Russ Allen writes:
> I was searching for a patent issued on June 9, 1925. That date is
> on the thumb screw of a broken router plane I bought. My original
> to recast a complete router was put on hold while patent search fever took
[snip-- another galoot bitten by the patent bug ]
> The next step was to determine the first and last
> patent issued on that date. I used the chart of first patents by year on
> the mwtca.org site. I calculated the mid point in 1925 and began searches
> by patent number.
Actually, one of the more useful pages on the DATAMP site is the
first-and-last patent pages. We calculate the first and last patents of
each day/month/year based on the patents we have in our DB. It is a very
useful starting point for starting these kinds of searches.
We have a lot better coverage for pre-1900 patents than for the later ones,
but it is still a useful thing for later patents.
> The next step? Get a cable modem and write a perl script to retrieve a
> thousand patents per day.
These already exist! Ace DATAMP steward Jeff Joslin wrote a Perl script
that processes a list of patent numbers and downloads the first page of the
specification. It's what he used for most of his data collecting. Since I
am a Python and Linux kind guy, I took Jeff's ideas and created a more
generalized Python script to do patent downloads. In addition to the
picture download, my script can create a PDF for an arbitrary patent (using
some open source graphics tools), and generate shell XML for DATAMP upload.
Both of these are pretty straightforward to use. Jeff's are simpler and
more robust, while mine are more for hard-core computer doinks. These will
eventually be moved to the datamp code repository, but for now are available
from the authors.
And now, a plug for DATAMP...
We are always looking for folks to do data collection-- if you'd like to
have the kind of fun Russ is talking about, come on down to the datamp
mailing list, and let us help you get started! It's pretty easy to do, and
absolutely fascinating to do this kind of research.
Check out the datamp mailing list at:
We are also looking for developer-types to help us enhance the site and
create new steward tools. If you've got PHP, Perl, Python, MySql, or
XML/XSL experience and want to help us develop patent searching tools, drop
me a line...
Come join the patent cabal!
Ralph Brendler, Chicago, IL - OTLM, ENB, FOYBIPO
"Science works even if you don't believe in it..." - Penn Jillette