The architecture of the iPhone shows that ESL tools fail to provide a comprehensive solution to the problem of system level design.
Hopefully every CEO of an EDA company has by now purchased an iPhone and, in the next 24 hours will also read the two "Under the Hood" articles found in the Product Center section of this web site. They will then be in a much better position to understand the reason that most system designers, editors, and analysts have not shared their enthusiasm for ESL during all these years. Introduced as the next growth segment for EDA, Electronic System Level (ESL) has had a disappointing history.
Although it brought us SystemC, allowed Forte and Coware to attain name recognition in EDA, provided conferences organizers with ample material for inconclusive panels, and offered editors various material for commentary, ESL failed to deliver where it counts. Some remain puzzled to this day, and others, mostly at Forte and Coware, continue to fiddle with the model in search of the pot of gold. Finally, Steve Jobs has shown to all of those who would look the reason for the failure. It took someone with the reputation of Mr. Jobs, who would be President if he could just rename a certain building at 1600 Pennsylvania Avenue "the iHouse", to provide us with an example of a "electronic" system that EDA companies can use to plan their growth in the system market.
To design the iPhone architects had to deal with more than just partitioning the hardware, or even build a mockup of the product using SystemC and TLM (Transaction Level Modeling). They had to model the human interface, the analog portions of the hardware, and they had to evaluate the role software would play in establishing the products parameters, including its size, its response speed, and its "look and feel" and they had to evaluate the mechanical characteristics of the touch screen. No EDA company can deliver such an environment at this point and thus designers are not spending a lot of money buying partial solutions to problems that increase in difficulty practically every month.
Since EDA companies have made most of their profits by selling tools for digital designs, they have had significant problems admitting what I have been saying for years: "The world is analog; digital is only a convenient approximation."
And so in the world of system design, we have seen Synposys, who acquired MAST and Saber when it purchased Avanti, ignore the possibilities the language and simulator offered for analog system design and just be content to use them to maintain its market position in the automotive industry. We have also witnessed the most depressing failure in the standard making world when we failed to develop a Verilog-AMS standard. And I do not see any indication that the industry learned anything from the failure in marketing/engineering collaboration that was the Alta division of Cadence.
So the iPhone presents two interesting conclusions. Either engineers can develop systems in a timely manner without the use of proper system level EDA tools, or Apple could have introduced the product sooner and lost revenue because its engineers did not have the proper tools. The financial success of The Mathworks proves the first assumption incorrect, so how much would have those tools been worth to Apple? Someone in EDA should find out.