I need to comment on the latest news and editorials from some of my colleagues surrounding the Electronic Design Automation Consortium (EDAC) because I feel strongly about the EDA industry and I have voiced my opinions about EDAC many times and in many venues.
It is not good news when the institution that is supposed to represent the EDA industry receives negative criticism, even when the criticism is meant to be constructive, because this means that it has failed in some ways. It all started with Peggy Aycinena's report of the EDAC Phil Kaufman award dinner. You can read it here. Peggy tells it like it is and can get to the true feelings as few industry editors can. Her description of the somewhat undisciplined behavior of the attendees, coupled with the mystifying speech by Aart de Geus, Synopsys CEO and Chairman of EDAC, read more like a frat party report than the most prestigious gathering of the year for EDAC.
Brian Fuller, editor-in-chief of EE Times )a sister publication of this web site from CMP Media), followed with an editorial that questions EDAC leadership capabilities when judged by Aart's keynote speech at the same event. You can read Brian's viewpoint here.
In it Brian questions how EDAC can promote the industry when it sees it as the doormat of the semiconductor industry.
And, to make things worse for the consortium, Michael Santarini resurrected his accusation that EDAC is fudging the numbers on its Market Statistics Services (MSS) publication by including IP revenues into the product mix. His article, published by EDN questions EDAC's intent in expanding the report to include IP vendors in the EDA market. In the feedback loop on the same page, David Maliniak, EDA editor at Electronic Design Magazine, attempts to provide a logical technical argument in support to Santarini's argument.
I did not attend the Phil Kaufman award dinner. This year, for the first time in many years, I did not receive an invitation, a likely indication that things are not as organized in EDAC offices as they once were. First of all I do not understand the reason for choosing the event to announce Pam Parrish "retirement". Pam has served EDAC for many years providing an able voice in communicating EDAC's message. And, although her replacement had already been chosen and he had accepted, Aart did not find it appropriate to announce it during his speech. The result was that those present heard that Pam was leaving and that a search was ongoing to find a replacement: thus there would be a void for some time. Instead, Bob Gardner, an industry veteran and for some years the volunteer treasurer of EDAC, had already agreed to replace Pam and his appointment was announced the following week. Read the coverage here. The fact that the news release announcing his appointment is not yet posted on the EDAC web site, or that Pam's smiling face is still above the "Executive Director" label on the Consortium Staff pages, can be taken as another indication of the ineffectiveness of the organization that does not seem capable of even updating a web page in a timely manner.
Of course we must allow the new executive director time to formulate the inevitable new message and goals for EDAC. What is obvious is that the EDAC board of directors went to great pain to choose a reliable team player for the position, so not to obfuscate the messages from the CEOs of the "big 3", or, in any way, indicate that there might be a different way to understand and implement EDAC's mission. But one of the most pressing action items facing Mr. Gardner will be to explain the contents of the MSS, which, following the demise of the Dataquest industry report (see this article), is now the only existing publication to offer an analysis of the industry.
Personally I support the inclusion of IP revenues within the EDA market. I do not share the views of David Maliniak that IP blocks are equivalent to electronic components. In fact many EDA tools, especially in the areas of ESL and verification, have IP blocks in them. Some of the IP is simply executable, but a significant percentage is in fact implementable and results in hardware that is included in the final product. To follow David's logic would mean that all scan tools, for example, should be excluded since they produce scan test logic which is a "component". And, for the same reasons, I think the Michael's arguments are invalid. One can hardly find any examples of SoC devices built without re-using IP. And EDA must include all tools and methods used to produce an SoC, including software by the way. I agree, it is difficult to do historical comparisons since the market is changing. But without change, death occurs. Unfortunately, both Aart's speech and Michael's wishes, seem to indicate that death is easier than continued growth of the industry.