 0: the width and height values are meaningfull, the clickable area is a
> rectangle
> 1-65534: the width and height values are to be ignored. Instead, this value
> describes are radius, and x and y describe the center point of a circular
> region. This type of region should not be included if XResolution and
> YResolution tags are in the IFD and have different values, since pixels in that
> case are not rectangular and a radius in pixels does not make sense.
> 65535: the widht and height values are to be ignored. Instead, after all regular
> rectangular regions (region type = 0) and circular regions (region type 1 to
> 65534) have been checked, and no region is found to comply, the x and y of this
> type of region that is nearest is said to be clicked. Again this type of region
> should not be included if XResolution is different from YResolution.
some refactoring :)
regiontype: name description
--------------------------------------
0000: [rectangle] the width and height values are just that
0001: [ellipse] the width and heigth values are the lengths
of the axis of the ellipse.
For circles they are identical
Note that if the X-Y res differs one still can
define perfect circles
FFFF: [nearest] see above 65535
This scheme include ellipses and leaves room for 60000+ other shapes and/or
attributes. The [not] and [doubleclick] are region attributes: E.g. the
[not rectangle] is the complement area of [rectangle]. all area outside the
rectangle.
> It can happend that multiple regions apply (either rectangular or circular, it
> cannot happen that more regions of type 65535 apply, except if these regions
> have identical x and y values). This case is undefined, and application code can
> de whatever it wants with this. Writers should avoid writing clickable area data
> with overlapping regions because the behavious in such case is not defined.
What to think about a 'nearest point with attracting factor' region type?
define a number of points with a different charge. Gravity. Automatically
creates area's. The ultimate region type is the polygon I guess.
> Just for the fun of it, though... ;-)
fun is the grandmother of creativity !
regards,
rob