Minutes of Weekly Meeting, 2014-11-10

Meeting called to order: 11:05 AM EST

1. Roll Call

Ian McIntosh
Carl Walker
Eric Cormack
Bill Eklow
Michele Portolan
Brian Erickson
Tim Pender (joined 11:07)
Peter Horwood (joined 11:10)
Brad Van Treuren (joined 11:13)

Excused:
Adam Ley
Heiko Ehrenberg

2. Review and approve previous minutes:

11/03/2014:

  • Draft circulated 11/03/2014.
  • No corrections noted.
  • Eric moved to approve, seconded by Peter. No objections,  Bill abstained.

3. Review old action items

  • All: do we feel SJTAG is requiring a new test language to obtain the information needed for diagnostics or is STAPL/SVF sufficient? See also Gunnar's presentation, in particular the new information he'd be looking for in a test language (http://files.sjtag.org/Ericsson-Nov2006/STAPL-Ideas.pdf)
  • Ian: Add the previously discussed lists to the 'master' template. Ongoing.
    • Some sections need further expansion that may take time to develop.
  • Ian: Publish Green Paper on website (today). COMPLETE
  • Ian: Publish Newsletter (tomorrow). COMPLETE

4. Reminders

  • Consider Adam's three points (from the action from the first weekly meeting) and suggest what is preventing us from answering those questions:
    • Establish consensus on goals and constraints
    • What are we trying to achieve?
    • What restrictions are we faced with?
  • Forum thread for discussion: http://forums.sjtag.org/viewtopic.php?f=3&t=172

5. Discussion Topics

a. Green Paper feedback.

  • As of Sunday evening (UK time) the Green Paper had been accessed 38 times, although it is clearly impossible to ascertain how fully the article was read each time.
  • The Newsletter tracking showed that 22 people had opened the message.  This represents about 45% of the mailing list and is the average for recent Newsletters.  Peter pointed out that this would not include those people who have notifications turned off.
  • Brad was due to make a post on LinkedIn to announce the Green Paper but had not yet done so.  Ian has supplied some instructional notes.  Getting onto the "JTAG & DFT Engineers" group would help exposure.

b. Demonstration "dry-run".

  • The response to the demonstration invite had been a little disappointing so far: Only Gunnar had requested to attend (via LinkedIn) and there had been no contacts via the Newsletter.  Ian had expected more interest but there was still some time.
  • As Michele had been concerned about the legibility of some windows in his demonstration, Ian invited him to try sharing {Michele shared his desktop}
    • The VirtualBox window was just about readable although the editor window was much clearer.
    • Eric noted that by right clicking on the screen and selecting "View" -> "100%" it became much more legible, although it might be necessary to scroll the window slightly to see everything.
    • Michele re-sized the font in the eMacs editor to make it more readable at default "zoom" settings.
    • Michele's GUI uses a larger font for readability.  However this also opens a lot of windows and it became easier to see everything if the "zoom" was set to 75%.
  • This was sufficient to show that the demonstration was workable.

c. Test Managers - Primary and alternate cases.

  • It was several weeks since we had last discussed this.  Brad recalled that we had decided to pick one case for the Test Manager and run with that.  Ian had made up the slide pack to show that {slides shared} and using the abstracted Test Manager case, with a progression from a generic illustration through a simple JTAG example to the EST example with two Test Managers below the External Controller.  The case with the Test Manager as a layer was also included to show a possible alternative.
  • Ian felt that what was missing was some textual descriptions to help people understand what was intended by each of the boxes.  Also, we had not discussed how we should get this out to people and get feedback.
  • Brad noted that an earlier discussion had suggested defining some of the primitives that might be used to communicate with and between Test Managers.  Those may well be independent of whether the Test Manager is an abstraction or a layer as from the External Controller's perspective it doesn't matter.  What the primitives might be will come down to how much delegation the External Controller can rely on.  Ian saw that meant that you had define some roles or behaviours for the Test Manager that the primitives would act on.  At the same time, that possibly also meant that you needed to define what the External Controller's roles were too.  Besides, we needed to clarify that, as previously discussed, "External" need not mean outside the system, but simply outside the current test environment.
  • What are the requirements that an SJTAG Test Manager must adhere to in order to be classed as a "Test Manager"?  Brad had tried to start on that with his OSI-like layers in 2006 paper but that needed fleshed out and fundamentally it didn't define what each of the boxes were and just identified the interfaces.
  • Ian wondered if the Test Manager was ever described in the meetings that pre-dated his joining the group. Brad replied that there was some definition of the high level and the lower levels but the Test Manager was just this "fuzzy thing" that was needed in the middle.
  • Brad recalled that Gunnar's 2005 paper showed some primitives, and the Shelf Master there may be equivalent to our External Controller.  Brad also suggested that two BTW 2005 presentations in our Stored Documents, by himself and Mike Westermeier, might also contain primitives but after scanning both {shared} this seemed to be incorrect.  The main point about Mike's presentation was that he showed that he could do everything he needed by specializing the STAPL EXPORT statement.  Gunnar's STAPL++ work extended the STAPL Player and added commands to support multiple threads, managing concurrency and dependencies.  Brad's 2002 BTW paper introducing TFCL and Gunnar's 2005 ITC paper both described some primitives but neither were immediately available to the group, although Brad's paper should be in the BTW archive.
  • The Test Manager needs to have some defined behaviours with associated access functions.  Brad was keen to point out that we cannot confine it to all be located on the same piece of hardware, and must allow for it to exist remotely; potentially controlled from outside of what we might call "the system".  This reminded Brad that was never completely clear on what Gunnar was defining as his "system"; although it was ATCA was he considering that as the system or looking a sub-system?  Brad will contact Gunnar for clarification {ACTION}.
  • Brad found he was continually wondering where to draw the line, where to have the locality split. This was especially true when managing data files; where do they reside and how are they managed?
  • In BScan Plug and Play, you pull the test from the UUT over a multidrop interface so the UUT "tells" the Shelf Controller how to test it and the Shelf Controller has a generic means of applying those tests.  This contrasts with Gunnar's scheme where a set of tests resides on the UUT and can be commanded to execute on the UUT itself although there is still a means to pass in additional tests. Both schemes do the basically the same thing - "I need you to run this test, at this particular time".

d. SJTAG Standards Roadmap - continue organising topics.

  • {Not discussed due to lack of time}

6. Key Takeaway for today's meeting

  • The definition of the External Controller, Test Manager, associated primitives need to be developed alongside each other.

7. Glossary terms from this meeting

  • None.

8. Schedule next meeting

  • November 17, IJTAG Engine demonstration.
  • November schedule: 24 (Thanksgiving week), Brad likely to be absent, Bill at NTF.

9. Any other business

  • None.

10. Review new action items

  • Brad: Contact Gunnar for clarification on the scope of his "system".

11. Adjourn

Eric moved to adjourn at 12:05 PM EST, seconded by Peter.

Respectfully submitted,
Ian McIntosh