Showing posts with label searching. Show all posts
Showing posts with label searching. Show all posts

Sunday, May 4, 2008

CASSIS - Not Just a Resort Town in the South of France

It's pretty remarkable how quickly the folks at the Patent and Trademark Office have moved the tools for patent searching from print to the Web. Less than ten years ago, most people would work with microfiche, CD ROMs, and usually some dusty old books for patent searching. The Patent and Trademark Depository Libraries were often only place people could get access to patent drawings and specifications, and the full patents often only went back to the 70's. For older patents, people needed to rely on the Official Gazette, which is still published, and shows only an abbreviated description of the invention with a few drawings. That solution proved pretty effective for people who could get to the libraries, but geography posed a pretty major obstacle for many people who didn't live anywhere near one of the 90 or so Depository libraries. (Most states only have one!)

Remote access was quite a boon for independent inventors. An electronic image of every patent is available over the internet; users can search the full text of patents from 1976 onward. And that's only using the tools created by the USPTO. PAT2PDF is a free web-based application that will retrieve patents by number and serve up a printer-friendly PDF. Google is in the process of making the entire set of patents searchable using their proprietary algorithms; what they have accomplished so far is pretty impressive.

This material didn't jump directly from books and fiche to the Internet, though. Like many early digital publishing ventures, there was a transition in which material was sent on discs, in this case DVD ROMs.

This system, called CASSIS (for Computer Aided Something Something Including Something), was a welcome arrival. Once we were set up, we had discs in our cabinets that gave us a 200 year back catalog of full text patents. We could keyword search, we could view the international classification orders, we could print images. With the CASSIS DVD-ROM set and the accompanying software, we finally had a complete set of patent drawings and specifications.

As I mentioned above, though, the images available through CASSIS became, shortly after CASSIS was released, available online. However, we still get the DVD ROMS: 4 discs of registered patents, 4 of patent applications, 4 Official Gazettes, and a handful of others relating to classification orders, trademarks, and other USPTO-related documents.

Is this CASSIS system a relic? This question comes up from time to time on both sides of the reference desk. Librarians, for our part, could use the space that's taken up by the DVD cases. We also have to make sure that everyone in the department knows how to help people use this sort of specialized computer system, or at least make sure that someone who knows how to use it is around. And library patrons sometimes wonder why all of the patent search books and websites insist that they physically come to the library to do a patent search. Isn't it all online? Shouldn't we dump the discs?

Yes to the former, no to the latter. Actually, canning the discs is not an option. As long as the USPTO considers the DVD's to be the official version of patents and applications sent to depository libraries, we have a legal requirement to keep them and make them available to everyone. Patents are an important legal and historical record, and the DVD's are a reliable and stable way, at least for now, for us to maintain an archive. All of the Web-based patent collections have some shortcomings, too -- the USPTO database, for instance, uses an unusual format for patent images. There's also very limited bibliographic data for patents issued before 1976. I wrote a post recently where I discuss Google Patent's problems. PAT2PDF is simply a retrieval tool with no search capabilities.

The shortcomings of these Internet-based patent collections speaks to the former question, as well. Technically speaking, every patent that is available on CASSIS is also available on the Internet. But there are still instances when physically coming into the library is still the best bet for a patent search.

One thing I often use CASSIS for is retrieving a list of patents. When you're doing a preliminary patent search, you often want to take a look at several patents. With the Web-based tools, you can retrieve one at a time. With CASSIS, you can create a list and retrieve them as a batch.

Another thing that CASSIS has going for it is that it's nearly devoid of typographical mistakes. Google is improving, but when correctness and comprehensiveness are important, as they often are in patent searching, it's useful to know that you're using the officially endorsed search tool.

I'll concede that CASSIS probably isn't necessary for most people. I know that I often warn readers about the dangers of a lot of the patent searching techniques that that make the process easier. I don't dislike the technology, and probably 90% of the people I help find patents could do just fine using Google or the USPTO website. I also realize, however, that reliable, accurate information is often more important than convenient access to information, and I'd like to make sure people know how find it, or at least know how to ask me how to find it.

If you would like to take your patent search to the next level, I'll be happy to demonstrate CASSIS here in the Government Information Center. And you don't have to hurry -- I think it'll be around for a while.

Thursday, April 24, 2008

Searching for Trademarks: An Annotated Bibliography

A few people have asked recently about what they need to do to register a mark for a prospective business. Aside from shelling out a few hundred bucks, all a person needs to do is to make sure that nobody else is using the mark in the same way.

Here are websites to help you do that, particularly if you live or work in San Francisco:

TESS (Trademark Electronic Search System) comes from the USPTO and contains every trademark currently registered, applications, and many abandoned trademarks from the past. Keep in mind that not all trademarks have to be registered, and those that are registered don't necessarily have to be registered with the Federal government. Keeping that in mind, this will be the largest database of registered trademarks you'll find.

California Business Search If there's a Corporation, Limited Liability Company, or Limited Partnership in California, it's registered here. That means you can use this to see if a business name is in use.

Fictitious Business Search always has a funny ring to it to me, but it's a good resource. Anyone doing business in San Francisco under a name other than their own must register that name with the county. You can access that registry here. This link is for San Francisco, but most counties have a similar database which you can probably find by typing "fictitious business name" and "(your county here)" into Google.

Allmusic Guide If your business is a band, check here and see if anyone's using the name. You may think that you're the only Harmonica Lewinsky out there. And you may be wrong.

D&B Million Dollar Directory
and Reference USA (library card required for both) I know, a business name isn't necessarily a trademark. But it's wise to know who else is using a name and what they're using it for before you decide on a business name. That way you're less likely to unknowingly infringe.

Finally, by all means do a Google search of your business or product name. Do several, and get clever with it. Almost everything that is sold is sold or advertised online, so that's actually a pretty useful way to find product names.

This is by no means a comprehensive list, but if you get through all of these steps, the chances of your mark begin unique are pretty good. To perform a very serious search, stop by the library and we'll take a look at state trademarks, business directories, and some product catalogs.

Wednesday, April 23, 2008

The Index to the USPC: Don't be Afraid

I usually tense up at the reference desk at when someone wants to know how you perform a preliminary patent search. Telling someone about the Index to the U.S. Patent Classification kind of feels like giving someone bad news.

Those who have never performed a patent search in earnest and those who regularly do this kind of research might wonder why I get this feeling. Is it really that bad?

Of course not. In fact, the Index to the USPC is a wonderful tool: it's well created, operates fairly smoothly, and users who are accustomed to indexes (like librarians and anyone who went to school in the pre-Google era) will probably find it to be very intuitive.

The problem is that keyword searching technology in other fields is so good that it makes people wary of indexes and what we librarians call controlled vocabulary. Often, librarians may scoff at keyword searching because of its obvious (to us, at least) shortcomings -- results are limited to the input of the user, relevancy may take a back seat to quick retrieval, searchers may forget synonyms; in general, we're worried that people are sacrificing accuracy for speed and convenience, perhaps without knowing it. Because index terms are assigned by a human being, not a computer, they are often going to be more comprehensive, more relevant, and more accurate. But they may be a pain to get to.

From the perspective of a non-librarian, however, I can understand that trading a little accuracy for the convenience and relatively flat learning curve of keyword searching might be just fine. I don't always go straight to subject headings when I search the library's catalog; though I greatly appreciate the work my colleagues have done creating a taxonomy for the mountains of information out there, sometimes I don't need to perform a comprehensive search. I just need something, and maybe something quick.

To get back to my point about the Index to the USPC and my unease when getting people started with it, it's that flat versus steep learning curve that tends to bum people out. Google, and, more specifically Google Patent Search, appear to work so well (and as a result create such confidence in users) because they get results. They get results for almost any search, even if you spell your terms wrong. If there isn't an exact match, Google will bring you a bunch of "close enough" matches. This type of research has virtually no learning curve.

Let me get off my soapbox and get back to patent searching. To properly do any kind of patent search beyond retrieving a known patent by number or an inventor's name, you have to rely on classes and subclasses, which you find through the Index. The Index, then, is scary because, unlike other research methods, there's a bit a slope to its learning curve. It's not much, mind you, but this type of searching is different than keyword searching.

There are a couple of reasons why this technique is necessary. Accuracy and comprehensiveness are crucial in patents: mistakes could mean that two people are granted the same patent for something. For the inventor, an incomplete search could mean missing a previous invention that disqualifies their invention from being granted a patent and wasting lots of money and time applying.

The reason keyword searching won't work is because a computer can only look for the words entered by its user. Since there's usually more than one way to describe any one thing, there has to be a way to link words and concepts that mean the same thing to the same patents. The Index is the way.

So don't be scared. It's just different, not much harder. Here's a quick rundown of the process:

  1. Jot down some keywords that describe your invention or the invention you're looking for. Think of nouns and verbs, single words or phrases. Any of these can be included.
  2. Words are listed alphabetically in the Index to the USPC. Click on a letter at the top, then either scroll down or use the "Find" function in your browser to seek out your word.
  3. If you find your word, there will be two numbers next to it. The one on the left is the class, which is a broader category of technology. The one on the right is the subclass, which is a more narrow area. Click those numbers, and it will bring you to a listing of subclasses within a given class.
  4. Click the "P" icon to the left of the subclasses, and you will find a list of patents that fall into that category.
As I mentioned before, the Index is actually quite an accomplishment. The creators knew that there are several different words for any given word or concept, and that they needed an accurate way to link these words to the appropriate group of patents.

For more information about patent searching using the Index, check out this tutorial from the University of Texas. And be free from fear!