Phone Conference P1450.3 Working Doc Subgroup

Friday, August 2, 1:15 to 2:30pm PDT

Attendees:
Daniel Fan
Greg Maston
Tony Taylor

Agenda:

DF-4 - resource allocation

FM-1 - tutorial (Dan's proposal)

JT-5 - LoopCharacteristics

TT-3 - TRC-Err definitions

Discussion:

The 1450.1 and 1450.3 working groups are planning to hold a joint working group meeting at ITC on Monday, Oct 7. Dan, Tony, and Greg indicated their interest in attending.

DF-4 - resource allocation

We first discussed the merits/de-merits of each of the three proposals (identified as JT, TT, DF) (please refer to the 6/21 minutes for the proposals):

• The JT proposal utilizes the InheritWaveformTable as a mechanism for structuring the waveforms such that only one actual definition is in the file and all others reference it.
  • PRO: This is a viable way of defining the waveforms and, if used, ensures same definitions
  • PRO: No change to existing syntax.
  • CON: Does not provide a way of handling files that are generated without adhering to this convention
  • CON: Does not provide a way to map different ATE's differently.
  • CON: Does not work for pattern resource allocation (i.e. time set mapping)
• The TT proposal leaves STIL syntax completely unchanged, and adds the definition of a resource tag - a label enclosed in angle brackets <xxx>.
  • PRO: No change to existing syntax
  • PRO: Can be superimposed (i.e., post processed) onto any STIL file
  • PRO: Can support multiple ATE architectures by having different labels for each
  • PRO: Works for timing and patterns
  • CON: The STIL code can be changed and obsolete the resource labelling.
  • CON: No guarantee that the waveforms are indeed compatible, just because they are labelled the same.
• The DF proposal uses regular labels (xxx:) to identify where things are defined and uses resource lables (<xxx>) to show references to them.
  • PRO: Guarantees that there is only one definition of a given waveform
  • PRO: Somewhat reduces the amount of code
  • CON: As shown in the example, it is not really STIL.0 compliant as it is not permissible to define multiple periods and multiple waveforms for a wfc within one wft.
  • CON: Supports only one ATE mapping.

In summary we decided that the JT example should be included in the document as an example of how to construct waveforms that define inherent re-use. We decided that the TT scheme of allowing resource labels to be inserted in front of any STIL keyword (block or statement) is the most flexible and will be added to the 1450.3 standard.

We need to contact the 1450.6-wg to see if there are other applications for this resource label. If so, it may be more appropriate to define it in 1450.1.

FM-1 - tutorial (Dan's proposal)

We briefly reviewed the state of the tutorial. We basically like the direction it is going and decided to: 1) get the latest version onto the web site, and 2) be prepared to discuss in detail at the next meeting.

JT-5 - LoopCharacteristics

This topic has been pretty much discussed in prior meetings. The changes should be added to the doc. One change that we identified is to not have a separate Repeat statement, but to handle it with a named Loop block. And, to allow the InstructionCharacteristics statement to have an optional label when pointing to a Loop block.

TT-3 - TRC-Err definitions

There was general agreement as to the TRC-error concept. One change is that we need to set up a documentation scheme such that error numbers remain constant. This is to be done by putting a numbered table as an Annex and refer to it from the body of the document. That way all new TRC-errors can be added to the end of the table.

Topics for next meeting (Friday 8/16/2002)

DF-4 - resource allocation

FM-1 - tutorial (Dan's proposal)

JT-5 - LoopCharacteristics

TT-3 - TRC-Err definitions
 

Action Items:

AI-1: Tony - update document for resource allocation

AI-2: Tony - update the document for LoopCharacteristics

AI-3: Dan - distribute the latest tutorial

AI-4: Tony - put all the latest docs onto the web