Minutes of Weekly Meeting, 2013-01-21

Meeting called to order: 11:07 AM EST

1. Roll Call

Adam Ley (left 11:56)
Brian Erickson
Harrison Miles
Ian McIntosh
Brad Van Treuren
Heiko Ehrenberg
Tim Pender (joined 11:10)

Eric Cormack (comment by proxy)
Peter Horwood
Carl Walker

2. Review and approve previous minutes:

1/7/2013 minutes:

  • One correction noted: Change 'Ian wondered it was best...' to 'Ian wondered if it was best...'
  • Brad moved to approve with the above correction, seconded by Heiko. No objections or abstentions.

1/14/2013 minutes:

  • One correction noted: Add Adam as Excused attendance.
  • Insufficient members present for approval.

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
  • Harrison will attempt to come up with a table of use cases and their associated layer and what can be done at that layer. Ongoing.
  • Ian: Split the Use Cases covered today onto a new slide and add annotations. 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

  1. Newsletter.
    • Ian had received no topic suggestions for the Newsletter. The opinion within the group was that the current exercise was not yet sufficiently developed for an article. Heiko wondered if the consideration of support for mobile devices on the forum was worthy of a small article. Ian thought it might be, especially if there was an easy way for readers to vote on the idea, so he would consider this.
  2. Proposals for SJTAG at ITC or BTW 2013
    • Arrangements for BTW apparently tend to be 'last minute', partly to fit in with the ITC schedule which has been 'oscillating' over the past few years. Although usually held in Intel's Design Center in Fort Collins there are suggestions of moving it to Oregon this year. BTW is less formal in both its submission and presentation processes, allowing it accept papers later than with ITC. Bill Eklow had not yet responded to Harrison's email and Brad noted that Bill was heavily involved in an issue he was coordinating that may make him less responsive.
    • Brian noted that while BTW was an interesting venue, it had a very much smaller attendance than ITC, perhaps only 20 people last year, and therefore offered less exposure. Tim and Brad felt that while it may offer less exposure, everyone in attendance would be interested in our subject, which has not been the case at ITC. There was an opinion that BTW was more 'forward looking' and broached subjects that would be picked up by ITC in subsequent years.
    • Brian suggested the we could support both events, if necessary. Ian felt that depend on what we proposed to offer for each. Brad pointed out that just offering an overview would not be adequate for BTW, but if it could start to describe the methodologies and architectures then that would be good. A benefit of BTW is that it is possible to get instant feedback on a concept due to the more intimate surroundings and the ease of extending into offline interactions.
    • Brad felt that 1581 benefitted from its participation in BTW, and Harrison added that by getting iNemi involved that helped to gain access to other venues.
    • Ian noted that setting the workshop arrangement details late did not help in planning for travel - a late request for travel had much less chance of success. Adam commented that Bill would likely provide a letter to assure an opportunity at BTW if that would help.
    • Proxy comment from Eric Cormack (unfortunately I received this too late for it to be included in the discussion):
    • For the workshop idea for 2013 ITC I think the first thing has to be who can attend from our working group, I hope to this year but will be late in confirming this, and the duration of the workshop. Most now expect a workshop to be a day so we have to think of how we fill 6-8 hours of presentation/discussion activities.
    • Also, as Peter mentioned last week, do we also need hardware/tools demonstrations as a side-bar or as part of the main feature?
  3. The relationship between lifecycle phases and Use Cases.
    • Ian had circulated updated slides by email prior to the meeting.
    • {Slides shared}
    • Slide 5:
    • While Harrison liked the slides he was unsure how to perceive the 'levels' from this view, that as the level increased the observability would typically decrease, so the table possibly represented the view on the top level. Ian replied that it was his observation from the last meeting that there were multiple dimensions that became apparent while trying to create this two-dimensional view. Brad noted that as an example there was a dependency ordering that doesn't get expresses here: You would check the infrastructure before trying to debug the interconnects. However the point of Grady Booch's method is to concentrate on a single perspective. Once you have several perspective then you try to assemble those and look for the commonalities and differences.
    • The consensus was that Slide 5 accurately reflected the conclusions of the previous meeting.
    • BIST:
    • Brad considered that the economics question applied here too: If a pluggable instrument was readily available then it was more viable. BIST could be an advantage in Development if it helped to meet a market window. Harrison felt that an organisation that heavily used mid and upper end FPGA would be well positioned to use BIST at almost every stage. Ian agreed but thought that possibly just a refinement of the more general case that Brad had described.
    • Brad observed that the tooling to support any instrumentation was also important; how easy it was to incorporate into the process flows. Typical vendor tools do not integrate easily. Often these are GUI based and the desire is to automate their use within some other process. Harrison commented that previously vendor tools had often focussed on maximizing routing efficiency, but now customers were demanding support for design verification.
    • Brad observed that it had taken 10 years for JTAG tools to get a good level of integration. Harrison agreed that it would take time, but not another 10 years; more likely between 3 and 5 years.
    • Fault Injection:
    • Ian explained that he viewed Fault Injection as primarily a Development activity. There was some potential use in Prototyping, while for Field Service and Workshop Repair it was more likely be to assist diagnosis of some other element.
    • Brad remarked that make effective use of Fault Injection the designer needs to consciously design where he wants to use Fault Injection, as that enables reuse at other stages.
    • Harrison thought that very few segments of industry would be using this, while Ian felt that may be simply that many people don't recognize that the opportunity exists.
    • {Adam left}
    • Brad suggested that the main use was for software validation and that it was generally considered to be outside the scope for hardware designers. It required an 'Aha! moment' to use it for other applications. Brad added that the Venn diagram shows Fault Injection is also in the Reliability area as it can be used during stress testing. It also helps to educate on what you can do with SAMPLE, such as monitoring alarms while the system is running.
    • Ian will split the table again. {ACTION}
    • Footnotes to add:
      BIST: Economics (as before); Instruments and support tooling.
      Fault Injection: Software debug but enabler for other things; Supplementing diagnostics (Field Service and Workshop Repair)

6. Key Takeaway for today's meeting


7. Schedule next meeting

Next Meeting:
January 28.

February schedule:
11 - Ian will miss.

8. Any other business


9. Review new action items

  • Ian: Split the two Use Cases covered today onto a new slide and add annotations.

10. Adjourn

Tim moved to adjourn at 12:06 PM EST, seconded by Brad.

Respectfully submitted,
Ian McIntosh