Hacker News Reader

Comment by AlbertCory | Hacker News Reader

Me: software engr, retired; also patent agent who worked in Google Patent Litigation.

A rookie mistake is reading the TEXT of a patent. A pro looks only at the claims. The text is only there to "support" the claims, which boils down to using the same words. Cynical, I know.

Claims unfortunately look like gibberish if you're unschooled, but "construing" them is an early step in a trial, and often if you get a bad claim construction you give up. I worked on tons of claim construction briefs.

I'll read those claims a little later this evening.

View on HN

> I'll read those claims a little later this evening.

I'll save everyone the trouble and tell you without reading them: they're bullshit and so are most software/computer patents. We really need to come together as an industry and fight this crap. Huge props to SparkFun for taking on these bottom feeders.

The paper in [1], which I wrote in 2014, was rejected (in a slightly different version) by CACM. The referee was a Microsoft researcher, who knew nothing about patents and relied on some "expert" who told him "the patent law has changed." This was completely bogus because the law on obviousness hadn't changed at all.

So the big companies, e.g. IBM and Microsoft, are not on our side at all. Neither are ACM and IEEE. I've been on this issue for a long, long time, and it really WILL take a bottom-up revolution. We will get no support whatsoever from the people who should be leading.

[1] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2399580

> A rookie mistake is reading the TEXT of a patent. A pro looks only at the claims. The text is only there to "support" the claims

That's how I have always looked at patents...until I ran into an attorney who insisted that you have to read both and that the text is just as important. The context was in reviewing patents for prior art. This confused the hell out of me because very often the text is where you see someone throw the kitchen sink at the patent. In other words, you often run into things in the text that are not really being claimed.

Not sure who's right any more. Color me confused.

Is the difference between the two cases one where, in the first case, one is using all of the patent (text and claims) as a source of prior art (the text being like a magazine article where someone discusses a bunch of ideas) vs. the second case, where one is looking for infringement of an existing patent based on what is claimed?

Bingo. You've got it. I was talking about SparkFun trying to determine if he's infringing. When we look at a patent that way, we care mainly what the claim says.

On the other hand, every word of everything ever published can be used as prior art. So if SparkFun was trying to invalidate the patent, they would look for anything published anywhere, including patents but also journals, books, etc.

Your comment about the kitchen sink is also true -- when you're trying to get the patent through the PTO, you may want to add claims later, and you cannot change the spec. So you put everything in the spec you might want to claim later.

Does someone offer a comprehensive "Patent it yourself" course?

This would be very useful, not necessarily to actually write patents but rather to truly understand the subject. It could save entrepreneurs time and money and, yes, some could actually use the knowledge to write their own patents.

Patents are very expensive. The small entrepreneur often has to choose between protecting IP and having cash to run the business. Spending $15K to $25K on a patent is out of the question for most small businesses. In that context, the idea of writing multiple patents is nothing less than fantasy.

Such a course, if well done, could easily be worth a few thousand dollars. It would also help equalize the disparity of IP "power" between large corporations and the little guy. When someone has the financials to write hundreds of patents per year, while the other side can barely (financially) approach one every five years, that creates a serious asymmetry that is hard to resolve.

Nolo Press, which is known for putting out do-it-yourself legal books, has a book for writing one's own patents [0]. At the link, there's a sample chapter.

They also have a number of pages on their website related to this topic, which might be worth reading first[1].

[0] https://store.nolo.com/products/patent-it-yourself-pat.html

[1] https://www.nolo.com/legal-encyclopedia/patent-law

The text is the authority on what the phrases in the claims mean. So, you can be surprised when the text sets out a much broader meaning of a term than is used in common parlance.

Ah, if only you'd seen some of the claim construction briefs and opinions I've seen. The text is not the ONLY authority, and if it contradicts common practice then the lawyers can argue about it.

For example, in chemical cases, you cannot have your own definition of "methylated." There are standard definitions of terms. Regrettably, in software not so much.

How did you learn to read the claims?

I took the Kayten course [1] to review for the Patent Bar exam. It was three days, and then I spent a couple months reviewing all the old exam questions.

[1] https://passingpatentbarwithprg.tumblr.com/post/9992218513/w...

Note that the patent bar exam tests you on knowledge of rules and regulations of the patent office and of the Patent Cooperation Treaty that many countries have signed on to. So you won't necessarily learn to draft a patent or how to read a patent just by studying materials for the patent bar exam.

True. Like anything, you only learn the REAL practice of it by doing it. In my case, I found a lawyer who made a practice of taking novices & training them (then they'd eventually go on to higher-paying jobs).

Reply on news.ycombinator.com