Back in 1996 you filed for a patent which issued in 2001 as U.S. patent 6,199,076. This actually sounds original for the time; it seems to be a system for providing hyperlinks that could be followed while listening to an audio program, along with a way to jump back to the previous program. Of course, we had those features already in web browsers; whether doing the same thing in an audio program was sufficiently innovative enough to deserve a patent is debatable (and presumably was debated a bit, since it took 5 years for the patent to be issued).
However, that patent in no way describes podcasting, which involves an ability to subscribe to a recurring series of audio programs, including ones not yet issued. That is instead covered by patent 8,112,504, which you filed in 2009 as a "continuation" of the much earlier patent application, one which had, in fact, already been issued as a complete patent for 8 years.
By definition, a continuation application is identical to the parent application, except for the claims. The specification and the figures must be word for word and line for line identical. The claims must also be supported by that specification and those figures. Accordingly, if the '504 continuation describes podcasting (which I'm not taking a position on), then by definition, the '076 patent describes podcasting.
The ability to go back and rewrite your old patent to include new features, and claim you invented them back when the old patent was filed (even if, as you noted, you're limited to collect damages on activity after the new version of the patent is issued) is one thing that is broken in the patent system. You basically saw something that people were doing, found an old patent which bore a little similarity, but which didn't have any claims against that activity you could enforce, and rewrote it so it covered the activity, after the fact. This should not be allowed.
It's not - if you rewrite any part of the specification, or add any "new features", then it's considered a continuation-in-part, not a continuation. A CIP application does not get the filing date of the parent application, so in this case, it would have a filing date of 2009. Anything that happened prior to that would be prior art against it.
But this wasn't a CIP. It's a continuation. It's identical to the parent application, and has that 1996 filing date.
Now I realize that there are legitimate reasons for continuations being considered a part of the original application. But you shouldn't be able to introduce new concepts outside the scope of the original patent application in a continuation. This sort of thing should either be rejected outright, or treated as a new application with priority date set to when the new concepts were first filed.
That's exactly how the system works. And this falls under your first sentence - a continuation that's part of the original application, lacking any new concepts.
Accordingly, going back to your original statement, since the system works exactly the way you say is proper, then nothing about continuation practice is broken.