Gregory Aharonian showed some statistics about quality patent numbers, and i
would like to discuss some numbers here. First of all, my intention is not
to defend any patent office, i am from Brazil, my english is not so good,
and my real intention is only to test the methods and the numbers used, so
that i could made some similar study here in brazil specially at high tech
areas as software industry, in my DSc thesis.
1. A registration system
The first table show that the average patents per examiner decline gradually
from 61 to 41.2 (1976-1999) and that at same period the issuance rate stays
steady at 65% so the total of patents examined would be at same period 94 to
63.4 patents examined per examiner. This number shows that the examiners
waste more time examining each patent, so it could be an indicator of an
improving of quality ? !
2.Non patent references
According to Todd Dickinson (www.derwent.com/news/dickinson.html) some USPTO
survey shows that over 2,145 recent computer-related patent applications,
65% cited no non-patent prior art. This number is even more dramatic than
the number around 48% showed in Grge statistics for 1998-9. But ths same
stistics show an gradual decrease in number of zero non patent references.
also, the number of non patent references jumped from 0.34 to 2.29
(1976-1999) that is 600% and at the same period the number of patent
references jumped from 4.81 to 10.72 an 123% increasing. So the numbers show
that non patent reference is growing more faster, even though still a low
number, in face of many other sources specially at software industry.
3. months/claims
The first table shows that the average patents per examiner is declining to
41.2 (1999), so when we look at table about average pedency, that shows
steady around 24 months for each patent, we can conclude that the examiners
are spending more time at each patent because they are becoming more
complex. This is exactly what we see when observe that the average number of
claims at same period jumped from 9.1 to 14.6 (mech). So the number
months/claims could be analysed correlated with patents per examiner. So
dividing 61.0/2.18 (1976, mech) we reach the same level at 1999 (41.2/1.64)
so I think that the number months/claims alone can not be used as a good
indicator of quality, but in conjunction with the average patent per
examiner we see that this quality indicator stays steady !!
a good indicator of the quality of the exam (not the quality of the patents)
could be the number of technical requirements made by the examiner,
discriminating the formal to technical ones?
4.jepson type
The decreasing number of Jepson type claims is surprisingly to me, because
we see that many applications in Europe (EPO) also are abandoning the
traditional "characterized for" by "comprising" so adopting the jepson type
claims !!??
Antonio
________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com
This archive was generated by hypermail 2b29 : Fri Aug 10 2001 - 15:58:23 EDT