How to Write a DAC Panel Proposal

• What Are the Goals and Structure of Your Proposed DAC Panel?
  • There are basically two types of DAC panels: "educational" and "debate".
  • Independent of which panel type you are proposing, the following generic questions always arise during committee discussions. The successful panel proposal will address these as well as the other issues compiled in this web page.
    • Is the topic of general interest? DAC organizers count the attendance at each session, and track from year to year how well certain topics tend to draw audiences. This is not to say that all DAC panel proposals must be on the latest fad or sexy issue - but audience appeal is a factor.
    • What will the audience come away with when the session is over? Will they have something to use at work the following Monday? The successful panel session will not only draw an audience, but will leave them satisfied that their time was well spent. (Who is your target audience, and does it match the DAC attendee profile?)
    • How are you going to maximize audience impact? Choice of panelists by position or constituency, structure of the session time budget, order of presentations, use of moderated and free (audience) Q&A, etc. are among the organizer's degrees of freedom. Explain in your proposal how these choices are aimed at maximizing success.
    • Along similar lines: What is the anticipated impact of this panel, and of its topic, on the EDA field in general?
    • If certain positions are the domain of "technologists", then have you really populated the panel with the most credible R&D people (as opposed to, say, marketeers)?
    • Do you have real customers and designers on the panel, or is it vendor-dominated? (Depending on the context, vendor-dominated can be all right - but usually some real-world "tales from the trenches" help keep the discussion interesting and relevant.)
  • If you have a "debate" panel:
    • How are you going to avoid the pitfall of "much heat, little light"?
    • It helps to identify the three or four key questions that will focus the discussion.
    • It also helps to identify the various constituencies (e.g., flat vs. hierarchical, synthesis-centric vs. physical-centric, formal vs. ad hoc, ...) and how the panel membership covers the various viewpoints.
  • If you have an "education" panel:
    • How are you going to avoid the pitfall of too much "motherhood and apple pie"? E.g., do you have a "naysayer" who will say that this isn't a critical problem after all, that it's already solved by the EDA customer base, that the trend is a mirage, ...?
    • Are you sure that you've covered all bases, all constituencies?
  • In addressing these issues, it's not a bad idea for your proposal to end up looking somewhat like the example summaries that appear in the Proceedings (see the first three links HERE).
  • Your proposal should also clarify the anticipated structure of the panel.
    • 3/4 of the sessions at DAC are 90 minutes long (the other 1/4 are 120 minutes long). How will you use the time? Following are two "tried and true" scenarios.
    • Scenario 1 begins a typical 90-minute session with a 5-10 minute introduction and context setting (what is the scope of the panel, what are the key questions to be illuminated during the 90 minutes, who are the panelists and what are their qualifications, what constituencies or viewpoints do the panelists represent, etc.). While the moderator can introduce the panelists and scope, sometimes in an "education" panel the first panelist will set the scope (e.g., an academic who flashes the ITRS slide, the laundry list of issues slide, etc. so as to avoid duplication in later talks).
    • Scenario 1 will then have around 7-minute position talks from each of 6 panelists. Counting transition times, this already uses up the first hour of the session. (Some moderators and organizers attempt to identify a few common questions that all panelists must answer -- or otherwise limit the position statements to, say, a maximum of 5 slides.)
    • Scenario 1 will then conclude with 30 minutes of moderated Q&A, often semi-scripted to highlight the distinctions between panelist viewpoints. It is nice to have audience Q&A, but this can be fairly random and unfocused, and is often less effective in terms of value to the audience. (The trend seems to be away from audience Q&A.)
    • In Scenario 2, the moderator leads off with a question which panelists will answer with their positioning statements. From there, the discussion begins. This is a less common organization, but on the occasions that it's been tried, it has worked well.
    • The above are just some possibilities. The main point is that session structure is a necessary component of a panel proposal.

  ---

• Rough Outline of the DAC Panels Process
  • In a given year, DAC will have approximately eight (8) panel sessions. One or two may be "plenary", e.g., the "CEO panel" or the "Embedded Systems panel". The DAC Panels Subcommittee works with the Executive Committee and the TPC, trying to strike the right balance between programmatic needs ("growth in embedded systems"), programmatic balance ("need to have a system-level verification methodology panel"), current interests ("DSM timing closure is a hot topic"), setting direction for the EDA community ("future trends"), and projected audience interest.
  • Panel proposals are due at the same time as other submissions (late October). However, there is much more flexibility with panels than with technical papers, so you should send in your proposals late rather than waiting an extra year. In previous years, Terri Alexander at MP Associates (terri@dac.com, 303-530-4333/-4562) and the DAC Panels Subcommittee Chair have been the appropriate points of contact.
  • Anywhere from 30 to 50 proposals are received in a given year. These are reviewed by a Panels Subcommittee of around 6-8 TPC members, starting Tuesday evening of ICCAD week (early November) and continuing via teleconference, email, and a face-to-face meeting in the Bay Area in mid-December. Depending on issues of balance, quality, programmatic need, etc. the Panels Subcommittee may also solicit new proposals from members of the EDA community.
  • By mid-December, there are typically around 12 to 16 "live" proposals. Each is assigned to a "shepherd" who is a member of the Panels Subcommittee. If you are associated with a "live" proposal (e.g., as organizer/proposer), then you will probably be contacted by the shepherd in late December and early January to refine your proposal. The purpose of this contact is to initiate a process of improvement and definition of your proposal, so that it can be presented to the DAC TPC in its best possible light. Please note that the concept of "your proposal" is somewhat vague. This is because there are often many similar proposals submitted to DAC. (For example, in recent years there are always multiple "perils of UDSM" proposals, "verification methodology" proposals, "new EDA business models" proposals, etc.) The Panels Subcommittee will tend to pick the strongest exemplar in each equivalence class, and use it as the basis for mixing and matching together a stronger proposal. Thus, your proposal may appear to be "rejected", even though essentially the same panel idea ends up in the program. Conversely, your proposal may appear to be "selected", but the reason may be as insignificant as the title or the moderator's name.
  • At the DAC TPC meeting, which typically takes place in the Boulder area on the third Thursday in January, all live proposals are presented by shepherds to the entire TPC. (Because of time constraints, typically each entire proposal is boiled down into 1 or 2 PowerPoint slides. Examples are (last three links) HERE. Feedback is obtained. Based on all inputs (including the number of available session slots), the Panels Subcommittee internally agrees on a recommended slate of panels. As noted above, there are typically around eight panels at this stage. These panels are cleaned up during the last part of January to address any TPC concerns, and to solidify title/abstract and participant data.
  • The cleaned-up slate of panels is subject to approval at the DAC Executive Committee meeting, which typically occurs on the second weekend in February. Until this approval is obtained, nothing is officially in the DAC program.
  • By mid-February, the Advance Program must be finalized. This means that all moderator and panelist names, affiliations and locations must be locked down. Typically, a 60- to 80-word panel abstract is written for use in web and/or printed advance publicity.
  • By the end of March, the Proceedings must be finalized. This means that a 1-2 page double-column summary of the panel (abstract, position statements, etc.) must be sent to the publisher in both softcopy and hardcopy form. Panel organizers work closely with shepherds on this. See the EXAMPLES (first three links) HERE.
  • Well before DAC, it is recommended that the moderator and panelists get together -- not only by email, but face-to-face -- in order to agree on the key issues that will be discussed, as well as the structure of the panel, order of presentations, contrast of position statements, etc.
  • Note that all panel presentations must be approved by DAC in advance. A number of guidelines re appropriate use of logos, level of overt product marketing, etc. must be followed. Organizers and shepherds must work closely with panelists so that presentation materials are available for review in a timely way.
  • Caveat: For DAC-2002, there is a plan to shift the due dates (and thus the schedule) later, e.g., the submission deadline will likely be at the beginning of December, and the rest of the Panels schedule will likely shift accordingly. Stay tuned for updates.

  ---

• Tips for a Successful Proposal and Pleasant DAC Panels Experience
  • The ideal number of panelists is either five (5) -- especially for a 90-minute session -- or perhaps six (6). Seven is too many for a 90-minute session (as noted above, 3/4 of all DAC sessions are 90 minutes long).
  • Please do not fudge the participant list. When you say that someone is "confirmed", they should be. Proposals lose credibility when a committee member contacts the "confirmed" moderator or panelists and finds out that they never heard of the proposal.
  • Please do not attempt to put two members of the same organization onto the panel, e.g., as Moderator and Panelist. "Same organization" encompasses any obvious close business or personal relationship. (Examples: investor/VC, board member, advisor, co-developer, co-marketer, future acquisition/merger, ...) Such relationships will not escape scrutiny, will weaken the proposal, and are an invitation for DAC to start changing your panel composition. The basic guideline is this: a DAC panel is not a commercial or a forum for product announcements, and should not be stacked with any given company's own customers. There are never any "slots" on panels that may be filled at the discretion of a company.
  • Please do not expect to "own" your proposal. DAC owns the panel proposals, and feels free to mix, match, re-spin, etc. according to what is best for the DAC program. The best-case outcome for a proposer/organizer is to still be listed as such (probably jointly with a DAC TPC member) in the program and proceedings.
  • If you are the proposer and have a public relations firm as your affiliation, please do not tell the shepherd that Company X must be on the panel because that is how you get paid. The argument just doesn't have any impact. (And even if it is injected into the committee discussion, someone will say, "Hey, those guys work on retainer anyway".)
  • Please do not "spam" low-quality proposals. Sometimes the Panels Subcommittee will see 5 proposals from the same "organizer", with each proposal having just a title ("EDA and the Third World", "What Your Mother Never Told You About Dealing With EDA Sales Organizations: The Fourth-Quarter Lever", ...) plus about 2 lines of content. Some of these might be good, but no one will ever know: they're likely to be thrown away immediately. Weak proposals also tend to damage viable proposals that come from the same source.
  • Please try to find a good title. Obviously, a title such as "The Purpose of A System-Level Design Language is to Maximize Productivity of Designers Who Are Used to Using MATLAB" won't pass muster... but you'd be surprised how many titles in proposals look like this. A good title is catchy and crisp, and clearly sets out the question or issue that the panel will address. Remember, all that the DAC audience has to go on -- in determining whether to attend a panel -- are the names of the panelists and the description in the program. Only rarely will the audience look to the proceedings for help. If you have some recognizable names, this will help, but in any case, make the title and description of the panel appealing. (Here are some building blocks of the generic DAC panel title: "From the Trenches", "What's Next", "Past Problems and Future Solutions", "Impact on EDA", ... you get the idea. Put a colon in there and you're set!)
  • Please think hard about the "Q" of the moderator and panelists. Of course they don't have to be telegenic, but good stage presence, quick repartee and humor, etc. are all helpful attributes. DAC organizers are concerned about this aspect of the panel, and proposals can lose out based on the perceived verbal strengths or weaknesses of the session participants.
  • Please do not submit a proposal if you are not able and willing to follow through. "Follow through" means (i) rapid turnaround time with your shepherd, (ii) flexibility with respect to panel composition (no attempted pocket-vetos, please), (iii) active help in tracking down position statements and PowerPoint presentations from panelists, (iv) coordinating setup meetings before DAC, etc. It is a lot of work and responsibility.
  • Please do not attempt to go around your shepherd or the DAC Panels Subcommittee in setting the content of, e.g., the Advance Program, the Final Program, or the Proceedings. Because MP Associates and the proceedings publisher are extremely busy during the February to April time frame, it is possible to pull a fast one (e.g., list a phantom panelist or organizer). However, this will be remembered, and your name will be mud for years to come.