From: "Lynne Green" <lgreen@cadence.com>
To: "Peters, Stephen" <stephen.peters@intel.com>; <bob_ross@mentorg.com>; <chris_reid@mentorg.com>; <ibis@lumbercartel.com>; <gerald.bannert@icn.siemens.de>; <apanella@molex.com>; <mhaque@ti.com>; <angulo@hyperlynx.com>; "Muranyi, Arpad" <arpad.muranyi@intel.com>; "Peters, Stephen" <stephen.peters@intel.com>; <scott@vasthorizons.com>
Cc: <lgreen@cadence.com>
Subject: Re: Agenda, IBIS Futures Meeting 11/1
Date: Thursday, November 01, 2001 4:12 PM


>Date:           Thursday Nov 1, 2001
>Time:           2:00 - 4:00 PDT
>Dial-In Number: 1-408-894-3000 (Meeting Place)
>Meeting ID:     42479 (IBISX)
>
>
>-------
>Roll Call                   5min   Peters

Stephen Peters, Lynne Green, Arpad Muranyi, John Angulo, Bob Ross
New: Jerry Lalk, Innoveda


>Minutes and AR review       5min   Peters
>  -- review of previous minutes
>  -- AR list
>   * John, upload minutes to web page [Done]
>   * Stephen, PDF current spec and get uploaded to web page [not done]

Stephen has LRM 5.0 posted.

>Opens for new Issues         5min   All

None.

>Old Business                     20min  Peters
>   -- Further update on 'power supply rail selection' method (how to
>      do generics).

How to make sure external power supplies are connected if required, or
internal supplies are used.
Much discussion about various possible implementations.
"select" state does not change during simulation.
Resolution:
Right now, "case" must get its values from the IBIS keywords.  Make it more 
general, and let user/tool pick any valid "case" value. Add attribute, data type of "enumerated list", noting [ ] => a keyword in the IBIS data file, anything else is part of model definition.
Lynne - put the "select" variable name as a parameter on the [Define Model] 
statement?

John: [Pin Mapping] has no macro right now.
(Whole component/pinlist/pkg values is open.)

>   -- Tline 'ref1' nodal format compatibility with N-port matrix
>       (further discussion)

Rlcg matrix in HSpice: 3 conductors + reference, 3x3 matrix.  Known convergence
problems if ref node is not at same DC bias at both ends.
Resolution:
# rows in Rlcg matrix = number of inputs and outputs.
Node list includes ref1 and ref2 for consistency with Spice format.
Attached to node number "0" (global ideal ground).
Document node "0" is not the same as node "gnd".

>IBIS-ML Spec Review           90min  All
>   -- review latest spec (IBIS-ML_06)
>   -- outline of chapter 6

pg 11 - add in the remaining circuit elements (trigger, alarm, driver, delay).
AR: John - Look at Al's list (reshape, invert, others in 3.2).

Table 3 & descriptions - V cannot depend on itself, ditto I.
<,| > for node lists.

OPEN: When trigger is "false", "time-trigger" evaluates to "time".
Using trigger=0 initially may present problems, but so does "NA".

Put in Stephen's updates to 5.4-5.6.

Section 6 limit to expressions.  No mention of functions.
Talk about what an expression is, what an equation is, logical and math 
operators.
Expressions to create a value Relational operators (if/else)

How deeply does this doc get into matrices?  Ref to another spec?  Put it 
in here? Don't want copies of the same info in two different places.

Put datatypes chapter ahead of expressions chapter.

Chapter 8 - add keywords, parameters, subparameters.

>Next Meeting Plans
Date:           Thursday Nov 15, 2001
Time:           2:00 - 4:00 PDT
Dial-In Number: 1-408-894-3000 (Meeting Place)
Meeting ID:     42479 (IBISX)


Respectfully submitted,
Lynne Green