The contents of DAC's technical program are as much if not more to blame for its declining attendance than the tools featured on the exhibit floor.
In a viewpoint published in SOC Central here Rick Munden offers his observations about the decline in attendance at DAC, and foresees a similar outcome for other electronics design conferences, like DATE and DesignCon. He points out that there is too much emphasis on IC design and little, or none in the case of DAC, on PCB or FPGA issues. Rick opinionates that the problem rests with the exhibitors who realize far greater profits licensing high priced IC design tools as opposed to FPGA and PCB tools that offer far less profit margins. He also points out that DATE this year dedicated far more time to system design issues than DAC, especially if you define a system as something bigger than an SoC. I agree that one would expect exhibitors to showcase the products they want to market more aggressively, and typically this means products that generate good profit margins and demonstrate technological leadership. In general this means IC design tools.
Yet, if DAC is to continue to be the representative conference of the EDA industry, it must be more receptive to the problems of the entire electronics design community, not just the IC part: I have not yet heard a proposal to change its name to ICDAC, especially since we already have ICCAD, a much smaller conference, I may add.
The problem, of course, is more complex than just exhibitors' profit motive. DAC is really three events in one. You have the exhibit floor, what people often call the "commercial" part of DAC, the "technical" program where academics and researchers show the fruits of their labors and newly graduated doctoral students compete for employment through a papers war, and the promotional part of the conference made up of all the activities (feeding opportunities, cocktails opportunities, dancing opportunities) designed to lure attendees to hear a promotional message specific to the activity sponsor.
It has been a few years since I have attended a promotional event about FPGA or PCB at DAC, in spite of the fact that, contrary to what Rick implies in his article, both FPGA and PCB tools are quite profitable market segments. Cadence, and Mentor in particular, make very good profit margins from their PCB market segments, and Mentor continues to invest in the FPGA sector, a sure sign that it is not a loss leader by any means. And we should not forget Synplicity, who is also a successful publicly traded company that sells mostly into the FPGA design market.
I am convinced that the root of the problem rests with the contents of the Technical Program. Unfortunately for DAC, in the last couple of years at least, the contents of the technical program cater more to researchers than they do to the majority of designers. While the "publish or perish" crowd continues by necessity to focus on exploring the problems facing leading edge IC designers and beyond, the majority of IC designs are being done using less demanding processes to serve consumer markets that require a company to be able to make good profit margins while selling less than one million units of a particular design in less than one year.
So the contents of DAC have become less relevant to the working class designers while continuing to serve exceptionally well its research and academic constituency. Since two of the three DAC sponsors (ACM and IEEE C-EDA) principally serve this community, it may be hard to modify the profile of the Technical Program in the short term. And there are very practical issues that make expanding the technical program difficult. There are only four days, and DAC already has a facilities problem. Only very few conference centers are big enough for DAC today: where can DAC find the room to offer more tracks to cover the design issues not cover today?
I think that the solution rests in new and creative approaches: this may not be just an opportunity for "out of the box" thinking. It may require "out of the planet" thinking. But isn't this just what the EDA community has proven to be good at?