Tuesday, December 15, 2009


Dissecting a patent

Around two and a half years ago, the U.S. Supreme Court unanimously limited patents for the “predictable results” of “ordinary innovation.” At the time, my comment on that spoke full agreement:

I’m thrilled to see this decision. I’ve personally reviewed many computer software patents that I’ve considered to be obvious extensions or combinations of existing technology, and I’ve shaken my head and said, “Geez, who decided that this patent should be issued?” Maybe there’ll be less of that now, and maybe we’ll actually have to do some real innovation to get something patentable.

Alas, no. Two and a half years later, the U.S. Patent and Trademark Office is still issuing patents for the predictable results of ordinary innovation, obvious combinations of existing technology. The IBM Research blog has just highlighted a recent example. Of course, they’re presenting it as an example of an innovative patent by an IBM researcher... and I don’t mean to criticize the researcher, here: the system described in the patent is a useful one, the pressure to patent is high in large companies, and this is a well written patent.

It’s just not a particularly striking bit of innovation.

U.S. patent 7,610,187 has just one independent claim, a claim to the following method (paraphrased here):

  1. A client requests content from a content feed (such as RSS or Atom).
  2. The system identifies the client’s desired language (perhaps from a tag in the request, or perhaps it’s been pre-configured).
  3. The system retrieves the content feed.
  4. The system translates the content from its language into the desired language.
  5. The system gives the translated feed to a feed aggregator.

From my point of view as a user of this, I would subscribe to a feed without regard to its language (perhaps I tell my feed reader to include a news feed from Le Monde). When I see the items in my feed reader, they’ll have already been translated into English for me. Machine-translated, of course, and I have to be aware of the limitations of that, but it’s still pretty useful.

Let’s look at what’s new here.

Machine translation, of course, has been around for a long time — decades, at some level, and on the web since pretty close to its beginning. Babel Fish was the earliest one on the web that I’m aware of; there are many more now.

Proxies that manipulate data have been around for a long time as well, also decades. Just looking at it from the point of view of web servers and services, we can find proxies that filter objectionable content, proxies that reformat content for different devices (make pictures smaller and rearrange layout to send to your mobile phone, for example), proxies that turn one data encoding into another, and so on. That’s not new.

In fact, here’s a patent by AT&T from 1999, U.S. patent 5,875,422, “Automatic language translation technique for use in a telecommunications network,” which does exactly this, in different words. It covers speech or text; communication by telephone, computer, or fax; addressing through telephone number or IP address; language preferences obtained in various ways; and so on.

What’s new in the IBM patent are the content feed and the feed aggregator (reader).

Is that significant innovation? Or is that “ordinary innovation”, showing “predicable results” of normal technology development? It’s clear to me that it’s the latter: it’s just a matter of where the data’s coming from and where it’s going after it’s translated. What’s stopped us from adopting this sort of thing before now is the quality and reliability of the translation, not the lack of innovation in putting the pieces together.

There are natural-language-translating web proxies available, which will basically behave as though you got the web pages and then told Google to translate them for you. Such a web proxy would now, if it were used to retrieve an RSS feed, be infringing on this patent. Of course, before this patent was issued, that same proxy infringed on the AT&T patent from 1999. And so it goes.

What this says is that the 2007 Supreme Court decision has had no effect on the U.S. Patent and Trademark office. The decision’s effect will be limited to court cases — if, for example, someone should challenge this new patent on the basis that it merely patented ordinary innovation — but these sorts of patents will continue to be issued, and their existence will continue to frighten away those with other, novel ideas, who are unwilling to risk being on the wrong side of a legal notice.

That’s not what the patent system was meant for.

1 comment:

Anonymous said...

I agree with this 100%. Slight variations of wording should not classify something as patentable-new.