I've read through the ArcGrid raster example as well as other
documentation, and I'm pretty clear on why and how to create an ImageIO
plugin. It'll take a little "stretching" of the traditional semantics
of the model to make it fit pulling images from a database, but it seems
that the API was largely designed to do this.
My one question is about this line in ArcGridReader.java:
RenderedOp asciiCoverage = JAI.create("ImageRead", pbjImageRead, hints);
I understand what this does, but I'm just curious as to why you decided
to do it *this* way. Couldn't you have just as easily called this
(assuming that you'd created an appropriate ImageReadParam object)?
Is that one "JAI" line the "JAI special sauce" that links the whole
thing into the wacky world of JAI? Is that the line that gives you
cached image tiles, stream-able PlanarImages and all the other JAI
goodness that you mentioned in our breakout a while ago (link below)?
Not that this really changes any implementation details for me, just
curious about that bit.