PDF and Gerber -- is this the best the industry and professionals who collectively all but created the Digital Revolution can put forth?
It is widely acknowledged that our society has been dramatically impacted by electronics since the PC revolution commenced in the mid-1970s and gained mainstream momentum in the early 1980s. As an electrical engineer, I often find myself bristling with pride in how much my profession and industry have helped society. I'm hard pressed to find any profession that isn't achieving elevated efficiencies as a result of the Digital Revolution.
A substantial behind-the-scenes enabler of our digitally empowered society has been the electronic design automation (EDA) industry. At the risk of dating myself, my hands were often coated with lead after drawing schematics and having laid out a few PCBs on a light board. I was nothing short of amazed when I used my first EDA system in the mid-1980s. I just don't think it is possible to fully appreciate today's EDA tools unless you have designed in a world without them.
In my experience, many designs in those days were actually redesigns. Collections of schematic, PCB layouts, chunks of embedded software, and supporting files were scattered throughout the directories of the organization's networked hard drives. All this data constituted an early but ad hoc design repository. Designs were frequently referenced and augmented as required to satisfy new mission requirements. Design throughput progressed as known-good modules could be augmented with new functionality, and incremental design enhancements were made to improve power consumption, performance, testability, reliability, BOM cost, etc.
It seems to me that the situation today is much the same. However, as companies develop new products, their internal design repository can be extended to the vast array of reference designs made available to the community by device vendors via the web. These reference designs are a mainstay of today's semiconductor vendors. They offer product- and circuit-level application of the vendor's devices. Considerably more complete than the predominantly circuit-level application notes of the past, today's reference designs support the promise of shortened development times and often are optimized for performance and BOM.
There are exceptions, but the primary distribution medium for these designs is a PDF -- essentially a drawing on a piece of electronic paper. Further, PCB design data is frequently provided on what can be thought of as a graphic plot on multiple pieces of electronic paper -- a Gerber file. Is this the best the industry and professionals who collectively all but created the Digital Revolution can put forth? The fact that the industry that elevated the sophistication of today's electronics landscape cannot (or will not) do the same for design data exchange is inexcusable. It should be an embarrassment to the electronics community.
The obvious complacency on electronic data exchange demonstrated by the primary technology providers over the last 20 years wasn't always characteristic of the industry. The idea that EDA tools would usher in an age of design data exchange across vendor independent tools was not lost on the industry's founding companies and key silicon vendors. Collectively, they established the EDIF standard in the early 1980s to support this fundamental vision of electronic design interchange. It is beyond the scope of this missive to explore all the reasons the EDIF standard (last updated in 1996) has failed to support data exchange, but few would argue that it has accomplished its primary mission.
When one looks to other design engineering domains, such as mechanical CAD, one sees a legacy of consistently increasing levels of design data exchange, starting with IGES and moving towards today's STEP, which is now an ISO standard. EDIF is now an IEC standard, but such a consistently improving data exchange scenario across tools is absent from the EDA industry's history. Other domains -- such as word processing, database/IT, graphics, and the web -- are admittedly less complex than electronics design, but they have all demonstrated a corresponding attentiveness to data exchange.
Vendor independent electronic design exchange remains elusive, and the situation is not consistent with what the electronics industry has accomplished as a whole. As competitive pressures continue to up the ante for design cycle efficiency, the community of OEMs, silicon vendors, and design engineers deserve, require, and (should) demand better.
In my next column, I will outline what I see as the reasons for today's lamentable state of design data exchange, as well as a potential solution. In the meantime, I welcome your thoughts.