Image specific restriction option

[Rick Brewster]

Unfortunately a feature like this doesn't really make sense when you consider the engineering ramifications. You can't just "disable transparency"; this is because it's an integral part of the composition engine. A feature like this would add combinatorial complexity to all the performance-critical portions of the code as well. Instead of writing code that says, "the resultant pixel equals this combination of layer 1's pixel and layer 2's pixel" it would have to be written to be, "if they've disabled feature A, then use this method of combining, but if they disabled feature B then use this method, and if they've disabled both A and B then use this other method, .... otherwise use the normal method." And then imagine what happens if "C" and D" are added into the mix. And then imagine that strewn across 86,000 lines of code.

It doesn't scale at all because anytime you introduce an extra layer of something that can be disabled you have to weave extra code all throughout the code base. You can't just put that code in 1 place, which is what good OO programming and maintainability dictate. It's also a very high-risk as a "bug hazard": if I have to add code in 20 places, that's 20 chances to introduce a new bug.