A few decades ago, scientists at Lawrence Livermore National Laboratory (LLNL) in California developed a computer program called the Numerical Electromagnetic Code, or NEC. This set of algorithms allowed researchers to build computer models of radio antennas so they could make realistic predictions about a given antenna design’s performance. As has happened in many other fields, this computer-aided approach drastically accelerated research and development. Instead of building working prototypes and trying them out on elaborate antenna test ranges, then repeating the whole process dozens of times to arrive at a final product, designers could now test and refine their plans on the computer first, and spot-check the results on the final product. LLNL describes the program this way:
NEC (Numerical Electromagnetic Code), written by Gerald Burke, is a popular antenna modeling code for wire and surface antennas and scatterers. Models can include wires buried in a homogeneous ground, insulated wires and impedance loads. The code is based on the method of moments solution of the electric field integral equation for thin wires and the magnetic field integral equation for closed, conducting surfaces.
Like all good algorithms, NEC has grown over the years. The latest release is version 4.1, and it incorporates a huge array of variables that allow users to model all sorts of antennas over realistic terrain. Whether you’re a cellular phone company, a physicist, or an amateur radio operator, this is a tremendously powerful tool.
As computers became faster, smaller, and cheaper, other programmers extended NEC, porting it to platforms beyond its original mainframe implementation and building user-friendly front-ends for it. Many of these ports and interfaces came from hobbyists who released their code for free, a tradition that continues today – the latest open-source Mac version, for example, is here.
However, if you download one of those excellent open source implementations, or even a commercial antenna design package built on NEC, you may notice a discrepancy in the version number: all of these programs use NEC-2, not the much more sophisticated NEC-4. That’s because LLNL, citing export-control regulations, forbids software developers from including NEC-4 in their products. Instead, anyone who wants to use the more advanced version of the algorithm has to submit an application and pay a fee directly to LLNL. The folks at Livermore will then provide a copy of the NEC-4 code, with the stipulation that the applicant can’t give it to anyone else.
I understand and support the need to keep America’s enemies from exploiting American research, and that’s what export-control laws are supposed to do. But it’s very hard to see how LLNL’s licensing policy accomplishes that. Take a look at the NEC-4 application form (PDF). There’s a lot of legalese telling the applicant that the code comes with no warranty, and that he/she/it is not to redistribute this software. Then there are blanks for the applicant’s information: name, address, phone, email. And that’s it. It’s clear that the most important – probably the only important – part of this form is “attach payment.”
How much? Well, NEC-4 will set you back $300 for an individual or academic institution, $500 for non-U.S. academic institutions, $1,100 for American companies, or $1,500 for non-U.S. companies. Considering that the software was developed on U.S. government grant funding, and that much of the heavy-lifting of interface design and code porting was done by others for free, exactly what is this money for? It’s certainly not for a thorough background check – LLNL isn’t collecting enough information to do one, at least not on the main form for U.S. citizens. You don’t even have to enter your Social Security number.
But wait, there’s a separate “Customer Screening” form (PDF) that non-U.S. entities have to fill out. Let’s take a look at some excerpts from it to see what kind of heavy-duty security screening the foreigners are getting:
Is your facility involved in any of the following activities?
Research on or development, design, manufacture, construction, testing or maintenance of any nuclear explosive device or components or subsystems of such a device?  Yes  No
Will the item(s) requested for export be used in the design, development, production or use of missiles?  Yes  No
Will the item(s) requested for export be used in the design, development, production, stockpiling, or use of chemical or biological weapons?  Yes  No
The rest of the form is in the same vein. I’m pretty sure that even the world’s stupidest terrorist would know not to check “yes” to any of these items. Even if that level of understanding escaped them, or if their mailing address seemed likely to arouse suspicion (e.g. “Hidden Base, Afghanistan”), they could simply find a U.S. citizen willing to fill out the form and forward the CD-ROM, perhaps in exchange for a small bribe.
It’s hard to see how LLNL is accomplishing anything with this silly sham, besides scamming the public out of some money and preventing legitimate users from accessing a useful tool. The cumbersome but pointless licensing process for NEC-4 is actively hindering innovation, as many users who might otherwise be able to contribute to the development of the next generation of antennas can’t afford the extra tax LLNL has chosen to levy on them. Even those who can pay the fee may run up against technical problems, such as the need to compile the code from source. If NEC-4 were available as openly as NEC-2, developers could not only do the compiling for their users, but also improve the code’s integration into antenna modeling packages.
I think it’s high time to channel some of this geek rage toward changing the situation. The first step is to unmask the scam. Who wants to take up a collection to submit a transparently bogus license application (with a valid payment) and blog about the results anonymously, to prove that LLNL isn’t doing background checks? I’ll put up the first $20.