[ibis-macro] Re: Minutes from the 25 August 2015 ibis-atm meeting

  • From: "Bob Miller" <dmarc-noreply@xxxxxxxxxxxxx> (Redacted sender "bob" for DMARC)
  • To: IBIS-ATM <ibis-macro@xxxxxxxxxxxxx>
  • Date: Wed, 26 Aug 2015 15:28:56 -0600

In the interest of amicable <wink!> progress on the "Model_Specific Info
parameter" issue, how about a proposal like this?

Optional Reserved_Parameter IBIS_AMI_Extentions (or whatever name is agreed
on) which defaults to "no" and can be set to "yes". (or whatever boolean
values IBIS likes).

if "yes", IBISCHK (at import time) throws a *warning* and outputs the
Description attached to this parameter. In the specification of this
parameter, we encourage the model maker to describe the extension(s) and
what limitations they may impose on interoperability, etc. The warning
could alternatively/optionally be thrown at run time by the EDA tool, but I
heard opposition to requiring this in the teleconference.

The problems I am trying to address with this are:

- *informing the user that a model may not function correctly in his
chosen platform *(primary goal)
- Neither overly "punishing" or "extolling" the use of extensions. There
are sometimes good reasons for using them and obvious reasons for
eliminating them.
- Leaving some incentive to standardize and/or work around extensions
(less explaining by the model maker to end users if warnings aren't thrown)

Things this doesn't solve:

- Allowing EDA platforms which understand the extension to suppress the
warning (no parsable syntax in the Description). Adding this capability
could significantly increase the complexity and aggravation factor (read:
minimize voluntary compliance) of the solution.
- Identifying whether a particular platform understands the extension
(unless the Description states this, which I am reluctant to recommend;
motivation for extension-savvy platforms to work with model makers and the
IBIS "pirate brethren" represented here to standardize/workaround the
extensions and eliminate the queries from users).
- Others?

I suspect this may not make anyone truly happy; feel free to try to do
better <grin>... It is possible a good compromise will make *no-one* truly
happy.

Regards,

Bob

On Tue, Aug 25, 2015 at 6:01 PM, Curtis Clark <curtis.clark@xxxxxxxxx>
wrote:

Minutes from the 25 August 2015 ibis-atm meeting are attached.


Other related posts: