Are we privatizing test and measurement standards?
AXIe, IVI, LXI, PXI, and VXI.
They have two things in common: They each manage standards for the test and measurement industry, and each of those standards is ruled by a private consortium.
It wasn’t always this way. When Hewlett Packard invented the 8-bit bus that would become known as GPIB (General Purpose Interface Bus), they were quick to bring the standardization effort to the IEEE where it was codified as IEEE-488. Subsequent work on the standard remained in the IEEE, such as the effort to define common commands within IEEE-488.2.
A National Instruments PXI-to-GPIB interface card.
Then, in 1987, something happened. Five test and measurement companies unveiled VXI (VME eXtensions for Instrumentation) as the first open modular instrument standard. Instead of pursuing standardization through the IEEE, the five companies formed a legal consortium, called the VXIbus Consortium, to manage the specifications. Though the IEEE later ratified VXI as IEEE 1155 in 1993, the real discussions occurred within the VXIbus Consortium. Other companies could join the consortium, typically requiring a fee as well as a public statement of endorsement.
Since that time, PXI, IVI, LXI, and AXIe standards have followed that exact same approach (minus any effort for a latter IEEE ratification). Why is this? Right or wrong, it comes down to speed of execution.
In full disclosure, I was intimately involved in the creation of the VXIbus Consortium, as I represented Hewlett Packard in the standard’s creation. After some exploratory meetings between the five vendors, we had created an outline of a specification that seemed achievable. However, we observed the long and rigorous process that IEEE standards were subject to, and had significant doubts about the timeframe to have a final specification, or even if the work so far would be accepted. Without that, it was risky to develop products. Our customers, largely from the aerospace and defense industry, were begging for products from the day the VXIbus was announced, so we had a short window of opportunity. Consequently, we formed a consortium dedicated to completing the specifications as quickly as possible.
This was the test and measurement industry’s first consortium, and it proved remarkably successful. Other similar standards, listed above, adopted the same approach.
But it isn’t the only approach. In contrast, IEEE 1149.x, commonly known as JTAG or boundary scan, relies heavily on the JTAG working group within the IEEE. Boundary scan is a very successful set of standards.
Going forward, which way would I choose? As someone who has helped create two of the above standards (VXI and AXIe), I have an admitted bias towards the consortium model. I know it works and can achieve results very quickly. It is a model that I am very familiar with. However, the JTAG success gives me pause. Perhaps it is equally effective.
Readers, any comments?