From tests I've run up to this point, no.

To duplicate the test, save a scene with 1 Michael[2] and 1 Victoria[2]... one with the null included, one without. Start a new session (to clear memory of instances within Poser). Import the saved scene that included the null... set values on Figure 1/Body/Tone, Heavy, Emaciated (channels that are named identically in both figures). You'll notice the second figure mimics the first and independent control is lost for identically named channels in the second figure. Now, start another new session (you must clear the instance count). Load the null from your library. Using the same steps you would in order to normally use the null (select Body, "No Actor" will be displayed), import the saved scene that the null was deleted from prior to saving. Do the same check as above with the like named channels. You'll notice the same affect.

The difference from the "normal" use of the null, is that in the mentioned scenarios the null is not being selected prior to the loading of each figure. Further tests also change my previous assessment of the extent that manual editing of a saved pz3 would rectify the situation (at least without the use of fixed names, which BTW dates as far back as the first day or so of the initial announcement of EMC [June 2000]). Inspecting the pz3 code for the conflicting ERC channels in the above scenarios would show that the proper links (figure exclusive) are established within. Additionally, inspecting the same code in saved scenes that didn't use the null, showed the ERC channels in any other figure than the initial to point at the initial (Figure 2/instancs of :2 point at Figure 1/instances of :1).

One would hope it does get fixed in 5... being that much of the testing, with a number of different scenarios, to pinpoint the problem has already been done. Admittedly, it is quite frustrating. Even more so... trying to find the solution with what we are currently limited to. ;)

-Rob
rbtwhiz.com