What technology companies are actually doing is off-shoring electrical engineering jobs.

This decision only accelerates the trend. You lawyers can comfort yourselves with the thought that the Ninth Circuit didn't screw up it's opinion too badly. Instead, the jury didn't understand the technology and reached an obscene result.

Oh, well. They were nice jobs while they lasted.

This further explication is illuminating. After rereading the opinion, I have a few questions about Altera's claims:

1) The opinion starts: "The case involves an infringement action". The section on copyright misuse reads: "there has been no allegation of copyright infringement". Is the distinction there between copyright in fringement and SCPA infringement? Or between infringement action and infringement claim?

2) The initial description of the Altera's business model implies that "the customer, working with Altera" designs the chip. The rest of the opinion claims that Altera owns the mask work represted in the bitstream. Is this because the particular claim involves cases where Altera did all the work in designing the chip? Or is Altera claiming ownership of any mask work produced with its software?

The bitstream used to program a "Complex Progammable Logic Device" (CPLD) doesn't represent the mask work of the CPLD. The bitstream isn't "etched" into the CPLD any more than a copy of Microsoft Word is "etched" into an Intel Pentium IV processor. These are programmable devices. The bitstream is the software used to program them.

See Altera's CPLD Device Family Overview.

ee is right; Altera's process doesn't etch anything. That's what makes their products, as you say, programmable. As to that detail, then, my bad; I was trying to follow the court, which does a poor job of explaining the technology.

But what does Clear Logic do with the bitstream? My understanding is that it converts the bitstream into instructions for a laser. As I read the court, it interprets the SCPA's definition of "mask work" as embracing those instructions; the fact that the instructions are in the form of a program is significant only in the sense that the program is a particular "fixation" or "encoding" of the mask work [note the bizarre mismatch between 20-year-old legislation and current semiconductor technology].

Who "owns" the work? If I use MS Word to write an essay, MS doesn't own what I write. Copyright law is pretty clear on that. And if the MS EULA said otherwise, a court would strike that down as preempted. The SCPA has a preemption provision comparable to the Copyright Act's preemption provision, but it lacks "ownership" definitions as (relatively) precise as those found in copyright law, and the Altera court really doesn't address ownership at all. It does assume that the work either belongs to Altera (for SCPA purposes) or that customers can't reuse the bitstreams (given the terms of the license).

All that said, I still think that the court comes out right. ee argues: Altera was outsourcing jobs, and CL's competition was limiting that. If CL had won and made Altera's business model unsustainable, would that tend to limit investment in FPGA technology? If so, how is that a good thing?

I wasn't arguing that Altera was outsourcing jobs.

Silicon Valley companies design and build limited production runs using (expensive) CPLDs or FPGAs and then—if they've got a succesful product—convert to (cheaper) ASIC for volume production. If that work can't be done in Silicon Valley at a competitive price, then it can be done in Korea, Taiwan, India and China.

Off-shoring.

I wasn't arguing that Altera was outsourcing jobs.

Companies design and build limited production runs using (expensive) CPLDs or FPGAs and then—if they've got a succesful product—convert to (cheaper) ASIC for volume production. If that work can't be done in Palo Alto, California or Portland, Oregon at a competitive price, then it can be done in Korea, Taiwan, India and China.

Off-shoring.

In terms of the economics -- the cheap(er) side of the process goes overseas -- is there a difference between "off-shoring" (company keeps the jobs in-house) and "outsourcing" (company contracts for the work)? I wasn't trying to mis-label what the company is doing; I was making a point about the relationship between the lawsuit and innovation policy.

This decision moves the design, prototyping and early production process offshore. That's not the cheaper side of the process.

To varying degrees, depending on the end-user industry, much of the work was already contracted out or "outsourced" to electronics engineering firms in Palo Alto and Portland. This decision puts a bullet in those firms.

Altera wants to keep the cost of converting from CPLD to ASIC technology higher, thinking that they'll sell more CPLDs that way. That's debatable. But just keeping the cost high in the Ninth Circuit is indisputably stupid.

What's Altera going to do next? Climb on board the "sue your own customers" business model?

To restate ee's point: I am skeptical that their business model should depend on seeking rent from the FPGA customer who is locked-in due to the difficulty of re-embodying the customer's design work in an ASIC. The customer who programmed the FPGA should be free to reuse the programming labor to create a n ASIC. As to Altera's business, actual features of their FPGA can sell it: for example, their MaxII line supports in-system programmability.

Open data formats can foster competition. If CL won, Altera could expect not just competition from CL, but from other FPGA vendors. With FPGA vendors competing on features, every player has incentive to do invest in the technology and patent advances. This scenario is especially plausible if Altera has large market share and is using the "sole use" provision to hold on to customers.

I understand the point. And as you've phrased the question, there's nothing novel about Altera's plan; it's coercing customers who want choice. But I phrase the question differently. If Altera simply took customer specs, then designed and sold chips, no one would complain about that business model. Because the technology has evolved, Altera has separated those two functions. Design and chips are sold separately; they're linked by the license, not by physics. Altera wants to mimic the economics of the older model in the context of the new one. Should IP law tell Altera that it can't? Is this "new value" that customers have created, or is it "existing value" that Altera should be able to keep? If you like, it's the Microsost browser antitrust case in reverse. Can Microsoft legally integrate browser and OS? Can Altera legally dis-integrate design and production?

Separately, I responded to ee off-blog, but ee hasn't responded back. So I'll post that message here:

ee wrote: "This decision moves the design, prototyping and early production process offshore. That's not the cheaper side of the process."

-- I understand sentence two. How would the outcome of sentence one be different if Altera had lost? If design, etc. is cheaper offshore, and if manufacturing is cheaper offshore, why doesn't all of it go offshore regardless of the lawsuit? I'm not trying to be argumentative; at this point I'm genuinely trying to follow the business model. My intuition was the opposite: Altera now has the ability to exclude competitors on the cheaper side of the process and keep prices high throughout its product line. Isn't that right? If it keeps early stage work in the US at this point, it reduces the profits that it would make if it moved that work offshore. But I would think that the pressure to move that work offshore would be even greater if Altera didn't secure the competitive advantage via the lawsuit. So I'm missing something, obviously.

ee wrote: "Altera wants to keep the cost of converting from CPLD to ASIC technology higher, thinking that they'll sell more CPLDs that way. That's debatable."

-- I follow that.

But just keeping the cost high in the Ninth Circuit is indisputably stupid.

-- I'm unclear here. Why wouldn't Altera use this decision as leverage elsewhere? From a legal standpoint, I would certainly expect that courts in other circuits would follow the Ninth Circuit's reasoning. The license aspect of the case is key to Altera's SCPA argument, though the court doesn't acknowledge that, and the Ninth Circuit's reading of the license is pretty consistent with what other circuits have been saying about software licenses.

AFAICT, the economics of the old model have the customer designing a chip, possibly using 3rd party EDA software, and then shopping around for a semiconductor fabricator. Altera modifies this by combining the EDA and the fabriaction. As long as the combination of IP and contract law allows the Altera software license their EDA software in such a manner, their claim of contract interference is legitimate.

Also of interest is that Altera's site mentions compatibility with 3rd party EDA packages. I'm wondering if future Altera customers will rely on those 3rd party software packages to a greater extent.

So it seems that my question was a right one (the legitimacy of the combination), but it got the history upside down (Altera is combining what was separate, not separating what was combined). Interesting.

To help understand the business models, take a look at Altera's Design Services “Partners” by business model. Particularly notice how many of those companies offer turnkey design services (column 1). Also take a look at Altera's “What Customers Are Saying About Quartus II Software”. Notice how many of those quotes trash Xilinx. Don't necessarily believe them (this is Altera's PR after all), but take away the impression that many companies design with competing tools and use competing devices.

Altera may use the Ninth circuit's decision that they control the use of their software's output to obtain the same result all over the U.S. That's irrelevant to an electronic design services firm in India. Altera might be able to leverage the decision into customs interdiction of products designed in India, but realistically how are they going to know how a third-party designed an imported product? Sue and obtain discovery? Does that mean that anyone anywhere in the world who has ever used an Altera device for prototyping is at risk for a lawsuit if they try to import into the US?

OK: The decision puts design jobs in the U.S. at risk because Altera customers will take their Altera devices to electronic design services overseas for prototyping rather than to electronic design services here. Correct me (again), please, if I've got that wrong.

As an answer to the last question: Basically, yes. If "many companies design with competing tools and use competing," then the industry sounds competitive. Again, if that's true, I don't see the harm from the judgment v. CL.

Just some comments regarding the statement
"Does that mean that anyone anywhere in the world who has ever used an Altera device for prototyping is at risk for a lawsuit if they try to import into the US?"
This should not happen because the IC Designers are creating their design using open sourced Hardware Description Language and these design files are portable to third party EDA Tool to convert to ASIC. The problem with CL is that it is using an output file compiled by Altera Software. If CL has its own software to take in the design files (even if they are prototyped using Altera Device) and generate its own output file to program CL Device, then the problem won't exist at all.