Software patents and disclosure

Techdirt has a useful article about how useful software patents actually are.

He points out that :
- companies will usually only patent stuff that would get disclosed anyway, relying on trade secret protection for the rest;
- Microsoft tells their employees to "never search, view, or speculate about patents", partly due to the worries over "willful infringement" and partly because you wouldn't learn anything from them anyway.

So if software patents aren't read by practitioners, don't disclose anything that wouldn't be disclosed anyway, and are written so as to be incomprehensible to those "skilled in the art", how can they possibly be a good thing ?

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Not so useful article

So many things wrong here, it's hard to know where to start.

So if software patents aren't read by practitioners,

You are basing this on a single data point from a single company. Hardly much of a trendline. In fact, they are read by practitioners. I've helped some of them find the patents they were looking for.

don't disclose anything that wouldn't be disclosed anyway,

Purely opinion, not backed up by any kind of facts or statistics. And in my experience, completely incorrect. There are many reasons why companies seek patent protection on inventions that may or may not be disclosed or independently discovered.

and are written so as to be incomprehensible to those "skilled in the art",

Some patents can be very difficult to understand, even by someone who knows the technology. If the inventor can't recognize their own invention, I would say the patent was not well written. They are not often written this way, particularly not the description and preferred embodiment sections. The claims - sure, leave those to the lawyers.

I disagree

You are basing this on a single data point from a single company

Well, they do happen to be the largest software company in the world. And my own personal experience is that it's very rare for somebody writing software to read patents unless they are in the process of filing one. So I'm not actually basing that on just one data point.

Some patents can be very difficult to understand, even by someone who knows the technology. If the inventor can't recognize their own invention, I would say the patent was not well written. They are not often written this way, particularly not the description and preferred embodiment sections.

In my experience, having read quite a few patents and filed for one myself, I'd say that the majority I read were pretty incomprehensible. They tend to either be so broad as to be ridiculous or so weirdly worded that you have no idea what they're talking about.

Do you write software for a living ?

Some do, some don't...

It's still just one person at one company. In my experience, directors and lawyers tend to be very conservative on this point for legal reasons. If the 'official' advice or policy is to avoid patents, then they aren't likely to be found guilty of willful infringement if it comes out that someone at the company happened to read the patent in question.

I have read dozens or hundreds of patents in detail in my career, and many thousands in less detail. It's a common misconception that most patents are 'so broad as to be ridiculous'. The bulk of a patent is directed at describing the current state of the art, or the preferred implementation of the invention. This implementation includes a lot of matter which is not patentable - it's only the unique part which is novel.

Legally, the most important part is the claims, and these cannot be easily and correctly interpreted without some background and training in patents. They are also the least important part for anyone concerned with actually understanding the invention.

I have also filed a few patents, in Canada, the US, and Europe, I believe. I think my first is due to be published any time now. Yes, I write software for a living.

Here is an important point that I think you've missed. When a company files for patent protection, they have disclosed how their technology works. They cannot rely on trade secrets anymore. Their employees are free to publish papers on it, discuss it in public, etc. Similarly, the company is free to use it in marketing collateral for their products. There's no point in trying to keep it a secret, and so can't. Even if no-one ever reads the patent, the fact that it is disclosed has let the cat out of the bag, so to speak.

Also, patents are used to understand the direction that a company is taking in its research. By understanding in which technology a company appears to be filing patents, we can obtain valuable competitive information on the research efforts taking place in the company. The OECD uses these statistics to infer innovation direction and practices at the national level, and many others use them at the corporate level for a similar purpose. This is actually fairly common, particularly for larger corporations.

What gave you that idea ?

Here is an important point that I think you've missed. When a company files for patent protection, they have disclosed how their technology works. They cannot rely on trade secrets anymore. Their employees are free to publish papers on it, discuss it in public, etc. Similarly, the company is free to use it in marketing collateral for their products. There's no point in trying to keep it a secret, and so can't. Even if no-one ever reads the patent, the fact that it is disclosed has let the cat out of the bag, so to speak.

I'm not sure why you think I missed this, when my original post said

companies will usually only patent stuff that would get disclosed anyway, relying on trade secret protection for the rest;

Actually, I think this is a very important point, because it naturally leads to :
a) patents written to disclose as little as possible, to maintain as much secrecy as possible (meeting the letter of the law as far as "disclosing the invention" is concerned, while doing as much as possible to minimise any competitive disadvantage from actual the disclosure).
b) patents written only for the least innovative stuff, with trade secret protection used for the really novel work (which is less likely to be figured out anyway by the competition).

You missed my point...

Perhaps I didn't explain it well. I was arguing against your claim that 'if no-one reads the patents, what good are they'? While I dispute the first part of that claim, I was trying to illustrate that even if no-one ever read a particular patent in question, they still serve a useful purpose. The inventor is free to write a paper about the invention, discuss it at a conference, sponsor further research, etc. without any fear of disclosing a trade secret. This is a different point than you were trying to make.

a) patents written to disclose as little as possible, to maintain as much secrecy as possible (meeting the letter of the law as far as "disclosing the invention" is concerned, while doing as much as possible to minimise any competitive disadvantage from actual the disclosure).

Based on this logic, is there any particular reason why so-called 'software' patents should be any different from those in any other industry?

b) patents written only for the least innovative stuff, with trade secret protection used for the really novel work (which is less likely to be figured out anyway by the competition).

You still have not presented any real data to back up this claim. It is based on supposition and 'logic', not fact. In my experience, this is factually incorrect.