« Angel investors deserve respect from VCs | Main | Top quartile VCs make most of the money »

July 19, 2007


Andy Freeman

Suppose that I invent a mechanism that makes an x86 processor run 10% faster.

What is the "fair market value"?

One way to determine the answer is to sell two processors, one with said mechanism and the other without and look at the difference in revenue.

That's an expensive decision procedure. However, there's a cheap approximation. If Intel (for example), doesn't want to pay me what I think that said mechanism is worth, they can simply continue to make processors that don't use said mechanism.

Unfortunately, that's exactly the sort of thing that "patent reform" is meant to stop.

Don Dodge

One reader had trouble posting a comment here, but sent it to me via email. Here it is;

My main gripe with software patents is that they fail to disclose the invention in a meaningful way. I remember being floored when a chemical engineer told me she regularly read patents in her field to keep up on what she could and couldn't do.

I asked about the liability for knowing infringement, and she said that they could understand the patent, decide if they could invent around, or license.

I think a lot of the issues with software patents stem from no one really knowing what they cover until you've spent millions in court.


I agree. Lawyers write patent descriptions so that they are broad, anticipate all possible future uses, and are left open to wide interpretation. Most of the software patents I have read have left me baffled as to what it actually does.

Don Dodge

Andy, In most cases you could license the technology to competitors to get an idea of what it is worth.

If an inventor patents something that is only applicable to one proprietary product...it makes it pretty hard to get a market value.

Even in this case the inventor could create a list of ALL the features of the product, assign a percentage of the sales price to each feature, and come up with a theoretical license value. This is what the courts should do with patent damage estimates.

Andy Freeman

I picked my example because there's only one feature - 10% faster AND that the value of that feature is open to considerable debate. Heck, whether the mechanism actually delivers 10% is open to debate.

Under "reform", I get to sue intel and negotiate value. Under the current system, I get to tell intel "pay my price or do without". Clearly the latter is better for me.

The "reform" argument is that the current system lets patents on pieces scuttle entire products. Except that that's not necessarily true. It isn't in this case - Intel has "done without" for 20 years so it's absurd to argue that continuing to do so is burdensome.

Note that this is the sort of case that the reformers are using to advocate for reform. Oops.

Ogan Gurel

I have written about - including in the Wall Street Journal - the implications of patent reform for medical technology focusing in particular, on the growing area of convergent medical technologies. Basically, this new patent reform bill will be more in favor of such combinatorial innovation. For a more detailed review of this, please see the article:

"Patent Reform Act of 2007: Innovation, Implications and the American Inventor" which you can find at:


Thanks again for posting on this important topic.

The comments to this entry are closed.

My Photo

Enter your email address:

Delivered by FeedBurner

Twitter Updates

    follow me on Twitter