I tend to agree with Ambrish on this one. This is the way I would summarize: If the input portion of an InOut is needed by the resolve function, we need to pass InOut parameters to it. The catch is that the resolve function should not depend on the output portion returned by the Init function of this InOut parameter. So I would prefer to allow InOut to be passed into the resolve function as well, but spell out in the specification clearly and loudly that the resolve function shall not rely on the output portion returned by Init. Thanks, Arpad ================================================================= From: ibis-macro-bounce@xxxxxxxxxxxxx [mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Ambrish Varma Sent: Monday, July 22, 2013 4:57 PM To: ibis-macro@xxxxxxxxxxxxx Subject: [ibis-macro] Re: input parameters to AMI_Resolve in BIRD155.1 Fangyi, If the model maker understands that a) The AMI_Resolve function is called pre-characterization, and b) The AMI_Init function is called post-characterization and c) Any change in the input parameters after AMI_Init call will not affect the dep parameters once they are resolved in AMI_Resolve - then we should be ok with having the same input parameter string to both the functions. We can add the following words (or something similar) to step 9 to make sure that the flow is clear: 'Any change in any of the AMI parameters out of AMI_Init or AMI_Getwave will not affect the already resolved parameters from the AMI_Resolve function.' Thanks, Ambrish. From: ibis-macro-bounce@xxxxxxxxxxxxx<mailto:ibis-macro-bounce@xxxxxxxxxxxxx> [mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of fangyi_rao@xxxxxxxxxxx<mailto:fangyi_rao@xxxxxxxxxxx> Sent: Monday, July 22, 2013 4:57 PM To: ibis-macro@xxxxxxxxxxxxx<mailto:ibis-macro@xxxxxxxxxxxxx> Subject: [ibis-macro] input parameters to AMI_Resolve in BIRD155.1 Hi, AMI experts; In the current AMI_Resolve API BIRD (155.1) we require that only usage In parameters are included in the input parameters string to the AMI_Resolve function as independent variables. At the last AMI meeting we had talked about whether InOut parameters should also be included. I'd like to collect your feedback on this topic. If you have a preference, please share your reasons. Below are the reasons I can see for a yes or no answer. Reason to not include: - Eliminates the possibility of AMI_Init modifying input parameters (independent variables) to AMI_Resolve after it's been called. Reason to include: - AMI_Resolve and AMI_Init have identical input parameter strings Thanks, Fangyi