Originally posted by stevebrot As a programmer, I would suggest that this evidence of a limit of design. Hyperprogram adheres to the program line. Changing the ISO dynamically introduces an indeterminate factor. While this would not be insurmountable, I am pretty confident that this "bug" was known during design phase and that the current behavior was deemed a reasonable compromise.
Well, as a software developer you should know that bug-free software does not exist.
I remember that some years ago, when we, Auto ISO users, were complaining about Exposure compensation defeating Auto ISO, many dpreview forum members also were talking crap like "it's not possible", "too difficult to implement", "too many variables" and so on.
Being software developer myself, I knew, that implementation of Auto ISO and exposure compensation working together is very easy. Instead of metering light and passing the value to routines calculating shutter speed, aperture and ISO, meter the light
adjust by the value of exposure compensation and pass the value to routines, calculating aperture, shutter speed and ISO. That's it. However many "engineers" like you were saying that it's impossible and so on. Pentax fixed it in K10D and I am sure that it was one line fix.
The same goes for Hyper Mode. Pentax
fixed it in K-7, however you still are arguing that it works OK in K10D/K20D?! Well, well...
By the way, did you look at those links which I gave to you? They explain what hyper program is.
Did you try to read, understand and repeat the steps which lead to clear demonstration of the bug, we're talking about? I am sure you did not.
And, please, don't catch me by my words, as some of you did with that "P works as I expect to work". English isn't my native language, it isn't even my second language. Therefore I often have difficulty expressing myself. Thank you for your understanding.