Minutes of Weekly Meeting, 2008-11-10

Meeting called to order at 8:20am EST

1. Roll Call

Carl Walker
Eric Cormack
Ian McIntosh
Peter Horwood
Adam Ley
Brad Van Treuren
Carl Neilsen

Heiko Ehrenberg

By Proxy:
Tim Pender

2. Review and Approve Previous Minutes

  1. 10/13 Minutes: Approved (moved by Ian, seconded by Peter).
  2. 10/22 Minutes: Approved (moved by Carl W., seconded by Ian).
  3. 10/30 Minutes: One correction noted: Affiliation of Paul Holowko should read "BAE Systems" (instead of "BEA Systems").
    Not enough members available to approve. Deferred until a later date.

3. Review old action items

  • Adam proposed we cover the following at the next meeting:
    • Establish consensus on goals and constraints
    • What are we trying to achieve?
    • What restrictions are we faced with?
  • Establish whether TRST needs to be addressed as requirements in the ATCA specification if it is not going to be managed globally (All)
  • Adam review ATCA standard document for FRU's states
  • Brad to contact Rohit regarding IEEE policy for interviews.
    Ian will take ownership of this action from Brad.
  • Ian contact Mentor Graphics for EDA support person.
    Done. See discussion topic 4e.
  • Patrick contact Cadence for EDA support person.
    Expect report if Patrick attends 26/11 meeting.
  • [Ian] Were some of the very old actions ATCA specific questions? Do we still need to continue these actions?
  • [Brad] Adam's action on FRU States was ATCA specific.
  • [Ian] I missed the meetings where the first actions were set - I wasn't sure how relevant they still were.
  • [Brad] The TRST issue is one that we still need to consider sometime - Do we have TRST or not.
  • [Ian] OK. We'll keep these actions on then.

4. Discussion Topics

    1. Election of Officers
      • Chair:
        The motion is that Ian McIntosh be elected as group Chair (Proposed by Brad, Seconded by Eric).
        There being no dissenting or abstained votes, the motion is carried and Ian McIntosh is hereby elected to the position of Chair.
      • Vice-chair
        The motion is that Heiko Ehrenberg be elected as group Vice-chair (Proposed by Ian, Seconded by Carl W.).
        There being no dissenting or abstained votes, the motion is carried and Heiko Ehrenberg is hereby elected to the position of Vice-chair.
      • Tim Pender voted by proxy in both elections.
      • [Ian] There had been some earlier discussion about giving Brad and Gunnar "Emeritus&quot status in recognition of service?
      • [Brad] That's right, but that's up to the group.
      • [Ian] I'm in favour of that. What effect does it have on Ben's Chair Emeritus status? Can we have two Chairs Emeritus?
      • [Brad] I don't think Ben would be too worried about it now.
      • [Adam] Emeritus is an honorary title, and can be held by as many people as you like.
      • [Ian] So do we agree that Brad becomes Chair Emeritus and Gunnar Vice-chair Emeritus?
      • [All] Yes.


    1. ITC summary (Brad)
      • [Brad] The Poster session was very busy. I didn't even manage to eat but Adam managed to bring me refreshments. There was a big spectrum of interest in SJTAG.
      • [Brad] Scott Davidson (Sun Microcomputers) is becoming involved with a government/military contract/committee that is looking to get a presentation on SJTAG sometime. Heiko is also expecting to be involved.
      • [Brad] The Fringe Meeting went well. Paul Holowko was keen for us to look at the APIs used in other industry sectors such as automotive - look at the interfaces and don't worry too much about what happens behind them.
      • [Ian] I only had a quick look at the references in the minutes. At first glance, I can't say how useful they will be.
      • [Brad] Yeah, I quickly put those links in so there was sort of reference. CANOpen looks like something we could take something from, as it's already used by 22 device manufacturers.
      • [Ian] CANOpen sits on a fairly rigidly defined network (CANBus) - I don't know if we can offer that?
      • [Brad] No, but it may help us to know how much we might need to tighten up our hardware definition. But I don't want things to go too tight and end up like another dot5 bus. Dot5 tried to do too much for too many people.
      • [Ian] Yes, and at the time many people found that a ScanBridge would do all they needed.


    1. Data Element validation against Structural Test Use Case
      • [Brad] Structural Test seemed like the best place to start trying to validate our data elements against a Use Case, as it involves ATPG, vector execution and diagnostics. It's also the Use Case most people understand best.
      • [Ian] Device Programming may have been simpler in terms of the required data, but I agree that Structural Test is the most familiar.
      • [Brad] I struggle with the partitioning - there is the aspect of ATPG, the aspect of the differences between EBST and XBST, the aspect of representation of results - deferred diagnostics versus real-time diagnostics.
      • [Brad] My experience is that tool vendors want to defer the diagnosis, but the embedded user often doesn't have that option and needs real-time diagnostics e.g. via a console interface.
      • [Brad] At the fringe meeting Paul Holowko pointed to the automotive industry and the use of diagnostic codes to indicate classes of failure.
      • [Ian] Ah, hence his reference to listing ambiguity groups that could cause a particular failure code?
      • [Brad] Yes, so for example, can we standardise on how failure codes are reported?
      • [Brad] Paul also commented that when they get fails on tanks, they record additional environmental data, so that they get a better picture of what the conditions at the time of the fail were.
      • [Ian] Not all systems will have that kind of instrumentation available. We often have several temperature sensors in our systems, but that's probably all.
      • [Brad] But more devices are appearing with some kind of environmental monitoring included - P1687 will also help. This aspect could be an extention to a standard. Maybe it's not in our scope - we need to decide.
      • [Ian] I can see it as an option, but not a mandatory element of SJTAG.
      • [Brad] I agree.
      • Brad had to break off at this point.
      • [Ian] The discussion has really looked at diagnostics so far. What about, say, the ATPG side?
      • [Ian] What data do we need to support test generation?
      • [silence]
      • Do people remember what Brad showed on the Data Elements diagram? We re-worked it a bit to try to make it easier to understand
      • Ian tried to share a slide of the simplified Data Elements diagram using the Cisco MeetingPlace tool, but only Carl W. and Ian were on-line. Ian tried to describe the slide to the others.
      • [Ian] This diagram tries to show the hierarchical structure of the system, with each assemly pulling in data like netlists, BoM, and part descriptions, but I'm sure that this isn't complete. Straight off, I'd comment that we don't have constraints like reset conditions that need managed.
      • [Carl W.] Yeah, I can see that. It's a good 50,000 foot overview, but it doesn't contain the detail.
      • [Ian] It's a more accessible view that the original UML, but the UML is probably easier to add the detail to. I guess we really need both versions.
      • [Carl W.] This one's good for getting the general idea across.
      • [Ian] Can anyone else identify any missing data elements?
      • [silence]
      • [Ian] I think this discussion is stagnating for now. I think we need to take a bit of homework and review the diagram offline, and try to brainstorm what might be missing to support Test Program Generation. I think the ATPG and diagnostics are going to be the hardest parts to figure out.
      • [Ian] I'll mail out copies of the diagram so every one has a copy.
      • [Carl W.] The meeting invitation included links for Cisco internal and external access to the MeetingPlace tool. You just follow the link and download a couple of bits of Java. You can then share and view applications or desktops.
      • [Ian] Yes. The installation was pretty painless. That surprised me because the IT security restrictions here at SELEX usually mean that anything remotely useful won't actually work!
      • [Carl W.] Yeah, the same here!


    1. "October" Newsletter (Ian)
      • [Ian] I held back on the October newsletter so I could get some material from ITC included, so I guess this will end up as an October/November issue.
      • [Ian] Where we are now it looks like we'll only have a few meetings in December, so maybe it makes sense for this to be the last one this year, then have a December/January issue?
      • [murmurs of agreement]
      • [Ian] I'll compose a draft and send it out. Hopefully we can approve this at the next meeting.


  1. EDA participation (Ian)
    • [Ian] I had about an hour's worth of discussion with Mark Laing at Mentor Graphics. He's interested in what we're trying to do, but doesn't believe that any one person will be able to supply the input we need: We may need some input on devices some on PCBs, etc.
    • [Ian] However, Mark felt that our first point of contact ought to be with the schematics people. He thought that there were some useful guys at Mentor in the UK he could put me in touch with, but as I was out of the office most of last week, I haven't been able to follow that up.
    • [Eric] Is there any reason we don't ask Synopsys?
    • [Ian] None.
    • [Eric] Well I can take an action to make a first contact - I'll be visiting them soon.
    • [Ian] Thanks, Eric, that would be really helpful.
    • Brad rejoined

5. Schedule next meeting

Monday, November 17th, 2008, 8:15am EST
Wednesday, November 26th, 2008, 8:15am EST
  • [Ian] I'll miss 11/26, as I'll be presenting at NTF that week.
  • [Ian] Following our natural cycle, December meetings would fall 12/8, 12/15 and 12/24. I guess not too many people will want a meeting on Christmas Eve?
  • [pause]
  • [Carl N] I think they all just answered!
  • [Ian] OK, so 12/8 and 12/15 will be our final meetings for the year?
  • [Eric] That seems fine.
  • [Others] Yes.

6. Any other business


7. Review new action items

  • Ian - Distribute Data Elements diagram (UML and simplified versions)
  • All - Consider what data items are missing from Data Elements diagrams for Test Program Generation
  • Eric - Make initial contact with Synopsys on possible EDA participation

8. Adjourn

Meeting adjourned at 9:13am EST (Moved by Eric, Second by Carl)

Respectfully submitted,