Ex Parte Menendez et al

Patent Trial and Appeal Board

May 2, 2017

12650113 (P.T.A.B. May. 2, 2017)

United States Patent and Trademark Office
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O.Box 1450
Alexandria, Virginia 22313-1450
www.uspto.gov
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO.
12/650,113 12/30/2009 Nereida Maria Menendez 51017-84536 6718
21888 7590 05/04/2017
THOMPSON OORT TRN T T P
EXAMINER
ONE US BANK PLAZA VIG, NARESH
SUITE 3500
ST LOUIS, MO 63101 ART UNIT PAPER NUMBER
3622
NOTIFICATION DATE DELIVERY MODE
05/04/2017 ELECTRONIC
Please find below and/or attached an Office communication concerning this application or proceeding.
The time period for reply, if any, is set in the attached communication.
Notice of the Office communication was sent electronically on above-indicated "Notification Date" to the
following e-mail address(es):
IPDOCKET@THOMPSONCOBURN.COM
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte NEREIDA MARIA MENENDEZ,
PAULA S. WILLIAMS,
and MICHAEL J. MANIS
Appeal 2015-002128
Application 12/650,113
Technology Center 3600
Before ANTON W. FETTING, MICHAEL W. KIM, and
NINA L. MEDLOCK, Administrative Patent Judges.
FETTING, Administrative Patent Judge.
DECISION ON APPEAL
STATEMENT OF THE CASE* 1
Appellants seek review under 35 U.S.C. § 134 of a final rejection of
claims 1-9, 11-13, 15-25, 27-32, 34—36, 38-80, 82-84, 90, 92-94, 100, and
1 Our decision will make reference to the Appellants’ Appeal Brief (“App.
Br.,” filed September 2, 2014) and Reply Brief (“Reply Br.,” filed
December 1, 2014), and Supplemental Brief (Suppl. Br.”, filed
September 23, 2016), and the Examiner’s Answer (“Ans.,” mailed October
1, 2014), and Final Action (“Final Act.,” mailed August 7, 2013).
Appeal 2015-002128
Application 12/650,113
104—114, the only claims pending in the application on appeal. We have
jurisdiction over the appeal pursuant to 35 U.S.C. § 6(b).
The Appellants invented a way of completing a rental agreement for
an item or service, such as a vehicle rental service. Specification 1:23—24.
An understanding of the invention can be derived from a reading of
exemplary claim 1, which is reproduced below (bracketed matter and some
paragraphing added).
1. A method for creating online and storing within a
server system a completed electronic rental contract for
engaging a rental vehicle from a rental car company's fleet to
thereby authorize a user to pick up a rental vehicle without the
user visiting a rental counter, the method comprising:
[1] receiving data
from a computer through a website during a visit
by a user of the computer to the website,
the data being representative of a user's future
rental of a rental vehicle;
[2] determining whether the user has a defined pre­
existing customer relationship with the rental car company;
and
[3] performing the following steps
even if the user is determined not to have the
defined pre- existing customer relationship:
[4] communicating an electronic rental proposal
to the computer through the website for
display to the user,
the electronic rental proposal being for a
rental transaction that is at least partially based on
the received data for the future rental;
2
Appeal 2015-002128
Application 12/650,113
[5] receiving an electronic acceptance of the
electronic rental proposal
from the user through the computer and
through the website;
[6] creating the completed electronic rental
contract for the rental transaction
in response to the received electronic
acceptance,
the completed electronic rental contract
resulting in the user being authorized to pick up a
rental vehicle in accordance with the completed
electronic rental contract
without the user visiting a rental
counter;
and
[7] storing the completed electronic rental contract
within the server system;
and
[8] wherein the data receiving step, the determining step,
the performing step, the communicating step, the electronic
acceptance receiving step, and the completed electronic rental
contract creating step are performed by the server system
executing software logic
during a single visit by the user of the computer to
the website.
The Examiner relies upon the following prior art:
Information on Hertz Corporation, archived web pages printed
through www.archive.org (1997—2000) (Hertz).
Avis Rent A Car — Rates and Reservations,
http://www.avis.com/rates_and_reservations/ (last visited March 03,
2000) (Avis).
Hertz Announces New, Elite Levels for #1 Club Gold Members in
the US, 11 July 2000, PRNewswire Association, Inc. (Hertz Gold).
3
Appeal 2015-002128
Application 12/650,113
In addition, the following references are newly cited:
Burch, Rental Car Companies Check Drivers’ Records, Sun
Sentinel,
http://articles.sunsentmel.coni/19930812/business/9301290359 1 cardri
versmotorvehicles, August 12, 1993 (Burch).
Gallian, Assigning Driver’s License Numbers, Mathematics
Magazine, pp. 13—22, v. 64, n. 1, February 1991 (Gallian).
Claims 1-9, 11-13, 15-25, 27-32, 34—36, 38-80, 82-84, 90, 92-94,
100, and 104—114 stand rejected under 35 U.S.C. § 101 as directed to
non—statutory subject matter.
Claims 1-9, 11-13, 15-25, 27-32, 34—36, 38-80, 82-84, 90, 92-94,
100, and 104—114 stand rejected under 35 U.S.C. § 112, second paragraph,
as failing to particularly point out and distinctly claim the invention.
Claims 1-9, 11-13, 15-25, 27-32, 34—36, 38-80, 82-84, 90, 92-94,
100, and 104—114 stand rejected under 35 U.S.C. § 103(a) as unpatentable
over Hertz, Avis, and Hertz Gold.
ISSUES
The issues of statutory subject matter turn primarily on whether the
claims do more than provide conceptual advice on how to rent a car. The
issues of definiteness turn primarily on whether one of ordinary skill in the
art would understand the metes and bounds of the claims. The issues of
obviousness turn primarily on whether it was predictable for a customer to
self-perform data entry, instead of going to a customer service counter for
such entry.
4
Appeal 2015-002128
Application 12/650,113
FACTS PERTINENT TO THE ISSUES
The following enumerated Findings of Fact (FF) are believed to be
supported by a preponderance of the evidence.
Facts Related to the Prior Art
Hertz
01. Hertz is a web site for Hertz, the car rental
company, which provides the contents of the legal requirements
for its agreements and provides screens for customers to enter
rental reservations. Hertz 1—61.
02. Hertz describes its system as allowing a customer
to make, modify, or cancel a reservation after checking rates.
Hertz 27.
03. Hertz describes the use of a customer profile for
entering data into a reservation. Hertz 17.
04. Hertz portrays radio button selection of entry by
customers with existing profiles and general customers. Hertz 36
05. Hertz describes an offer for a rental vehicle for
value containing the material terms of the agreement, and
requesting acceptance by the customer. Hertz 44.
06. Hertz shows separate screens for initially
requesting dates and locations (Hertz 18) and for auto selection
and pricing (Hertz 24) and for additional promotions. Hertz 26.
07. Hertz describes using the master agreement data
for filling in a reservation. Hertz 27.
5
Appeal 2015-002128
Application 12/650,113
Avis
08. Avis describes entry and storage of a vehicle rental
reservation. Avis 1—13.
Hertz Gold
09. Hertz Gold describes offering members an
automatic invitation to move up to #1 Club Gold after completing
four rentals. Hertz Gold 1.
10. “In 1972, Hertz became the first car rental
company to recognize the strategic value of maintaining a
customer profile database, with the introduction of Hertz #1 Club.
The service created a data file for each customer, by maintaining
driver’s license information, home address, car class information
and credit card information for instant recall. For Hertz’
customers; #1 Club helped reduce time spent making reservations
and sped the process when renting a car.” Hertz Gold 1.
11. “After implementing #1 Club service, Hertz
recognized the additional customer benefit to not only keeping an
active profile on customers, but also allowing] customers to
bypass the counter altogether, with the keys and completed rental
agreement ready and waiting for them.” Hertz Gold 2.
Burch
12. Burch describes how rental car companies are
checking driver’s records from their licenses. Burch 1.
6
Appeal 2015-002128
Application 12/650,113
13. Burch describes that the driving record check is
performed online in less than a minute, so as not to delay the
rental transaction. Burch 1.
Gallian
14. Gallian describes how driver license numbers are
assigned. Gallian 13:Title.
15. Gallian describes how driver license numbers are
encoded for possible detection of forgery or errors.
Gallian 13: Introduction.
ANALYSIS
This is one of three applications before us with similar claims to a car
rental system and method. The other two applications are Serial
Nos. 12/650,040 and 09/698,502. The claims in all three applications
generally recite performing an online car rental reservation that results in a
rental contract. The claims in all three applications also recite doing so in a
single session, without a customer having to visit a customer service counter
and without requiring a pre-existing relationship or master rental agreement
between the customer and the car rental company.
Claims 1-9, 11-13, 15-25, 27-32, 34-36, 38-80, 82-84, 90,
92—94, 100 and 104—114 rejected under 35 U.S.C. § 101 as
directed to non—statutory subject matter
The Supreme Court
set forth a framework for distinguishing patents that
claim laws of nature, natural phenomena, and abstract ideas
from those that claim patent-eligible applications of those
concepts. First, [ ] determine whether the claims at issue are
7
Appeal 2015-002128
Application 12/650,113
directed to one of those patent-ineligible concepts. ... If so, we
then ask, “[w]hat else is there in the claims before us? ... To
answer that question, [ ] consider the elements of each claim
both individually and “as an ordered combination” to determine
whether the additional elements “transform the nature of the
claim” into a patent-eligible application. [The Court] described
step two of this analysis as a search for an “‘inventive
concept’”—i.e., an element or combination of elements that is
“sufficient to ensure that the patent in practice amounts to
significantly more than a patent upon the [ineligible concept]
itself.”
Alice Corp., Pty. Ltd. v CLSBankInt’l, 134 S. Ct. 2347, 2355 (2014) (citing
Mayo Collaborative Services v. Prometheus Laboratories, Inc., 132 S. Ct.
1289 (2012)).
To perform this test, we must first determine whether the claims at
issue are directed to a patent-ineligible concept.
Although the Court in Alice made a direct finding as to what the
claims were directed to, we find that this case’s claims themselves and the
Specification provide enough information to inform one as to what they are
directed to.
The preamble to claim 1 recites that it is a method of creating and
storing an electronic rental contract for a rental vehicle. The steps in claim 1
result in creating an electronic rental contract for (??) a user’s future rental
of a rental vehicle. We do not discern any functional language in claim 1,
for example, any verb that would only have been relevant to the realm of
vehicle reservations. The Specification at 1:23 recites that the invention
relates to completing and storing an electronic rental agreement. Thus, all
this evidence shows that claim 1 is directed to completing and storing an
electronic rental agreement for (??) a user’s future rental of a rental vehicle.
8
Appeal 2015-002128
Application 12/650,113
Further, although the words “electronic” and “rental” are used, again, we do
not discern that removing the “electronic” and “rental” terminology
appreciably affects the functional language of independent claim 1. The
“electronic” limitations, such as “computer” and “website,” are generic, and
the “rental” limitations are merely descriptive, and do not affect how any of
the steps are performed. For example, the “receiving ...” step would
function in the same way, whether the data created are “a user’s future rental
of a rental vehicle,” as recited, or some other form of data. Additionally, we
discern that, after removing “rental” limitations, a “user’s future rental of a
rental vehicle” is merely previously gathered data. We find that all
contracts, at some level, must be based on previously gathered data.
Accordingly, we find that independent claim 1 is directed to entering into an
agreement or contract..
It follows from prior Supreme Court cases, and Bilski in particular,
that the claims at issue here are directed to an abstract idea. Like the risk
hedging in Bilski, we find that the concept of entering into a contract (based
on previously gathered information) is a fundamental business practice long
prevalent in our system of commerce. We find also that the creation and use
of contracts is a building block of our legal system. Thus, entering into a
contract, like hedging, is an “abstract idea” beyond the scope of §101. See
Alice Corp. Pty. Ltd., 134 S. Ct. at 2356.
As in Alice Corp. Pty. Ltd., we need not labor to delimit the precise
contours of the “abstract ideas” category in this case. It is enough to
recognize that there is no meaningful distinction in the level of abstraction
between the concept of risk hedging in Bilski and the concept of entering
into a contract at issue here. Both are squarely within the realm of “abstract
9
Appeal 2015-002128
Application 12/650,113
ideas” as the Court has used that term. See Alice Corp. Pty. Ltd., 134 S. Ct.
at 2357. We conclude that the claims at issue are directed to a patent-
ineligible concept.
The introduction of a computer into the claims does not alter the
analysis at Mayo step two.
[T]he mere recitation of a generic computer cannot
transform a patent-ineligible abstract idea into a patent-eligible
invention. Stating an abstract idea “while adding the words
‘apply it’” is not enough for patent eligibility. Nor is limiting
the use of an abstract idea “‘to a particular technological
environment.’” Stating an abstract idea while adding the words
“apply it with a computer” simply combines those two steps,
with the same deficient result. Thus, if a patent’s recitation of a
computer amounts to a mere instruction to “implement [t]” an
abstract idea “on ... a computer,” that addition cannot impart
patent eligibility. This conclusion accords with the preemption
concern that undergirds our §101 jurisprudence. Given the
ubiquity of computers, wholly generic computer
implementation is not generally the sort of “additional
feature[e]” that provides any “practical assurance that the
process is more than a drafting effort designed to monopolize
the [abstract idea] itself.”
Alice Corp. Pty. Ltd., 134 S. Ct. at 2358 (citations omitted). “[T]he
relevant question is whether the claims here do more than simply instruct the
practitioner to implement the abstract idea ... on a generic computer.” Id. at
2359. They do not.
Taking the claim elements separately, the function performed by the
computer at each step of the process is purely conventional. Using a
computer to accept and store data, and make simple “yes” or “no”
determinations, to create a digital file amounts to electronic data query and
retrieval—one of the most basic functions of a computer. All of these
computer functions are well-understood, routine, conventional activities
10
Appeal 2015-002128
Application 12/650,113
previously known to the industry. In short, each step does no more than
require a generic computer to perform generic computer functions.
Considered as an ordered combination, the computer components of
Appellants’ method add nothing that is not already present when the steps
are considered separately. Viewed as a whole, Appellants’ method claims
simply recite the concept of entering into a contract as performed by a
generic computer. The method claims do not, for example, purport to
improve the functioning of the computer itself. Nor do they effect an
improvement in any other technology or technical field. Instead, the claims
at issue amount to nothing significantly more than an instruction to apply the
abstract idea of entering into a contract using some unspecified, generic
computer. Under our precedents, that is not “‘enough’ to transform an
abstract idea into a patent-eligible invention.” See Alice Corp. Pty. Ltd.,
134 S. Ct. at 2360.
As to system claim 24, it is
no different from the method claims in substance. The method
claims recite the abstract idea implemented on a generic
computer; the system claims recite a handful of generic
computer components configured to implement the same idea.
This Court has long “wam[ed] . . . against” interpreting
§ 101“in ways that make patent eligibility ‘depend simply on
the draftsman’s art.’
Id. at 2360.
The dependent claims recite conventional data entry operations and
user interface features or recite the automotive rental context features, which
add nothing to overcome the abstract nature of the claims.
We are not persuaded by Appellants’ argument that claim 1 is directed
toward a particular machine under the “machine-or-transformation” test,
11
Appeal 2015-002128
Application 12/650,113
because claim 1 recites a server system that is specially configured with
software logic to perform recited steps of the method. Reply Br. 4. Alice
Corp. explicitly responded to such an argument. Appellants’ argument, by
its own terms, admits that a general purpose computer is programmed to
perform the steps of the abstract process. Accordingly, Appellants’ method
steps are no more than instructing the general purpose computer to perform
the method steps by expressing the idea in a language the general purpose
computer understands. Absent implementation details in claim 1 itself, we
are unpersuaded that the computer process is any less abstract than the
process itself.
We are not persuaded by Appellants’ argument that
claim 1 does not merely recite a “generic computer” in
combination with an abstract idea. Independent claim 1 recites
not only that “the data receiving step, the determining step, the
performing step, the communicating step, the electronic
acceptance receiving step, and the completed electronic rental
contract creating step are performed by the server system
executing software logic” but further recites that the server
system performs these steps “during a single visit by the user of
the computer to the website”; thus defining a method tied to a
non-generic computer design.
Reply Br. 5. Appellants essentially argue that only part of the abstract idea
is computer implemented. Alice Corp. informs that computer
implementation per se does not turn an abstract process into a concrete
process, and that this is also generally true for any combination where part
of the underlying process is not computer implemented. We are
unpersuaded that independent claim 1 is any different.
We are not persuaded by Appellants’ argument that
[i]n a dramatic improvement on this inefficiency, claim 1
defines a server system that functions better than the cited art
12
Appeal 2015-002128
Application 12/650,113
by removing the technical feature in the cited art which
conditioned the ability to conduct a rental counter bypass rental
transaction on the user providing proof a pre-existing customer
relationship.
Reply Br. 6. Conditioning a process on some criterion is not a technical
step; it is conceptual advice to restrict the conditions in which to perform a
process. In the instant case, the criterion is a convention that is not required
to rent a car; it merely provides marketing opportunities that one might
choose to forgo. FF 6, 11. Appellants have not identified any information,
provided at a customer service counter that could not have been entered over
a computer instead. Going contrary to a convention is a voluntary choice,
and not a technical hurdle.
We are not persuaded by Appellants’ argument that
that there are myriads of machine-based methods falling
outside the scope of claim 1 which are capable of generating an
electronic rental contract authorizing a user to pick up a rental
vehicle without the user visiting a rental counter. . . . Given
that the abstract idea can be performed outside the scope of
claim 1, claim 1 by definition does not pre-empt the abstract
idea identified by the Examiner.
Reply Br. 6. We disagree. That the claims do not preempt all forms of the
abstraction, or may be limited to the abstract idea in the car rental setting,
without visiting a customer service counter, does not make them any less
abstract. See OIP Technologies, Inc. v. Amazon.com, Inc., 788 F.3d 1359,
1360-1361 (2015).
We are not persuaded by Appellants’ argument that the Examiner
cites no evidence that the steps performed were conventional as of the filing
date. Suppl. Br. 3. The steps recite using a computer to accept and store
data, and make simple “yes” or “no” determinations, to create a digital file
13
Appeal 2015-002128
Application 12/650,113
which amounts to electronic data query and retrieval— one of the most basic
functions of a computer.2 To the extent Appellants argue the context and
mental interpretation attached to the data labels, these are of no patentable
significance, because they only exist in the mind of the beholder. See In re
Bernhart, 417 F.2d 1395 (CCPA 1969).
We are not persuaded by Appellants’ argument that the claims recite
performing steps in a single visit, and even if there is no pre-existing
relationship between the customer and the car rental company, and that the
Examiner has failed to give this limitation due weight in considering what
the claims are “directed to.” Suppl. Br. 4. Whether the steps are performed
in a single or plural set of visits is a matter of convention and tactical design,
not technical capacity, and, thus, is subsumed within “entering into a
contract.” Even on a computer, merging two sets of programming steps into
a single step is no more than merging two sets of textual computer code.
Bypassing a customer service counter, in the absence of a pre-existing
relationship between the customer and the car rental company, is, again,
conceptual advice.
We are not persuaded by Appellants’ argument that efficiency is
improved by merging the two code streams. Suppl. Br. 5. This is no more
than conventional computer operation in serially combining program
instructions, and we are unpersuaded that it adds anything substantive to
“entering into a contract” or amounts to “significantly more.”
We are not persuaded by Appellants’ argument that the claims do not
preempt all ways of performing creating a contract. Suppl. Br. 7. That the
2 IBM and Sperry Rand mainframe computers did as much in the 1960’s.
14
Appeal 2015-002128
Application 12/650,113
claims do not preempt all forms of the abstraction, or may be limited to the
abstract idea in the e-commerce setting, does not make them any less
abstract. See OIP Technologies, Inc. v. Amazon.com, Inc., 788 F.3d
at 1360-1361.
We are not persuaded by Appellants’ argument that the claims allow
for improvements realized by the invention. Suppl. Br. 10. Adapting any
conceptual advice does as much. Following the advice, which would have
been known to one of ordinary skill at the time of the invention, that “haste
makes waste” allows for improvement realized by avoiding the hazards of
waste. This is not a case of such advice being applied to, for example,
something as specifically and concretely rooted in computer technology as
synchronizing lip animations, created by computer technology, that
introduces asynchronous behavior to begin with. McRO, Inc. v. Bandai
Namco Games America, Inc., 837 F.3d 1299 (Fed. Cir. 2017). The instant
claims do no more than advise one to avoid a customer service counter,
which is not computer-specific behavior.
We are not persuaded by Appellants’ argument that
there is no evidence in the record that the pre-Internet
world provided for the creation of electronic rental contracts
that permitted counter bypass without leveraging a pre-existing
MRA, much less doing so during a pre-Internet world analog to
a single website visit.
Suppl. Br. 12. The claims do no more than the electronic equivalent
of converting a reservation into a contract, which has been the foundation of
the automobile rental industry since its inception.3 FF 10—11. Although it is
3 This has been a past practice of the automotive rental market since the
1960’s.
15
Appeal 2015-002128
Application 12/650,113
true that before computers, a rental agent had to transcribe from the
reservation to the contract, computers generically transfer data as a matter of
course. Thus, Appellants have not shown sufficient why saying that the
invention allows bypassing a counter is any more than saying that the
invention uses a computer. As to reliance on a master rental agreement,
there is nothing about a master rental agreement that mandates its use. It is
simply a tool that is optional. Choosing not to use the tool is no more than
choosing when to enter the data that would otherwise be stored in such a
tool, and is both “abstract” and does not amount to “significantly more.”
Claim 24 is argued on the same basis as claim 1. Reply Br. 7.
As to claim 15, reciting a selectable option to convert the claim 1
reservation into a contract, we are not persuaded by Appellants’ argument
concerning claim 15’s one touch, two transactions. Reply Br. 13—14. This
is no more than saying a computer stores data that are reusable. Simply
reusing data across plural transactions is no more than abstract conceptual
advice and, again, is not “significantly more.”
Also as to claim 15, we are not persuaded by Appellants’ argument
that this is a technical innovation. Suppl. Br. 8. This limitation only makes
the implication of claim 1 ’s limitation of receiving acceptance being
voluntary more explicit, and we are unpersuaded something voluntary is
technical.
We are not persuaded by Appellants’ argument that, as to claim 18, it
automates license validation. Reply Br. 14—16. Appellants dispute that data
validation is routine. Id. Data validation was too pervasive to seriously
contend its ubiquity. FF 12—15.
16
Appeal 2015-002128
Application 12/650,113
As to claim 107, reciting data to associate the reservation and
contract, and to flag the agreement, we are not persuaded by Appellants’
argument that this is a technical innovation. Suppl. Br. 10. Absent a recited
technical implementation, this is no more than conceptual advice to file the
information together and highlight it, which is not technical.
As to claim 5, reciting prefilling contract data from the reservation,
we are not persuaded by Appellants’ argument that this is a technical
innovation. Reply Br. 12. Absent a recited technical implementation, this is
no more than conceptual advice to copy or otherwise reuse the data, which is
not technical.
The remaining claims are argued in a manner similar to claim 1, or
based on claim 1.
Claims 1-9, 11-13, 15-25, 27-32, 34-36, 38-80, 82-84, 90, 92-94, 100,
and 104—114 rejected under 35 U.S.C. § 112, second paragraph, as failing
to particularly point out and distinctly claim the invention
We are persuaded by Appellants’ argument that the significance of the
determining limitation does not render the claims indefinite. App. Br. 19—
21.
Claims 1-9, 11-13, 15-25, 27-32, 34-36, 38-80, 82-84, 90, 92-94, 100,
and 104—114 rejected under 35 U.S.C. § 103(a) as unpatentable over Hertz,
Avis, and Hertz Gold
The independent claims describe the basic steps in contracting for a
rental car. First the customer provides the information for basic availability,
such as dates and location of pickup and return. The vendor then responds
with what is available, and allows the customer to refine the request within
those parameters, such as class of car and price levels. The customer then
asks for a specific quote, and then, if the customer accepts, the customer
17
Appeal 2015-002128
Application 12/650,113
provides license and credit card information if it is not already of record and
contracts for the rental. The screen prints of the Hertz web site of record are
evidence of this flow. FF l.4
That the prior art describes the above is not in dispute. Instead, what
is disputed is the timing and location of these steps. In particular, the claims
recite that both the “user’s future rental of a rental vehicle” (hereinafter
“reservation”) and contract are made in a single session, and this is done
without the customer going to a customer service counter, and is done
irrespective of whether there is some pre-existing relationship or agreement
between the customer and the car rental company. Appellants argue that the
prior art does not describe these limitations. Appellants also nominally
contend the limitations regarding the reservation process are separated by
some customer data entry steps, and the reservation and contract steps are
separated by customer acceptance, but these arguments appear to be related
to the single session argument as, again, whether Hertz discloses the basic
flow of data is not in dispute.
Before analyzing the specific comparisons between the claims and the
prior art, we will summarize the issues that these claims raise, because of
their number and complexity. First, the claims recite data entry and little
else. Data per se are no more than arbitrary (based on arbitrary coding
conventions) strings of binary digits irrespective of the labels attached, and,
without some explicit functionality dependent on the differences between the
actual labels, the labels are afforded no patentable weight. Absent some
4 We discern that the above flow comports with the common experience of
anyone who has rented a car.
18
Appeal 2015-002128
Application 12/650,113
functional effect, there is no patentable distinction between a master rental
agreement, a reservation, and a contract, as all are printed matter or binary
data. Absent some functional effect, any legal effect attaching to a contract
is a nullity in patent law. What is left is conventional data entry. Second,
we find that it would have been predictable to consolidate data entry even if
the labels were given patentable weight. FF 1—11. Third, we find that it
would have been predictable for a customer to perform the same data entry
that an agent at a customer service counter would perform. FF 10-11.
Fourth, the above-cited prior art describes a promotional program that allows
the customer to do just that and avoid a customer service counter. Id. Fifth,
even if weight were afforded to the label of a master service agreement, we
find that there is nothing about the absence of such an agreement or
relationship that would present a logistical hurdle to entering the data related
to a master agreement and reservation in a single session. FF 1—11. Sixth,
even if the master rental agreement contains data necessary for a rental, we
find that it would have been predictable to enter that same data during the
reservation and contracting phases in the absence of a master rental
agreement. Id. Seventh, we find that it was predictable to have a customer
contract for the car immediately after and in the same session as obtaining a
reservation. Id.
The dependent claims further recite driver license data entry and
validation. We enter Gallian and Burch as new art that shows it was known
to ask for such data and validate it online at the time of the invention.
Even before comparing the claims to the prior art, we discern that
Appellants have a high hurdle in terms of claim breadth and patentable
weight. The instant claims do not recite limitations not described in the
19
Appeal 2015-002128
Application 12/650,113
prior art. Rather, the claims recite not requiring portions of the car rental
experience that are described in the prior art. Removing something known
from a known process, without changing the result of the process, is
invariably predictable, unless the absence of that something is a technical
barrier to the process. This is a corollary to the volition involved in
performing the process and its steps. Nothing in the claims presents a
technical barrier that is overcome by the steps recited. The claimed
procedure is entirely volitional, and is unconstrained by inhibitory structure,
physical properties, or attributes. Equally, the claims do not recite removing
such a technical barrier. Instead, Appellants posit an industry convention of
using a customer service counter and separating the reservation and contract
in time, and it is this separation that they suggest circumventing. As a
convention, there is nothing about using a service counter that makes
avoiding it atypical. One of ordinary skill would have known that such a
modification has readily apparent costs and benefits, and using a service
counter or not using a service counter is simply based on weighing those
readily apparent costs and benefits. KSR Int’l Co. v. Teleflex Inc., 550 U.S.
398, 421 (2007) (“A person of ordinary skill is also a person of ordinary
creativity, not an automaton.”); Winner v. Wang, 202 F.3d 1340, 1349 (Fed.
Cir. 2000) (“the district court did not clearly err in finding that one of
ordinary skill in the art would not have reasonably elected trading the benefit
of security for that of convenience. Trade-offs often concern what is
feasible, not what is, on balance, desirable. Motivation to combine requires
the latter.”)..
Appellants have not identified any technological hurdle in avoiding a
customer service counter. The claims do not recite any action that would
20
Appeal 2015-002128
Application 12/650,113
ordinarily otherwise require, as contrasted with prefer, such a counter.
Appellants have not identified any technological hurdle to performing two
sets of data entry in a single session. The claims do not recite any action that
would ordinarily otherwise require, as contrasted with prefer, plural
sessions. Appellants have not identified any technological hurdle to
performing a second set of data entry in the same session as a first,
irrespective of the existence of some relationship between the customer and
the car rental company. The claims do not recite any action that would
ordinarily otherwise require, as contrasted with preferr, premising a second
data entry on the existence or lack thereof of such a relationship.
As to the issue of adding data, we find that data editing is a
notoriously well-known mechanism where data, that were not there prior to
the editing, are added; this is not a significant issue. More to the point, the
distinct items of data that must be in place for a rental contract, but not
necessarily for a reservation, are a credit card and driver license. FF 6, 10—
11. We find that adding such data, after seeing that a rental is available in
the form of a reservation, would not only have been predictable, but would
have been immediately envisaged by one of ordinary skill. FF 10.
Not only would have avoiding a counter been predictable, HertzGold
shows it was known to combine a reservation and contract in a single visit,
and that avoiding such a counter was preferable. FF 10-11. Appellants
counter that HertzGold requires a master agreement. But a master
agreement is just data collected prior to a rental, as is a reservation. So, the
issue devolves to the patentable weight afforded any distinction between
data gathered prior to rental and data gathered at rental time at a customer
service counter, and the predictability of combining such data gathering.
21
Appeal 2015-002128
Application 12/650,113
As to structural inventions, such claims must be distinguished from
the prior art in terms of structure rather than function, see, e.g., In re
Schreiber, 128 F.3d 1473, 1477—78 (Fed. Cir. 1997). In order to satisfy the
functional limitations in an apparatus claim, however, the prior art apparatus
as disclosed must be capable of performing the claimed function. Id. at
1478. When the functional language is associated with programming or
some other structure required to perform the function, that programming or
structure must be present in order to meet the claim limitation. Typhoon
Touch Techs., Inc. v. Dell, Inc., 659 F.3d 1376, 1380 (Fed. Cir. 2011)
(discussing Microprocessor Enhancement Corp. v. Texas Instruments, Inc.,
520 F.3d 1367 (Fed. Cir. 2008)). In some circumstances, generic structural
disclosures may be sufficient to meet the functional requirements, see Ergo
Licensing, LLCv. CareFusion 303, Inc., 673 F.3d 1361, 1364 (Fed. Cir.
2012) (citing Telcordia Techs., Inc. v. Cisco Sys., Inc., 612 F.3d 1365, 1376—
77 (Fed. Cir. 2010)).
Also, a structural invention is not distinguished by the work product it
operates upon, such as data in a computer. “[Expressions relating the
apparatus to contents thereof during an intended operation are of no
significance in determining patentability of the apparatus claim.” Ex parte
Thibault, 164 USPQ 666, 667 (Bd. App. 1969). Furthermore, “inclusion of
material or article worked upon by a structure being claimed does not impart
patentability to the claims.” In re Otto, 312 F.2d 937, 940 (CCPA 1963).
As to process claims, claim 1 recites seven steps, viz., receive X,
determine whether Y is true, irrespective of whether Y is true, communicate
Z, receive W, and create and store U, where X is labeled as data being
representative of a user's future rental of a rental vehicle, Y is labeled as the
22
Appeal 2015-002128
Application 12/650,113
user has a defined pre-existing customer relationship with the rental car
company, Z is labeled as an electronic rental proposal, and W is labeled as a
completed electronic rental contract. Thus, the claim is really to accepting
data and soliciting additional data, creating data and determining whether
some criterion of data existence is met, and accepting, creating, and
displaying more data. Nothing in the claim depends on or enforces the
perceptual labels the claim suggests. Mental perceptions of what data
represent are non—functional and given no weight. King Pharm., Inc. v. Eon
Labs, Inc., 616 F.3d 1267, 1279 (Fed. Cir. 2010) (“[T]he relevant question is
whether ‘there exists any new and unobvious functional relationship
between the printed matter and the substrate.’”) (citations omitted). See also
In re Lowry, 32 F.3d 1579, 1583 (Fed.Cir.1994) (describing printed matter
as “useful and intelligible only to the human mind”) (quoting In re Bernhart,
417 F.2d 1395, 1399 (CCPA 1969)). Data labels are just examples of such
mental perceptions. Data, being a succession of binary digits, are just those
digits, not perceptual labels of those digits. The binary digits, at times, may
impose some functional consequence, but absent some recitation of how so,
such consequence is not an issue.
To put this in terms of the test put forward in King, nothing in the
steps depends on or enforces the perception of the data labels and the data
label perceptions do not functionally affect the steps. The Federal Circuit in
King explicitly urged against non-functional limitations allowing the
indefinite patenting of existing known processes by merely putting those
processes in the context of such non-functional limitations. King Pharm.,
616 F.3d at 1279.
23
Appeal 2015-002128
Application 12/650,113
In particular, there is no patentable distinction between a master
agreement, a rental reservation, and a rental contract. We are not persuaded
by Appellants’ argument that there are such distinctions. App. Br. 15—16.
Although there may be legal distinctions between a reservation and a
contract, absent some functional effect, which Appellants have not
persuasively identified here, any such legal distinctions arise only in the
mind of the beholder, and are relegated to the legal arts, as contrasted with
the useful or technological arts. All are physically the same, viz. printed
matter or an arbitrary succession of binary digits (in that any encoding
scheme is arbitrary). Appellants cite evidence by a Declarant that additional
information is needed in a contract. Id. As we find supra, absent some
functional effect, which Appellants have not persuasively identified here,
any distinctions between such data are, again, only in the mind of the
beholder and do not affect the operations recited of receiving and sending
data.
As a result, when giving the data labels recited in the claims no
patentable weight, the independent claims devolve to data entry which is
anticipated by any data entry reference, including those applied by the
Examiner. FF 1—5. And even if the reservation and contract labels are
afforded weight, we find that it would have been predictable to consolidate
data entry operations. FF 7, 10—11. And even if the master rental agreement
label is afforded weight, it would have been similarly predictable to
consolidate data entry operations to that agreement, such that the entered
data would not have been pre-existing, but instead would have been
contemporaneously entered. We discern that consolidating data entry
operations would have been predictable, because we find that there is a cost
24
Appeal 2015-002128
Application 12/650,113
in terms of time separation and data entry overhead, in splitting such
operations, that is saved with consolidation. FF 7, 10-11. As a corollary to
data entry consolidation, we discern that it would have been predictable to
have the customer enter the same data at his computer that a rental agent
would have entered at a rental counter, for the same reason of cost reduction.
Id. And although Hertz Gold requires a master rental agreement, there is
nothing about such an agreement that would preclude entering the equivalent
information during the reservation data entry, again for the benefits of data
entry consolidation. Id. Both a rental counter and a master rental agreement
are marketing and sales tools that provide an opportunity to upsell car class
and accessories and instill loyalty. FF 6, 11. We find that their raison d’etre
is more marketing related than due to technical considerations. Id. It would
have been predictable to avoid such marketing opportunities, where the
benefits in data consolidation were available, particularly as we find that it
was known to use a rental counter without a master agreement and vice
versa. KSR, 550 U.S. at 421 (“A person of ordinary skill is also a person of
ordinary creativity, not an automaton.”); Winner, 202 F.3d at 1349 (“the
district court did not clearly err in finding that one of ordinary skill in the art
would not have reasonably elected trading the benefit of security for that of
convenience. Trade-offs often concern what is feasible, not what is, on
balance, desirable. Motivation to combine requires the latter.”).
We are not persuaded by Appellants’ argument that a pre-existing
master rental agreement (MRA) is a technical requirement in Hertz Gold.
App. Br. 18. Appellants only show that the Hertz Gold software happens to
be programmed to require the MRA, not that the software must be so
programmed. The distinction is that Hertz Gold programs its software to
25
Appeal 2015-002128
Application 12/650,113
perform what Hertz desires, rather than what is technically required
irrespective of Hertz’s desires. The choice to predicate a step on an optional
criterion is just that, a choice, based on tactical considerations. Anyone with
minimal programming skills would have known how to overcome such an
obstacle, viz, remove the check for the criterion. To the extent patentable
weight were to be given to the data content in an MRA and reservation, we
find that combining data entry was both known and predictable. FF 10—11.
Appellants have not identified any timing issues related to the availability of
information that were overcome by the claims. All the information needed
from the customer in the MRA and contract is available to the customer at
the time of reservation. The known benefit of combining data entry in
avoiding delay and plural logs would have been sufficient reason for such a
combination.
We are not persuaded by Appellants’ argument that the prior art
requires a two pass, rather than one pass, system, for similar reasons.
Appellants have not identified any technical hurdle to overcome in
combining data entry operations; it is only a question of volition. Hertz
Gold shows that such combining of reservation and contract was known and
practiced. FF 10—11. That Hertz Gold required an MRA is no more than
additional data entry that one of ordinary skill would have immediately
envisaged combining with the reservation.
We are not persuaded by Appellants’ argument that the rejection
failed to articulate the scope and content of the prior art, and differences
from the claims. App. Br. 24. We summarized the Examiner’s findings of
such supra.
26
Appeal 2015-002128
Application 12/650,113
We are not persuaded by Appellants’ argument that the prior art fails
to describe a single visit. Id. We find that combining data entry would have
been predictable to one of ordinary skill at the time of the invention, supra.
We are not persuaded by Appellants’ argument that the Examiner
cited the wrong pages of Hertz. Id. We find that this was an inadvertent
typographical error, and that the appropriate portions of Hertz and Hertz
Gold that the Examiner intended to cite are clear.
We are not persuaded by Appellants’ argument that Hertz Gold
discloses nothing about electronic rental contracts and how such electronic
rental contracts can be created. App. Br. 30. The claim only recites
receiving and sending data to create contracts. Hertz Gold describes as
much. Contracts in computer memory are binary digits. How they are
interpreted is a matter discemable only in the human mind. Thus, all we are
left with is the creation of some data, which would have been known to one
of ordinary skill at the time of the invention. Absent some functional effect,
which Appellants have not persuasively identified here, whether the binary
data, as interpreted, meets legal requirements is not a matter of patentability,
but is a matter of legal rather than technological arts.
We are not persuaded by Appellants’ argument that the Examiner
failed to establish the level of skill of the person of ordinary skill. Id. It is
well—settled that the presentation of the prior art itself reflects an appropriate
level of ordinary skill in the art. See Chore-Time Equip., Inc. v. Cumberland
Corp., 713 F.2d 114, 779 (Fed. Cir. 1983). See also Okajima v. Bourdeau,
261 F.3d 1350, 1355 (Fed. Cir. 2001); In re GPAC, Inc., 57 F.3d 1573, 1579
(Fed. Cir. 1995); In re Oelrich, 579 F.2d 86, 91 (CCPA 1978).
27
Appeal 2015-002128
Application 12/650,113
We are not persuaded by Appellants’ argument that the Examiner
failed to articulate why one of ordinary skill would have combined the
references. Id. We find that the three references all describe the same
process of car rental, albeit from differing perspectives. We are unpersuaded
that the Examiner’s proffered combination is anything more than a
combination of familiar elements according to known methods that does no
more than yield predictable results. KSR, 550 U.S. at 401.
As to separately argued claim 24, we are not persuaded by Appellants’
argument that patentable weight must be afforded a claim element that
provides functionality with respect to the end use of a claimed system. App.
Br. 32. As set forth above, the test for whether a limitation is afforded
patentable weight is whether the limitation affects or is affected by the
recited operations, or whether the recited operations affect or are affected by
the limitation. King Pharm., 616 F.3d at 1279. Here, the recited operations
are those of sending, editing, creating, and receiving data. The operations
are not conditioned on the presence of certain data, as the claim and
arguments expressly perform the operations irrespective of the presence or
absence of certain data. As we find supra, these operations are insensitive to
the labels attached to the data, and the data labels are unaffected by these
operations.
We are not persuaded by Appellants’ argument that
the “electronic rental contract” authorizes the user “to
pick up a rental vehicle in accordance with the completed
electronic rental contract without the user visiting a rental
counter” are inescapably intertwined with a functional effect of
claim 24. The system of claim 24 can create an electronic
rental contract for a user that permits such user to avoid
creating a rental contract at the rental counter when arriving at
28
Appeal 2015-002128
Application 12/650,113
the car rental facility to pick up a rental vehicle in accordance
with the electronic rental contract, which ameliorates
inconveniences for the user when picking up a rental vehicle.
Thus, the system of claim 24 provides a user with opportunities
and capabilities that did not exist before because the system of
claim 24 lets a user create an “electronic rental contract” that
authorizes the user “to pick up a rental vehicle in accordance
with the completed electronic rental contract without the user
visiting a rental counter” even if the user does not have a pre­
existing MRA with the subject rental car company.
App. Br. 34. Authorization for a manual act is in the mind of the beholder; it
is not a functional effect or result. There is nothing about a customer service
counter that caused authorization in conventional rentals; the counter was
just the conventional site for the transaction. Nothing about a rental
transaction dictates such a site. Avoiding a counter does not affect the
operations of sending, receiving, editing, and creating data. To the extent
avoiding a counter affects the timing of those operations, we find that this
would have predictable to one of ordinary skill, as described in Hertz Gold
supra. FF 10-11.
We are not persuaded by Appellants’ argument similar to the one just
referred to, but with respect to avoiding the Master Rental Agreement. App.
Br. 35. As we find supra, absent some functional effect, which has not been
identified by Appellants, there is no patentable distinction between an MRA
and a reservation, as both are merely binary data entered into a computer.
And again, to the extent the argument is that the MRA contains data that are
necessary for a contract that might not be entered into a reservation, this is
not in the claim, and it would have been predictable to combine two separate
data entry sessions into a single session where all the information is known.
As by the terms of the claim, an MRA is pre-existing at the time of a
29
Appeal 2015-002128
Application 12/650,113
reservation, all of the information would have been known at the time of the
reservation. It would further have been predictable to enter the data for an
MRA in the same session as a reservation, so that it is not pre-existing at the
time of the session, for the purpose of meeting any such requirement for an
MRA the first time a reservation is desired.
Claim 3 is argued on the basis of claim 1. App. Br. 36.
As to separately argued claim 15, we are not persuaded by Appellants’
argument that the prior art fails to describe an option for converting a
reservation into a contract. The argument refers to this as a one touch, two
transaction combination. App. Br. 36—37. First, the claim does not recite a
one touch, two transaction car rental limitation, as such. Second, the claim
does not recite any car rental transaction per se, but only data entry,
transmission, and reception. Third, again, absent some functional effect,
which Appellants have not persuasively identified here, there is no
patentable distinction between the contents of distinct data entry sessions
based on their labels. Fourth, again, we find that it would have been
predictable to combine two separate data entry sessions into a single session,
to avoid time separation and data entry overhead. FF 7, 10-11. Fifth, Hertz
Gold shows it was known to combine the reservation and contracting in a
single session. FF 10—11
Claim 38 is argued on the basis of claim 15. App. Br. 37.
As to separately argued claim 18, we are not persuaded by Appellants’
argument that the prior art fails to describe validating a driver license and
conditioning the contract on this. App. Br. 38—39. We find that contracting
for a car requires a valid driver license and credit card that are not
30
Appeal 2015-002128
Application 12/650,113
necessarily required just to reserve a car.5 FF 10. Automation of a known
manual activity is obvious. Leapfrog, id. It is unnecessary to actually see a
driver license to validate it because automated systems provide such
validation services.6 FF 12—15. One of ordinary skill would have known
that it was the marketing promotional activities that prompted Hertz to
require trips to the customer service counter, at least as much as having a
suitable system electronically perform such validation, in view of the
notoriety of much more secure and certain ways of automated validation.
FF 6, 12—15. Auto rental companies were validating driver licenses online
and checking driver records with those license numbers as early as 1993.
Burch 1. Gallian shows that validating licenses based on the number pattern
was known.7 FF 15. One of ordinary skill would have known that a driver
license number is just a data string that can be passed through to the
validation services from the data entered by the customer.
Claim 41 is argued on the basis of claims 18 and 24. App. Br. 40.
We are not persuaded by Appellants’ argument, as to claims 107 and
113, about data configured to flag the reservation and rental transaction as
eligible for the user to pick up the rental vehicle, in accordance with the
completed electronic rental contract, without the user visiting a rental
counter. App. Br. 40-41. As we find supra, absent some functional effect,
which has not been identified by Appellants, there is no patentable
5 We discern that this comports with the experience of anyone renting a car.
6 This is known to anyone who has ever been pulled over in a car by a
policeman or seen a representation of this on television.
31
Appeal 2015-002128
Application 12/650,113
distinction among a reservation, transaction, and contract. As the claims
recite no particular implementation for how the configuration is flagged,
such flagging is no more than some convention adopted by the programmer.
As the claims do not recite flagging as an alternative to ineligible, any
convention showing that the data may be usable as a contract would be
within the claim scope. Again, we find that it was predictable to store the
data as being usable for a contract.
We are not persuaded by Appellants’ argument as to claims 5, 69, 82,
105, and 109, and claims 28, 106, 110, 111, 112, and 114, aboutprefilling
data from a master agreement. App. Br. 41 42. Hertz describes using the
master agreement for a reservation (FF 7), and we discern that implementing
the reverse, prefilling the reservation from the agreement, would have been
known to and within the abilities of one of ordinary skill. In re Gazda, 219
F.2d 449 (CCPA 1955). As we find supra, there is no patentable distinction
among a reservation, transaction, and contract. As the claims recite no
particular implementation for how the data is prefilled, any disclosure using
the master agreement data would be within the claim scope. Hertz describes
using such master agreement data for a contract. FF 7.
We are not persuaded by Appellants’ argument as to claims 6, 70, and
29 about modifying data without modifying a master agreement. App. Br.
42-43. As we find supra, absent some functional effect, which has not been
identified by Appellants, there is no patentable distinction among a
reservation, transaction, and contract. As the claims recite no particular
7 Beyond that, we discern that the notoriety of hashing as an additional
check on validity was too notorious at the time of filing to consider this
32
Appeal 2015-002128
Application 12/650,113
implementation for how the data are modified, or that the data must come
from master agreement, any disclosure showing that the data would have
been entered from the reservation, or from customer data entry, would be
within the claim scope. FF 2, 3, 6, 10-11. Again, we find that it would have
been predictable to enter data usable for a contract.
We are not persuaded by Appellants’ argument as to claims 7, 71, and
30 about an option to keep modifications or revert data. App. Br. 43 44. As
we find supra, absent some functional effect, which has not been identified
by Appellants, there is no patentable distinction among a reservation,
transaction, and contract. As the claims recite no particular implementation
for how the data are modified, or that the data must come from the master
agreement, any disclosure showing that the data would have been edited
from the reservation, customer data entry, or the master agreement would be
within the claim scope. Again, we find that it would have been predictable
to edit the data from any available source (FF 2, 3, 6, 10-11), and the
notoriety of UNDO as an editing feature obviates any issue of the
predictability of reverting data. Although Appellants are correct that UNDO
is a generic command for cancelling the most recent command, which may
or may not be modification of data, we find that for data modification, which
is disclosed by the cited references, as indicated above (FF 2), user selection
of the UNDO feature would result in a data reversion.
Arguments as to the remaining claims are repetitions of the above
arguments.
unpredictable.
33
Appeal 2015-002128
Application 12/650,113
We have reviewed the Declarations by Mr. Smith and find them
unpersuasive.
The Board has broad discretion as to the weight to give to
declarations offered in the course of prosecution. See Velander
v. Garner, 348 F.3d 1359, 1371 (Fed. Cir. 2003)
(“[A]ccord[ing] little weight to broad conclusory statements [in
expert testimony before the Board] that it determined were
unsupported by corroborating references [was] within the
discretion of the trier of fact to give each item of evidence such
weight as it feels appropriate.”); cf. Ashland Oil, Inc. v. Delta
Resins & Refractories, Inc., 776 F.2d 281, 294 (Fed. Cir. 1985)
(“Opinion testimony rendered by experts must be given
consideration, and while not controlling, generally is entitled to
some weight. Lack of factual support for expert opinion going
to factual determinations, however, may render the testimony of
little probative value in a validity determination.” (citations
omitted)). Although there is “no reason why opinion evidence
relating to a fact issue should not be considered by an
examiner,” In re Alton, 76 F.3d 1168, 1175 n.10 (Fed. Cir.
1996), the Board is entitled to weigh the declarations and
conclude that the lack of factual corroboration warrants
discounting the opinions expressed in the declarations, see
Velander, 348 F.3d at 1371; Ashland Oil, 776 F.2d at 294.
In re American Academy of Science 367 F.3d 1359, 1368 (Fed. Cir. 2004).
We find that Mr. Smith’s Declarations describe conventional practices in the
rental industry and not technical hurdles that were difficult to overcome.
Choosing whether to allow a computer or a human to interact in converting a
reservation to a contract is no more than a decision weighing (i) the
marketing opportunity and adaptability of human intervention against (ii) the
speed and efficiency of automation. One of ordinary skill would have
known to weigh one more heavily than the other based on various factors,
such as corporate strategy in demonstrating handholding and tactics in
reducing time a customer has to spend in administration. Simply because a
34
Appeal 2015-002128
Application 12/650,113
practice was conventional does not mean a different practice was unknown,
unpredictable, or disparaged.
Because we introduce additional art and apply reasoning beyond that
of the Examiner, we denominate this as a new ground of rejection to afford
Appellants an opportunity for response.
CONCLUSIONS OF LAW
The rejection of claims 1—9, 11—13, 15—25, 27—32, 34—36, 38—80,
82—84, 90, 92—94, 100, and 104—114 under 35 U.S.C. § 101 as directed to
non—statutory subject matter is proper.
The rejection of claims 1—9, 11—13, 15—25, 27—32, 34—36, 38—80,
82-84, 90, 92-94, 100, and 104-114 under 35 U.S.C. § 112, second
paragraph, as failing to particularly point out and distinctly claim the
invention is improper.
The rejection of claims 1—9, 11—13, 15—25, 27—32, 34—36, 38—80,
82-84, 90, 92-94, 100, and 104-114 under 35 U.S.C. § 103(a) as
unpatentable over Hertz, Avis, and Hertz Gold is proper, but with the
addition of new art is denominated as new grounds of rejection.
DECISION
The rejection of claims 1—9, 11—13, 15—25, 27—32, 34—36, 38—80,
82-84, 90, 92-94, 100, and 104-114 is affirmed.
Our decision is not a final agency action.
35
Appeal 2015-002128
Application 12/650,113
This decision contains new grounds of rejection pursuant to
37 C.F.R. § 41.50(b). Section 41.50(b) provides “[a] new ground of
rejection pursuant to this paragraph shall not be considered final for judicial
review.” Section 41.50(b) also provides:
When the Board enters such a non-final decision, the
appellant, within two months from the date of the decision,
must exercise one of the following two options with respect to
the new ground of rejection to avoid termination of the appeal
as to the rejected claims:
(1) Reopen prosecution. Submit an appropriate
amendment of the claims so rejected or new Evidence relating
to the claims so rejected, or both, and have the matter
reconsidered by the examiner, in which event the prosecution
will be remanded to the examiner. The new ground of rejection
is binding upon the examiner unless an amendment or new
Evidence not previously of Record is made which, in the
opinion of the examiner, overcomes the new ground of rejection
designated in the decision. Should the examiner reject the
claims, appellant may again appeal to the Board pursuant to this
subpart.
(2) Request rehearing. Request that the proceeding be
reheard under § 41.52 by the Board upon the same Record. The
request for rehearing must address any new ground of rejection
and state with particularity the points believed to have been
misapprehended or overlooked in entering the new ground of
rejection and also state all other grounds upon which rehearing
is sought.
Further guidance on responding to a new ground of rejection can
be found in the Manual of Patent Examining Procedure § 1214.01.
36
Appeal 2015-002128
Application 12/650,113
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a).
See 37 C.F.R. § 1.136(a)(l)(iv) (2011).
AFFIRMED; 37 C.F.R, $ 41.50(b)
37
Application/Control No. Applicant(s)/Patent Under Patent
Notice of References Cited 12/650,113
Appeal No.
2015-002128
Examiner Art Unit
3622 Page 1 of 1
U.S. PATENT DOCUMENTS
* Document Number
Country Code-Number-Kind Code
Date
MM-YYYY Name Classification
A us-
B us-
C US-
D US-
E US-
F US-
G US-
H US-
1 US-
J US-
K US-
L US-
M US-
FOREIGN PATENT DOCUMENTS
* Document Number
Country Code-Number-Kind Code
Date
MM-YYYY Country Name Classification
N
O
P
Q
R
S
T
NON-PATENT DOCUMENTS
* Include as applicable: Author, Title Date, Publisher, Edition or Volume, Pertinent Pages)
U A.D.BURCH, RENTAL CAR COMPANIES CHECK DRIVERS' RECORDS, AUGUST 12, 1993
V JOSEPH A. GALLIAN, ASSIGNING DRIVER'S LICENSE NUMBERS, VOL. 64 NO. 1, FEBRUARY 1991
w
X
*A copy of this reference is not being furnished with this Office action. (See MPEP § 707.05(a).)
Dates in MM-YYYY format are publication dates. Classifications may be US or foreign.
U.S. Patent and Trademark Office
PTO-892 (Rev. 01-2001) Notice of References Cited Part of Paper No.
2/25/2017 Rental Car Companies Check Drivers' Records - tribunedigital-sunsentinel
Homs ---> Collections Motor Vehicies
[
Rental Car Companies Check Drivers'
Records
August 12.1993 j ByA.D. BURCH, Staff Writer
Your driving record affects more than your pride or pocketbook under policies adopted recently by the
nation's largest car rente! agencies.
Budget, Avis, Hertz and National are refusing to rent vehicles in some states to drivers with poor or
tainted records, a move to lower the cost of soaring liability costs.
The leasing agencies are now checking state motor vehicle records by computer before consumers
can get the keys. The driving record check is supposed to take less than a minute, so it is not expected
to delay getting a car rental.
Related Articles
Fort Lauderdale-based Alamo Rent A Car, is also considering a policy involving drivers' checks, but the
details are being worked out, said Liz Clark, director of public affairs.
Check Safety Recalls On Cars, Boats, Medicine
July 27, 2009
Law Bans Open Containers Of Alcohol In Motor
Vehicles
April 2, 1986
Failure To Pay Child Support Could Mean Losing
License
May 4 1998
Drivers May Catch A Break On Auto Tags
January 5, 2000
Leasing officials point to this multimillion-dollar phrase: "vicarious liability," as the reason for the new
policies. Vicarious liability is a law in New York and Florida allowing the owners of the vehicle to be held
liable even if the customer caused the accident.
"The cost of injuries is skyrocketing," said The Hertz Corporation spokesman Joe Russo, although he
did not have specific figures. "Just by virtue of us owning the car, we have to take the hit. So we are
now trying to identify the high-risk drivers and (say) they are no longer entitled to the privilege of renting
a car."
The program is limited to drivers in states where the program is operating, including Florida, New York,
Maryland and Ohio. For example, a New York- registered driver would be screened when attempting to
rent a car in Florida but not Utah, where the program is not being used.
Unlicensed Driver Dies In Accident On 1-95
j ,,6 lgg? Hertz, National and Budget report a denial rate that peaked at 10 percent of the drivers whose licenses
were checked. That rate has stabilized at 6 percent. The companies declined to release total leasing
numbers.
Find More Stories About
Motor Vehicles
Drivers License Check
See Anyones Drivers
License Record. Enter A
Name & Search For Free!
p t; b i icre co rd 5 re v ie ws .com
: Free Public Record Vehicle License Plate
ti
$2.99 Vehicle History
: Search Search - Enter Any License Check the Vehicle's
: 1) Enter Name - Check Plate Number History Before You Buy
! 100% Free. 2) Get View License Plate Number With instaVIN. Only $2.99
; Background Report Records. Get Unlimited Searches & inctavin con
i Instantly! Reports.
: oackgrounGaiert.ccrr: scaichquarry.ccrri
Hertz, based in Park Ridge, N.J., began its program in July 1992 in New York. It brought Florida on-line
in April and the greater Washington, D.C., area in June. The company has made preliminary plans to
add California later this year. Avis launched its program for Florida and New York drivers about a month
ago. Likewise, National and Budget adopted their policies late last year.
"The (Division of Motor Vehicles) is a tool to help us overcome difficulties in doing business," said
Michael Olsen, spokesman for the Minneapolis-based National Car Rental System Inc. "It also helps
with the safety issue."
Companies do not have access to motor vehicle records so they contract with TML Informational
Services, of Forest Hills, N.Y., which had direct access to the DMV databases. Rental agents enter the
driver's license number into the computer system. A TML representative runs the number and kicks
back a computer answer to the rental agent within 30 seconds, based on a series of criteria.
Although the rules vary among the companies, the red flag or denial usually bleeps across the
computer screen if a driver has a suspended, expired or revoked licenses, a history of driving under the
influence, driving without insurance or a license, or posssession of a stolen car.
In beginning the safe-driving program, the companies also launched a consumer awareness campaign.
http://articles.sun-sentinel.com/1993-08-12/business/9301290359_1_car-drivers-motor-vehicles 1/3
2/25/2017 Rental Car Companies Check Drivers' Records - tribunedigital-sunsentinel
"We have posted signs at the rental counters and our reservation sales agents (headquartered) in
Tulsa (Okla.) will tell Florida and New York consumers they are likely to run a check on their licenses so
they are not surprised when it pops up," said Ray Noble, spokesman for the Garden City, N.Y.-based
Avis Inc.
BAD DRIVERS BEWARE
The criteria for vehicle leasing vary by company. Here are the basic criteria used by Hertz and Avis:
1. Driver's license suspended, revoked, invalid, surrendered or expired, if not cleared.
2. Eight or more points on driving record within the past 24 months (48 months for reckless disregard
for life and property). Department of Motor Vehicle Accident Prevention credits are taken into
consideration.
3. Conviction for DWI/DWAI/DUI within past 72 months.
4. Three or more convictions for moving violations within past 36 months.
5. Two or more accidents within past 36 months.
6. One or more accidents resulting in fatality or bodily injury within past 48 months. (This applies only to
rentals in New York area; Avis does not use this criterion.)
7. Failure to report an accident or leaving the scene of an accident within the past 48 months.
8. Operating a motor vehicle without insurance or a valid driver's license within the past 48 months.
9. Permitting operation of a motor vehicle without insurance or unlicensed operation of a motor vehicle
within past 48 months.
10. Possession of stolen vehicle or use of a vehicle in a crime within 48 hours.
Drivers License Check
See Anyones Drivers License Record. Enter A Name & Search Eor Ereei Go :o s-.orn
Search DMV Driving Record
Type Name & State Then Get Driving Records Instantly Go to persopo.com
From the Web Sponsored Links by Taboola
How the U.S. Cut Pollution While Growing the Economy
The NetvTtrjrk Times for Chevron
Free Quick Start Developer Guide
Microsoft
Where Tabitha From ’’Bewitched” Ended Up Is So Incredible
Trend Chaser
Howdy Pilgrim, Can You Pass the John Wayne Movie Quiz?
Zoo.oom
Is Your Pet the Love of Your Life?
USA Today j A'estfHjSA
25 American Cities Where You Can Live Like a King
Insights Into Pig Health And Industry Developments.
7 Reasons Why Glasses Should be Bought Online
Gi a ris es USA.co :ti
http://articles.sun-sentinel.com/1993-08-12/business/9301290359_1_car-drivers-motor-vehicles 2/3
VOL. 64, NO. 1, FEBRUARY 1991 13
Assigning Driver's License Numbers
JOSEPH A. GALLIAN
University of Minnesota
Duluth, MN 5S812
"You know my name
look up the number”
John Lennon and Paul McCartney, You know my name,
single, B-side of Let it be, March 1970.
Introduction
.Among the individual states, a wide variety of methods are used to assign driver’s
license numbers. The three most common methods, a sequential number, the social
security number, and a computer-generated number, are uninteresting mathemati­
cally. On the other hand, many states encode data such as month and date of birth,
year of birth, and sex in ways that involve elementary mathematics. Seven states go so
far as to employ a check digit for possible detection of forgery or errors. Several states
assign driver’s license numbers by applying complicated hashing functions to the first,
middle, and last names and formulas or tables for the month and date of birth.
Surprisingly, the assignment of numbers is not always injective. In Michigan, for
instance, there are 56 numbers whose inverse image has two or more members. New
Jersey incorporates eye color into the number. Some states keep their method
confidential. In a few instances, administrators of the license bureaus do not know the
method used to assign numbers in their state! In this paper we discuss some of the
methods we have uncovered.
Check Digit Schemes in General Use
Schemes for the assignment of identification numbers are extremely varied in
methodology' and in the information encoded. Most interesting to mathematicians are
those that incorporate an extra digit for the detection of errors or fraud. Although the
purpose of this paper is to analyze the methods used for driver’s license numbers, it is
worthwhile to begin with a brief survey of the methods employed to assign check
digits to the most ubiquitous numbers in use and to provide a theoretical result that
delineates their limitations.
The simplest and least effective method for assigning a check digit is to use the
remainder or inverse of the remainder of the identification number modulo some .
number. For airline tickets, UPS packages, and Federal Express mail the check digit
is the identification number modulo 7. At the bottom of Figure 1 we see the airline
ticket number 17000459570 (the airline code 012 is not used in the calculation) is
assigned a check digit 3 since 17000459570 * 3 mod 7.
U.S. postal money orders use the remainder modulo 9 while VISA travelers checks
use the inverse of the number modulo 9. Thus, the check digit for the VISA number
1002044679091 is 2 since 1002044679091 * 7 mod 9.
14 MATHEMATICS MAGAZINE
■ ’ZBnT£ZX2X‘aXlTtX.t2X£Tn^rrr?’rrr’rjYTrriTTTn'T7TriTt
I 1KMTMUFST *lt» IMPS \RFFT/f5-7t 7000:459:570
DULUTHTVU AOCr'
CAMPUS-'W S
fi fset
FIGURE 1
Airline ticket with number 17000459570 and check digit 3.
The modulo 7 schemes detect all errors involving a single digit except those where
b is substituted for a and \a-b\ —1. Likewise, an error of the sort at - a
... _* ■ ■ ■ Qj • ■ ■ a. • • • will go undetected if 'a,: - dj =7 or if 6 divides j - i.
The modulo 9 schemes are slightly better at detecting single-digit errors: Only a
substitution of a 9 for a 0 or vice versa goes undetected. On the other hand, the only
errors of the form • ■ • a, ■ ■ ■ a. • • • ■ * • Oj * * • a, • • * that are undetected are
those involving the check digit itself. (A quick proof of this is to observe that the
residue of a number modulo 9 is the residue of the sum of its digits modulo 9.)
Nearly all methods for assigning a check digit to a string of digits involve a scalar
product of two vectors and modular arithmetic. For a string axa2 ••• ak_x and a
modulus n, many schemes assign a check digit ak so that
(aj, a2,..., ak) • (wk, w2, ■.,, wk) = Omod n.
We call such schemes linear and we call the vector (tf w2__ _ wk) the weighting
vector. The Universal Product Code (UPC) used on grocery items employs the
weighting vector (3,1,3,1,3,1,3,1,3,1,3,1) with n =* 10: the International Standard
Book Number (ISBN) utilizes (10,9,8,7,6,5,4,3,2,1) and n = 11; banks in the U S.
use (7,3,9,7,3,9,7,3,9) with n = 10; many Western countries use (7,3,1,7,3.1___)
with n = 10 to assign check digits to numbers on passports. Notice that the division
schemes mentioned at the outset of this section are also linear with weighting vectors
of the form (10*-2,10t-3,.... 10°, ± 1).
The error-detecting capability of linear schemes is given by the following theorem.
Theorem. Suppose a number ata.2 • • ■ ak satisfies the condition (ax,a2>..., at) ■
(t£j. w2, • • •, wk) = Omod n. Then the single error occasioned by substituting a\ for a,
is undetectable if and only if (a| — ai)wi is divisible by n and a sole error of the form
• • ■ a( ■ ■ ■ Oj ■ ■ ■ -* aj ■■■ a, ■ is undetectable if and only if (ai — aj)(wi —
wf is divisible by n.
Proof If a\ is substituted for at, then the dot product of the correct number and
the incorrect number differ by ^ — aJtVf. Thus, the error is undetected if and only if
(o' — afw, = Omod n.
Now consider an error of the form * * * at ■ * • aj • • • • • • a, * * • at • * * . Here
the dot product of the correct number and the incorrect number differ by
VOL. 64, NO, 1, FEBRUARY 1991 15
(a,u.-, + OjWj) - (OjtCf + a,Wj) = (ai-aJ)(wi- Wj)
The conclusion now follows as before.
Since the most common moduli are 10 and 11, the following corollary is worth
mention.
Corollary. Suppose an identification number axaz • • • ak satisfies
(Gj. a2...... dj.) ■ (wx,w2,...,wk) = Omodn
where 0 < at < n for each i. Then all single-digit errors occurring in the ith position
are detectable if and only if ic, is relatively prime to n and all errors of the form
■ • ■ at • ■ ■ Qj ■ • ■ aj ai - • ■ are detectable if and only if w( — u>j is rela­
tively prime to n.
The above theorem verifies our claims about the error-detection capability of the
schemes used on money orders and airline tickets. It also explains why the bank and
passport schemes will detect some errors of the form • • • abc • • * -* • * • cba • • •
while the UPC code will detect no such errors. Observe that because 11 is prime the
ISBN code detects 100% of all single-digit errors and 100% of all errors invoking
the interchange of two digits. But there is a price to pay for using the modulus 11:
The number at needed to satisfy the condition may be 10, which is two digits. In this
case, an alphabetic character such as X or A is used or such numbers are not issued.
As we will see below there are schemes that use the modulus 11 that do not resort to
an alphabetic character, but there is a price to pay for this too: Not all transposition
errors are detectable. More information about check digits schemes in use can be
found in [1], [21 [3]. [4], [6], [7], [8],
Check Digits on Driver's License Numbers
The state of Utah assigns an eight-digit driver’s license number in sequential order,
say {jjfl, * * • a8, then appends a check digit a9 using a linear scheme with weighting
vector (9,8,7,6,5,4,3,2,1) and modulus 10. This method is identical to that used by
the American Chemical Society for its chemical registry numbers. Assuming that all
errors are equally likely,1 this method detects 73/81 or 90.1% of all single-digit errors
and 100% of all transposition errors (i.e., errors of the form • * • ab • * • -♦ ■ • ■ ba
■ ■ ■ ).2
To verify the single-error detection rate, observe from our theorem that in positions
2, 4, 6, and 8 substitution of b for a will go undetected when |a — i| * 5; in position
5, a substitution of b for a will go undetected when a and b have the same parity.
Thus in each of positions 2, 4, 6, and 8 there are 10 undetected errors among 90
possible errors while in the fifth position, 40 of the 90 possible errors are undetected.
1 In practice all errors are not equally likely. One study [7, p. 15] revealed that a substitution of a "5” for
a “3” was 17 times as likely as a substitution of a “9” for a “1.” However, available data are insufficient to
assign reliable probabilities to the various error possibilities,
2A highly publicized error of this kind recently occurred when Lt. Col. Oliver North gave U.S. Assistant
Secretary of State EUiot Abrams an incorrect Swiss bank account number for depositing $10 million for the
contras. The correct number begins with “386”; the number North gave to Abrams begins with ‘‘368.“
16 MATHEMATICS MAGAZINE
So, in all, 80 of 810 errors are undetected.
Someone working for the Canadian province of Quebec, probably having seen a
scheme like the one used by Utah somewhere, came up with the laughable weighting
vector (12,11,10,9,... ,2,1) with modulus 10 to assip a check digit. Of course, any
error in the third position is undetectable and weights of 12 and 11 have the same
effect as the weights 2 and 1.
Newfoundland uses the weighting vector (1,2,3,4,5,6,7,8,1) with modulus 10.
This is nearly identical to the Utah scheme except that it will not detect the event
that the first and last digit are interchanged.
Three states use a modified linear scheme with modulus 11. New Mexico and
Tennessee append a check digit aH to ala2 ■ • • a7 as follows: First calculate
x = -(aj.tij. ... ,a-) • (2,7,6,5,4,3,2)modll.
If .v = 0. aH is 1; if x = 10, as = 0; otherwise as = .r. This method catches 100% of all
single-digit errors. Furthermore, the only errors of the form a, Oj ■■ ■ -* * - •
aJ - ■ ■ a. ■ ■ ■ that go undetected are those where i = 1 and j = 7 (an unlikely error
indeed) and some involving the check digits 0 and 1. Assuming that all transposition
errors are equally likely,3 this method detects 98.2% of such errors. The Vermont
scheme is the same as the one used by New Mexico except that when x = 0, the
letter ‘A" is the check. This method, like the ISBN method, yields a 100% detection
rate for both single-digit and transposition errors, but utilizes two formats for
numbers. Notice that there would be nothing lost if the weighting vector began with
8 instead of 2 and there would be a slight gain since errors of the form axat • • • a7as
—*■ a7a.i ■ ■ ■ a,as would be detectable.
The state of Washington and the province of Manitoba use a check digit scheme
devised by IBM in 1964 to assip a check digit. The license number is a blend of 12
alphabetic and numeric characters. To compute the Washington check digit, alpha­
betic characters are assigned numeric values as follows: * -»pr, A -*■ 1. B -»2......./ ->
9, / -* 1. K -*2,..., R -* 9. S -» 2, T -» 3.......Z ->9. (Notice the aberration at S.) The
12-character license number, after an alphabetic to numeric conversion, then corre­
sponds to a string of digits ala1 ■ ■ • av2 with a10 as the check digit calculated as
; ■■■ ha • • •, suppose
such an error is undetected. We consider four cases. For simplicity, assume a in the
correct number is in position 2, 4 or 6. The alternative case gives the same result.
Case 1. a, b < 5
Then 2a + b = 2b 4* a mod 10.
Thus a - b =* 0 and a = b.
Case 2. a < 5, b 3> 5
Then 2a + b**2b — 9 + a mod 10.
It follows that b — a = 9 so that b = 9 and a = 0.
Case 3. a > 5, b < 5
Then 2a — 9 + b & 2b + a mod 10.
So, a — b = 9 and a = 9 and b = 0.
Case 4. a > 5, b > 5
Then 2a — 9 + b = 2b-9+a mod 10.
Thus o - b = 0 and a = b.
So all transposition errors except 09 <-» 90 are detected. Since there are 90 possible
transposition errors, the error detection rate is 88/90 or 97.8%.
It is worth noting that Gumm [4] has shown that it is not possible to improve upon
these rates with any system that uses addition modulo 10 to compute the check digit
without utilizing an extra character, as was the case for the New Mexico scheme.
Wisconsin appends a check digit to a 13-digit string. Unfortunately, I have not
been able to figure out how this scheme works. I know it isn’t linear; for if so, the
weighting vector (tOj,tv2,wm, wu) could be determined by gathering up a large
number of valid license numbers to produce a system of linear equations with the
m/s as the unknowns. I have done this for modulo 10 and 11 to no avail. To
circumvent any peculiarity that might arise involving a check digit of 10 in a modulo
11 scheme (e.g„ New Mexico), I avoided numbers with a check digit of 0 or 1.
Encoding Personal Data
Here is the driver’s license number of a Wisconsin resident: E 425-7276-9176-07.
What information about the holder can you deduce from this number: year of birth,
day and month of birth, sex, name? None of these is obvious. Let’s go the other
direction. I am a resident of Minnesota. I was bom on January 5, 1942, and my
middle name is Anthony. From this can you deduce my driver’s license number?
Eleven states assign their driver s license numbers with hashing functions applied
solely to personal data. A good hash function should be fast and minimize collisions
(see [5, pp. 506-544] for a detailed discussion of this topic). Of course, there will be
occasions when two or more individuals have enough personal data in common that
collisions will occur. Most states have a tie-breaking mechanism to handle this
situation. Coding license numbers only from personal data enables automobile
insurers, government entities, and law enforcement agencies to determine the num­
bers when necessary.
Washington uses a complicated blend of name, check digit, and codes for the
month and date of birth to assign its numbers. This 12-digit identifier consists of the
18 MATHEMATICS MAGAZINE
first five letters of the surname; the first and middle initials (* is used when a name
has less than five characters, or there is no middle initial); the year of birth subtracted
from 100 (we suspect this is done to disguise the year of birth); a check digit; a code
for the month of birth; and a code for the day of birth. For instance, Fielding Mellish
(no middle name) bom on 10/29/42 receives the identifier MELLI F* 587P9. When
checked against a file of 1.6 million items, this scheme yielded duplicates at the rate
of 0.03% and only one number appeared as many as four times. (Most of the
duplications represented twins.) To ensure that the correspondence between individ­
uals and numbers is injective, 17 alternate codes for month and year of birth are
available. For example, an S can be used instead of a B for January or a Z instead of
a 9 for the year of birth. Interestingly, the check digit is invariant under all alternate
coding. The primary code and one alternate for months is given in Table 1 and the
code for the days is given in Table 2. Notice the absence of completely predictable
patterns.
TABLE 1. Washington code for months.
Months Codes Alternate Codes
January B S
February C T
March D U
April } 1
May K 2
June L 3
July M 4
August N 5
September O 6
October P 7
November Q 8
December R 9
TABLE 2. Washington code for days.
I - A 7 - G 13 -L 19-R 25-5
2 - B 8-H 14 -M 20-0 26-6
3 - C 9 - Z 15 -N 21-1 27 - 7
4- D 10 -S 16 -W 22-2 28-8
5 - E ii 17-F 23-3 29-9
6 - F 12 -K 18 -Q 24-4 30 - T
Illinois, Florida, and Wisconsin encode the surname, first name, middle initial, date
of birth, and sex by a quite sophisticated scheme. The first character of the license
number is the first character of the name. The next three characters are obtained by
applying the “Soundex Coding System" to the surname as follows:
1. Delete all occurrences of h and to.
2. Assign numbers to the remaining letters as follows:
b, f,p,v~* 1 4
c, g, j, k,q, s, x,s -* 2 m,n~* 5
d, t —* 3 r —* 6
(No values are assigned to a, e, i, o, u, and y.)
3. If two or more letters with the same numeric value are adjacent, omit all but the
first. (Here a, e, t, o, u, and y act as separators.) For example, Schworer becomes
Sorer and Hugh gill becomes Ugil.
VOL. 64, NO. t, FEBRUARY 1991 19
4. Delete the first character of the original name if still present.
5. Delete all occurrences of a, e, i, o, u, and y.
6. Use the first three digits corresponding to the remaining letters; append trailing
zeros if less than three letters remain.
Here are some examples: Schworer -» S-660; Hughgill -* H-240; Skow -» S-000;
Sachs -* S-200; Lennon —» L-550; McCartney —*M-263.
We parenthetically remark that the Soundex System was designed so that likely
misspellings of a name would nevertheless result in the correct coding of the name.
For example, frequent misspellings of my name are: Gallion, Gillian, Galian, Galion,
Gilliam, Gallahan, and Galliam. Observe that all of these yield the same coding as
Gallian. We also mention that the above method differs somewhat from the system
called Soundex by Knuth in [5, p. 392],
The next three digits are determined by summing numbers that correspond to the
first name and middle initial. The scheme for doing this begins with the block 000 for
the letter A and makes jumps of 20 for especially common names and each
subsequent letter of the alphabet. A small portion of this scheme is. given in Table 3.
The values assigned to the middle initial are given in Table 4.
So Aaron G. Schlecker would be coded as S426-007 (S426 from Schlecker; 000 for
Aaron +7 for “G"), while Anne P. Schlecker would be coded as S426-053.
The last five digits of Illinois and Florida numbers capture the year and date of
birth as well as the sex. In Illinois, each day of the year is assigned a three-digit
number in sequence beginning with 001 for January 1. However, each month is
assumed to have 31 days. Thus, March 1 is given 063. These numbers are then used
to identify the month and day of birth of male drivers. For females, the scheme is
identical except January 1 begins with 601. The last two digits of the year of birth,
separated by a dash (probably for camouflage), are listed in the 5th and 4th positions
from the end of the driver’s license number. Thus, a male bom on July 18, 1942.
would have the last five digits 4-2204 while a female bom on the same day would
have 4-2804. When necessary, Illinois adds an extra character to avoid duplications.
TABLE 3. Illinois. Florida, Wisconsin given name or first initial code.
000 — A
020 - Albert, Alice
040 - Ann, Anna, Anne, .Annie. Arthur
060 B
080 _ Bernard, Bette, Bettie, Bettv
100 — C
TABLE 4. Illinois, Florida, Wisconsin middle initial code.
0 - none 10-]
1 - A 11 -K
2 - B 12-L
3 - C 13 -M
4 - D 14-N.O
5-E 15- P,Q
6 - F 16-JR
7-G 17 -S
8-H 18 - T, U. V
9-1 19 - w, x, y, z
20 MATHEMATICS MAGAZINE
Among the 9,397,518 licenses on file on January 1, 1987, this occurred in 14,856
instances. Of these, 55 numbers corresponded to three individuals (excluding the
extra digit). No number corresponded to more than three people.
The scheme to identify birthdate and sex in Florida is the same as Illinois except
each month is assumed to have 40 days and 500 is added for women. For example,
the five digits 49583 belong to a woman bom on March 3, 1949.
Wisconsin employs the same scheme as Florida to generate the first 12 of their 14
characters. The thirteenth character is an integer issued sequentially beginning with 0
to people who share the same first 12 characters. The fourteenth character is a check
digit.
A Missouri driver’s license number has 16 characters. The first 13 characters are
obtained by applying a hashing function to the first five letters of the surname, the
first three letters of the first name and the middle initial. (The method of encoding is
similar to that used by Florida.) The final three characters are a function of the month
and day of birth and sex. For a male bom in month m and day d the three digits are
63m + 2d. For a female, the corresponding formula is 63m + 2d + l. Thus the
number of a woman bom on March 4 has the final three characters 198. To avoid
duplications, Missouri assigns a 17th character. Among the first 3,921,922 numbers
issued, 31,719 have a 17th character.
Last, we discuss the scheme employed by Minnesota, Michigan, and Maryland.
The number is a function of last name, first name, middle name, month and date of
birth. The first four characters are determined by the Soundex System, as was the
case for Illinois, Florida, and Wisconsin. The first and middle names account for the
next six characters and the same algorithm is applied to both names. In the majority
of cases the first two characters of the name determine the desired three digits for
each name (see Table 5 for a sample); for common pairs of leading letters such as Al
or Ja, the third letter is invoked (see Table 6); 11 three-digit numbers are uniquely
assigned to the 11 most popular names (e.g., 189 *•* Edward; 210 «-* Elizabeth).
TABLE 5. Minnesota, Michigan, Maryland code for first and middle names
beginning with A except Al.
A 027
Aa 028 Aj 037 As 072
Ah 029 Ak 038 At 073
Ac 030 Al — Am 074
Ad 031 Am 066 Ae 075
Ae 032 An 067 Au.’ 076
Af 033 Aa 068 Ax 077
Ag 034 Ap 069 Ay 078
Ah 035 Aq 070 Ax 079
Ai 036 At 071
The final three digits are based on month and date of birth (but not year). Each day
of the year is assigned a three-digit number in a monotonicallv increasing fashion.
Although the usual pattern is to alternate increments of 3 and 2, there are numerous
seemingly random increments at unpredictable dates. The month of March illustrates
this behavior well. Notice from Table 7 that March 1 is assigned 159. Subsequent
days are assigned values by increments of 3 and 2 in alternating fashion until March
8. Then there is mi increment of 5. Notice the jump of 20 between March 19 and
March 20.
These gaps serve a practical purpose. In the event that there are two or more
individuals bom on the same month and date and with names so similar that the
VOL. 64, NO. 1 . FEBRUARY 1991 21
TABLE 6. Minnesota, Michigan, Maryland code for first and middle names
beginning with Al.
Al 039
Ala 040 Mi 049 Als 058
Aft 041 Aik 050 Alt 059
Ale 042 All 051 Alu 060
AW 043 Aim 052 Ale 061
Ale 044 Ain 053 Alw 062
Alf 045 Alo 054 Mx 063
Alg 046 Alp 055 Altj 064
Alh 047 Alq 056 Alz 065
Alt 048 Air 057
TABLE 7. Minnesota, .Michigan. Maryland code for dates in March.
1 -159
March - 158
11 - 187 21 - 229
2 - 162 12 - 189 22 - 232
3 -164 13 - 192 23 - 234
4 -167 14-194 24 - 237
5 -169 15 - 197 25 - 239
6 -172 16 - 199 26 - 242
7 - 174 17 - 202 27-244
8 -177 18-204 28 - 247
9 -182 19 - 207 29 - 249
10-184 20 - 227 30 - 252
31-254
hashing function does not distinguish between them (e.g„ Jill Paula Smith and Jimmy
Paul Smythe), the first person who applies for a license is assigned the number given
by the algorithm while the second person is assigned the next higher number thereby
using one of the numbers in the gap for birthdays. For example, if Jill Paula Smith is
bom on March 2 and is the first to receive the combination S530-441-675-162 as
determined by the algorithm, then the next person who yields the same number is
assigned S530-441-675-163 instead. Once all of the higher numbers in a gap have
been assigned, lower numbers are used. Thus the third applicant with a name
yielding the combination S530-441-675 bom on March 2 would be assigned the last
three digits 161. As of 1984, this scheme had not yielded any duplications among
4,468,080 people in Maryland while of Michigan’s 6,332,878 drivers by 1987 there are
56 that have a number not uniquely their own. In fact, Michigan has two numbers
that are each shared by four individuals and three that are each shared by three
individuals. A common cause of duplication is the custom of naming a son after the
father. When both share the same birthday a duplication occurs.
Summary
Table 8 summarizes the information the author has discovered about the coding of
driver’s license numbers. Unfortunately our knowledge is incomplete. Several states
(e.g., Florida, New York, Minnesota, Missouri, Wisconsin) keep their methods confi­
dential. In some of these cases we were able to determine the coding scheme by
examining data. A question mark after the letter X indicates the corresponding item
is used in the coding, but we do not know the method involved. The expression (A)
after an X indicates that the corresponding item is part of a scheme that is an
alternative to the social security number.
O. 1, FEBRUARY 1991 21
.E 6. Minnesota, Michigan, Maryland code for first and middle names
ining with Al.
Al 039
la 040 Alj 049 Als 058
lb 041 Aik 050 Alt 059
1c 042 All 051 Alu 060
Id 043 Aim 052 Ale 061
le 044 Ain 053 Alw 062
If 045 Alo 054 Alx 063
ig 046 Alp 055 Aly 064
Ih 047 Afo 056 Alz 065
li 048 Air 057
TABLE 7, Minnesota, Michigan. Maryland code for dates in March.
■ March - 158
1 - 159 11 - 187 21 - 229
2 -162 12- 189 22 - 232
3 -164 13- 192 23 - 234
4 -167 14- 194 24 - 237
5 - 169 15 - 197 25 - 239
6 -172 16- 199 26 - 242
7 -174 17- 202 27 - 244
8 -177. 18- 204 28 - 247
9 - 182 19- 207 29 - 249
10 - 184 20- 227 30 - 232
31 - 254
motion does not distinguish between them (e.g., Jill Paula Smith and Jimmy
the), the first person who applies for a license is assigned the number given
iorithm while the second person is assigned the next higher number thereby
of the numbers in the gap for birthdays. For example, if Jill Paula Smith is
March 2 and is the first to receive the combination S530-441-675-162 as
:d by the algorithm, then the next person who yields the same number is
S530-441-675-163 instead. Once all of the higher numbers in a gap have
gned, lower numbers are used. Thus the third applicant with a name
he combination S530-441-675 bom on March 2 would be assigned the last
its 161. As of 1984, this scheme had not yielded any duplications among
people in Maryland while of Michigan’s 6,332,878 drivers by 1987 there are
ave a number not uniquely their own. In fact, Michigan has two numbers
;aeh shared by four individuals and three that are each shared by three
s. A common cause of duplication is the custom of naming a son after the
hen both share the same birthday a duplication occurs.
y
ummarizes the information the author has discovered about the coding of
cense numbers. Unfortunately our knowledge is incomplete. Several states
ida. New York, Minnesota, Missouri, Wisconsin) keep their methods confi-
i some of these cases we were able to determine the coding scheme by
; data. A question mark after the letter X indicates the corresponding item
the coding, but we do not know the method involved. The expression (A)
X indicates that the corresponding item is part of a scheme that is an
5 to the social security number,
O
O
M
O
JC
n *■
M
»
f
22 MATHEMATICS MAGAZINE
TABLE 8. Summary of Schemes for Assigning Driver's License Numbers.
Stale
Socud
Security
.Vernier
Computer or
Sequential
Number
Cheek
Dipt
Last
Verne
Coded
First
iVa me
CM
THuT
Usm
CM
Ytar of
Birth
CM
Month of
Birth
Coded
Birth
Coded
Sex
Coded
Alabama X
Alula. X
Ariwoi X
Arkansas X Xf AI
California X
Colorado X
Connecticut X X
Deiaware X X
Florida X X X X X X X
Georgia X XiA.
Hawaii X
Idaho X X..., at) ■
(t£j. w2, • • •, wk) = Omod n. Then the single error occasioned by substituting a\ for a,
is undetectable if and only if (a| — ai)wi is divisible by n and a sole error of the form
• • ■ a( ■ ■ ■ Oj ■ ■ ■ -* aj ■■■ a, ■ is undetectable if and only if (ai — aj)(wi —
wf is divisible by n.
Proof If a\ is substituted for at, then the dot product of the correct number and
the incorrect number differ by ^ — aJtVf. Thus, the error is undetected if and only if
(o' — afw, = Omod n.
Now consider an error of the form * * * at ■ * • aj • • • • • • a, * * • at • * * . Here
the dot product of the correct number and the incorrect number differ by
VOL. 64, NO, 1, FEBRUARY 1991 15
(a,u.-, + OjWj) - (OjtCf + a,Wj) = (ai-aJ)(wi- Wj)
The conclusion now follows as before.
Since the most common moduli are 10 and 11, the following corollary is worth
mention.
Corollary. Suppose an identification number axaz • • • ak satisfies
(Gj. a2...... dj.) ■ (wx,w2,...,wk) = Omodn
where 0 < at < n for each i. Then all single-digit errors occurring in the ith position
are detectable if and only if ic, is relatively prime to n and all errors of the form
■ • ■ at • ■ ■ Qj ■ • ■ aj ai - • ■ are detectable if and only if w( — u>j is rela­
tively prime to n.
The above theorem verifies our claims about the error-detection capability of the
schemes used on money orders and airline tickets. It also explains why the bank and
passport schemes will detect some errors of the form • • • abc • • * -* • * • cba • • •
while the UPC code will detect no such errors. Observe that because 11 is prime the
ISBN code detects 100% of all single-digit errors and 100% of all errors invoking
the interchange of two digits. But there is a price to pay for using the modulus 11:
The number at needed to satisfy the condition may be 10, which is two digits. In this
case, an alphabetic character such as X or A is used or such numbers are not issued.
As we will see below there are schemes that use the modulus 11 that do not resort to
an alphabetic character, but there is a price to pay for this too: Not all transposition
errors are detectable. More information about check digits schemes in use can be
found in [1], [21 [3]. [4], [6], [7], [8],
Check Digits on Driver's License Numbers
The state of Utah assigns an eight-digit driver’s license number in sequential order,
say {jjfl, * * • a8, then appends a check digit a9 using a linear scheme with weighting
vector (9,8,7,6,5,4,3,2,1) and modulus 10. This method is identical to that used by
the American Chemical Society for its chemical registry numbers. Assuming that all
errors are equally likely,1 this method detects 73/81 or 90.1% of all single-digit errors
and 100% of all transposition errors (i.e., errors of the form • * • ab • * • -♦ ■ • ■ ba
■ ■ ■ ).2
To verify the single-error detection rate, observe from our theorem that in positions
2, 4, 6, and 8 substitution of b for a will go undetected when |a — i| * 5; in position
5, a substitution of b for a will go undetected when a and b have the same parity.
Thus in each of positions 2, 4, 6, and 8 there are 10 undetected errors among 90
possible errors while in the fifth position, 40 of the 90 possible errors are undetected.
1 In practice all errors are not equally likely. One study [7, p. 15] revealed that a substitution of a "5” for
a “3” was 17 times as likely as a substitution of a “9” for a “1.” However, available data are insufficient to
assign reliable probabilities to the various error possibilities,
2A highly publicized error of this kind recently occurred when Lt. Col. Oliver North gave U.S. Assistant
Secretary of State EUiot Abrams an incorrect Swiss bank account number for depositing $10 million for the
contras. The correct number begins with “386”; the number North gave to Abrams begins with ‘‘368.“
16 MATHEMATICS MAGAZINE
So, in all, 80 of 810 errors are undetected.
Someone working for the Canadian province of Quebec, probably having seen a
scheme like the one used by Utah somewhere, came up with the laughable weighting
vector (12,11,10,9,... ,2,1) with modulus 10 to assip a check digit. Of course, any
error in the third position is undetectable and weights of 12 and 11 have the same
effect as the weights 2 and 1.
Newfoundland uses the weighting vector (1,2,3,4,5,6,7,8,1) with modulus 10.
This is nearly identical to the Utah scheme except that it will not detect the event
that the first and last digit are interchanged.
Three states use a modified linear scheme with modulus 11. New Mexico and
Tennessee append a check digit aH to ala2 ■ • • a7 as follows: First calculate
x = -(aj.tij. ... ,a-) • (2,7,6,5,4,3,2)modll.
If .v = 0. aH is 1; if x = 10, as = 0; otherwise as = .r. This method catches 100% of all
single-digit errors. Furthermore, the only errors of the form a, Oj ■■ ■ -* * - •
aJ - ■ ■ a. ■ ■ ■ that go undetected are those where i = 1 and j = 7 (an unlikely error
indeed) and some involving the check digits 0 and 1. Assuming that all transposition
errors are equally likely,3 this method detects 98.2% of such errors. The Vermont
scheme is the same as the one used by New Mexico except that when x = 0, the
letter ‘A" is the check. This method, like the ISBN method, yields a 100% detection
rate for both single-digit and transposition errors, but utilizes two formats for
numbers. Notice that there would be nothing lost if the weighting vector began with
8 instead of 2 and there would be a slight gain since errors of the form axat • • • a7as
—*■ a7a.i ■ ■ ■ a,as would be detectable.
The state of Washington and the province of Manitoba use a check digit scheme
devised by IBM in 1964 to assip a check digit. The license number is a blend of 12
alphabetic and numeric characters. To compute the Washington check digit, alpha­
betic characters are assigned numeric values as follows: * -»pr, A -*■ 1. B -»2......./ ->
9, / -* 1. K -*2,..., R -* 9. S -» 2, T -» 3.......Z ->9. (Notice the aberration at S.) The
12-character license number, after an alphabetic to numeric conversion, then corre­
sponds to a string of digits ala1 ■ ■ • av2 with a10 as the check digit calculated as
; ■■■ ha • • •, suppose
such an error is undetected. We consider four cases. For simplicity, assume a in the
correct number is in position 2, 4 or 6. The alternative case gives the same result.
Case 1. a, b < 5
Then 2a + b = 2b 4* a mod 10.
Thus a - b =* 0 and a = b.
Case 2. a < 5, b 3> 5
Then 2a + b**2b — 9 + a mod 10.
It follows that b — a = 9 so that b = 9 and a = 0.
Case 3. a > 5, b < 5
Then 2a — 9 + b & 2b + a mod 10.
So, a — b = 9 and a = 9 and b = 0.
Case 4. a > 5, b > 5
Then 2a — 9 + b = 2b-9+a mod 10.
Thus o - b = 0 and a = b.
So all transposition errors except 09 <-» 90 are detected. Since there are 90 possible
transposition errors, the error detection rate is 88/90 or 97.8%.
It is worth noting that Gumm [4] has shown that it is not possible to improve upon
these rates with any system that uses addition modulo 10 to compute the check digit
without utilizing an extra character, as was the case for the New Mexico scheme.
Wisconsin appends a check digit to a 13-digit string. Unfortunately, I have not
been able to figure out how this scheme works. I know it isn’t linear; for if so, the
weighting vector (tOj,tv2,wm, wu) could be determined by gathering up a large
number of valid license numbers to produce a system of linear equations with the
m/s as the unknowns. I have done this for modulo 10 and 11 to no avail. To
circumvent any peculiarity that might arise involving a check digit of 10 in a modulo
11 scheme (e.g„ New Mexico), I avoided numbers with a check digit of 0 or 1.
Encoding Personal Data
Here is the driver’s license number of a Wisconsin resident: E 425-7276-9176-07.
What information about the holder can you deduce from this number: year of birth,
day and month of birth, sex, name? None of these is obvious. Let’s go the other
direction. I am a resident of Minnesota. I was bom on January 5, 1942, and my
middle name is Anthony. From this can you deduce my driver’s license number?
Eleven states assign their driver s license numbers with hashing functions applied
solely to personal data. A good hash function should be fast and minimize collisions
(see [5, pp. 506-544] for a detailed discussion of this topic). Of course, there will be
occasions when two or more individuals have enough personal data in common that
collisions will occur. Most states have a tie-breaking mechanism to handle this
situation. Coding license numbers only from personal data enables automobile
insurers, government entities, and law enforcement agencies to determine the num­
bers when necessary.
Washington uses a complicated blend of name, check digit, and codes for the
month and date of birth to assign its numbers. This 12-digit identifier consists of the
18 MATHEMATICS MAGAZINE
first five letters of the surname; the first and middle initials (* is used when a name
has less than five characters, or there is no middle initial); the year of birth subtracted
from 100 (we suspect this is done to disguise the year of birth); a check digit; a code
for the month of birth; and a code for the day of birth. For instance, Fielding Mellish
(no middle name) bom on 10/29/42 receives the identifier MELLI F* 587P9. When
checked against a file of 1.6 million items, this scheme yielded duplicates at the rate
of 0.03% and only one number appeared as many as four times. (Most of the
duplications represented twins.) To ensure that the correspondence between individ­
uals and numbers is injective, 17 alternate codes for month and year of birth are
available. For example, an S can be used instead of a B for January or a Z instead of
a 9 for the year of birth. Interestingly, the check digit is invariant under all alternate
coding. The primary code and one alternate for months is given in Table 1 and the
code for the days is given in Table 2. Notice the absence of completely predictable
patterns.
TABLE 1. Washington code for months.
Months Codes Alternate Codes
January B S
February C T
March D U
April } 1
May K 2
June L 3
July M 4
August N 5
September O 6
October P 7
November Q 8
December R 9
TABLE 2. Washington code for days.
I - A 7 - G 13 -L 19-R 25-5
2 - B 8-H 14 -M 20-0 26-6
3 - C 9 - Z 15 -N 21-1 27 - 7
4- D 10 -S 16 -W 22-2 28-8
5 - E ii 17-F 23-3 29-9
6 - F 12 -K 18 -Q 24-4 30 - T
Illinois, Florida, and Wisconsin encode the surname, first name, middle initial, date
of birth, and sex by a quite sophisticated scheme. The first character of the license
number is the first character of the name. The next three characters are obtained by
applying the “Soundex Coding System" to the surname as follows:
1. Delete all occurrences of h and to.
2. Assign numbers to the remaining letters as follows:
b, f,p,v~* 1 4
c, g, j, k,q, s, x,s -* 2 m,n~* 5
d, t —* 3 r —* 6
(No values are assigned to a, e, i, o, u, and y.)
3. If two or more letters with the same numeric value are adjacent, omit all but the
first. (Here a, e, t, o, u, and y act as separators.) For example, Schworer becomes
Sorer and Hugh gill becomes Ugil.
VOL. 64, NO. t, FEBRUARY 1991 19
4. Delete the first character of the original name if still present.
5. Delete all occurrences of a, e, i, o, u, and y.
6. Use the first three digits corresponding to the remaining letters; append trailing
zeros if less than three letters remain.
Here are some examples: Schworer -» S-660; Hughgill -* H-240; Skow -» S-000;
Sachs -* S-200; Lennon —» L-550; McCartney —*M-263.
We parenthetically remark that the Soundex System was designed so that likely
misspellings of a name would nevertheless result in the correct coding of the name.
For example, frequent misspellings of my name are: Gallion, Gillian, Galian, Galion,
Gilliam, Gallahan, and Galliam. Observe that all of these yield the same coding as
Gallian. We also mention that the above method differs somewhat from the system
called Soundex by Knuth in [5, p. 392],
The next three digits are determined by summing numbers that correspond to the
first name and middle initial. The scheme for doing this begins with the block 000 for
the letter A and makes jumps of 20 for especially common names and each
subsequent letter of the alphabet. A small portion of this scheme is. given in Table 3.
The values assigned to the middle initial are given in Table 4.
So Aaron G. Schlecker would be coded as S426-007 (S426 from Schlecker; 000 for
Aaron +7 for “G"), while Anne P. Schlecker would be coded as S426-053.
The last five digits of Illinois and Florida numbers capture the year and date of
birth as well as the sex. In Illinois, each day of the year is assigned a three-digit
number in sequence beginning with 001 for January 1. However, each month is
assumed to have 31 days. Thus, March 1 is given 063. These numbers are then used
to identify the month and day of birth of male drivers. For females, the scheme is
identical except January 1 begins with 601. The last two digits of the year of birth,
separated by a dash (probably for camouflage), are listed in the 5th and 4th positions
from the end of the driver’s license number. Thus, a male bom on July 18, 1942.
would have the last five digits 4-2204 while a female bom on the same day would
have 4-2804. When necessary, Illinois adds an extra character to avoid duplications.
TABLE 3. Illinois. Florida, Wisconsin given name or first initial code.
000 — A
020 - Albert, Alice
040 - Ann, Anna, Anne, .Annie. Arthur
060 B
080 _ Bernard, Bette, Bettie, Bettv
100 — C
TABLE 4. Illinois, Florida, Wisconsin middle initial code.
0 - none 10-]
1 - A 11 -K
2 - B 12-L
3 - C 13 -M
4 - D 14-N.O
5-E 15- P,Q
6 - F 16-JR
7-G 17 -S
8-H 18 - T, U. V
9-1 19 - w, x, y, z
20 MATHEMATICS MAGAZINE
Among the 9,397,518 licenses on file on January 1, 1987, this occurred in 14,856
instances. Of these, 55 numbers corresponded to three individuals (excluding the
extra digit). No number corresponded to more than three people.
The scheme to identify birthdate and sex in Florida is the same as Illinois except
each month is assumed to have 40 days and 500 is added for women. For example,
the five digits 49583 belong to a woman bom on March 3, 1949.
Wisconsin employs the same scheme as Florida to generate the first 12 of their 14
characters. The thirteenth character is an integer issued sequentially beginning with 0
to people who share the same first 12 characters. The fourteenth character is a check
digit.
A Missouri driver’s license number has 16 characters. The first 13 characters are
obtained by applying a hashing function to the first five letters of the surname, the
first three letters of the first name and the middle initial. (The method of encoding is
similar to that used by Florida.) The final three characters are a function of the month
and day of birth and sex. For a male bom in month m and day d the three digits are
63m + 2d. For a female, the corresponding formula is 63m + 2d + l. Thus the
number of a woman bom on March 4 has the final three characters 198. To avoid
duplications, Missouri assigns a 17th character. Among the first 3,921,922 numbers
issued, 31,719 have a 17th character.
Last, we discuss the scheme employed by Minnesota, Michigan, and Maryland.
The number is a function of last name, first name, middle name, month and date of
birth. The first four characters are determined by the Soundex System, as was the
case for Illinois, Florida, and Wisconsin. The first and middle names account for the
next six characters and the same algorithm is applied to both names. In the majority
of cases the first two characters of the name determine the desired three digits for
each name (see Table 5 for a sample); for common pairs of leading letters such as Al
or Ja, the third letter is invoked (see Table 6); 11 three-digit numbers are uniquely
assigned to the 11 most popular names (e.g., 189 *•* Edward; 210 «-* Elizabeth).
TABLE 5. Minnesota, Michigan, Maryland code for first and middle names
beginning with A except Al.
A 027
Aa 028 Aj 037 As 072
Ah 029 Ak 038 At 073
Ac 030 Al — Am 074
Ad 031 Am 066 Ae 075
Ae 032 An 067 Au.’ 076
Af 033 Aa 068 Ax 077
Ag 034 Ap 069 Ay 078
Ah 035 Aq 070 Ax 079
Ai 036 At 071
The final three digits are based on month and date of birth (but not year). Each day
of the year is assigned a three-digit number in a monotonicallv increasing fashion.
Although the usual pattern is to alternate increments of 3 and 2, there are numerous
seemingly random increments at unpredictable dates. The month of March illustrates
this behavior well. Notice from Table 7 that March 1 is assigned 159. Subsequent
days are assigned values by increments of 3 and 2 in alternating fashion until March
8. Then there is mi increment of 5. Notice the jump of 20 between March 19 and
March 20.
These gaps serve a practical purpose. In the event that there are two or more
individuals bom on the same month and date and with names so similar that the
VOL. 64, NO. 1 . FEBRUARY 1991 21
TABLE 6. Minnesota, Michigan, Maryland code for first and middle names
beginning with Al.
Al 039
Ala 040 Mi 049 Als 058
Aft 041 Aik 050 Alt 059
Ale 042 All 051 Alu 060
AW 043 Aim 052 Ale 061
Ale 044 Ain 053 Alw 062
Alf 045 Alo 054 Mx 063
Alg 046 Alp 055 Altj 064
Alh 047 Alq 056 Alz 065
Alt 048 Air 057
TABLE 7. Minnesota, .Michigan. Maryland code for dates in March.
1 -159
March - 158
11 - 187 21 - 229
2 - 162 12 - 189 22 - 232
3 -164 13 - 192 23 - 234
4 -167 14-194 24 - 237
5 -169 15 - 197 25 - 239
6 -172 16 - 199 26 - 242
7 - 174 17 - 202 27-244
8 -177 18-204 28 - 247
9 -182 19 - 207 29 - 249
10-184 20 - 227 30 - 252
31-254
hashing function does not distinguish between them (e.g„ Jill Paula Smith and Jimmy
Paul Smythe), the first person who applies for a license is assigned the number given
by the algorithm while the second person is assigned the next higher number thereby
using one of the numbers in the gap for birthdays. For example, if Jill Paula Smith is
bom on March 2 and is the first to receive the combination S530-441-675-162 as
determined by the algorithm, then the next person who yields the same number is
assigned S530-441-675-163 instead. Once all of the higher numbers in a gap have
been assigned, lower numbers are used. Thus the third applicant with a name
yielding the combination S530-441-675 bom on March 2 would be assigned the last
three digits 161. As of 1984, this scheme had not yielded any duplications among
4,468,080 people in Maryland while of Michigan’s 6,332,878 drivers by 1987 there are
56 that have a number not uniquely their own. In fact, Michigan has two numbers
that are each shared by four individuals and three that are each shared by three
individuals. A common cause of duplication is the custom of naming a son after the
father. When both share the same birthday a duplication occurs.
Summary
Table 8 summarizes the information the author has discovered about the coding of
driver’s license numbers. Unfortunately our knowledge is incomplete. Several states
(e.g., Florida, New York, Minnesota, Missouri, Wisconsin) keep their methods confi­
dential. In some of these cases we were able to determine the coding scheme by
examining data. A question mark after the letter X indicates the corresponding item
is used in the coding, but we do not know the method involved. The expression (A)
after an X indicates that the corresponding item is part of a scheme that is an
alternative to the social security number.
O. 1, FEBRUARY 1991 21
.E 6. Minnesota, Michigan, Maryland code for first and middle names
ining with Al.
Al 039
la 040 Alj 049 Als 058
lb 041 Aik 050 Alt 059
1c 042 All 051 Alu 060
Id 043 Aim 052 Ale 061
le 044 Ain 053 Alw 062
If 045 Alo 054 Alx 063
ig 046 Alp 055 Aly 064
Ih 047 Afo 056 Alz 065
li 048 Air 057
TABLE 7, Minnesota, Michigan. Maryland code for dates in March.
■ March - 158
1 - 159 11 - 187 21 - 229
2 -162 12- 189 22 - 232
3 -164 13- 192 23 - 234
4 -167 14- 194 24 - 237
5 - 169 15 - 197 25 - 239
6 -172 16- 199 26 - 242
7 -174 17- 202 27 - 244
8 -177. 18- 204 28 - 247
9 - 182 19- 207 29 - 249
10 - 184 20- 227 30 - 232
31 - 254
motion does not distinguish between them (e.g., Jill Paula Smith and Jimmy
the), the first person who applies for a license is assigned the number given
iorithm while the second person is assigned the next higher number thereby
of the numbers in the gap for birthdays. For example, if Jill Paula Smith is
March 2 and is the first to receive the combination S530-441-675-162 as
:d by the algorithm, then the next person who yields the same number is
S530-441-675-163 instead. Once all of the higher numbers in a gap have
gned, lower numbers are used. Thus the third applicant with a name
he combination S530-441-675 bom on March 2 would be assigned the last
its 161. As of 1984, this scheme had not yielded any duplications among
people in Maryland while of Michigan’s 6,332,878 drivers by 1987 there are
ave a number not uniquely their own. In fact, Michigan has two numbers
;aeh shared by four individuals and three that are each shared by three
s. A common cause of duplication is the custom of naming a son after the
hen both share the same birthday a duplication occurs.
y
ummarizes the information the author has discovered about the coding of
cense numbers. Unfortunately our knowledge is incomplete. Several states
ida. New York, Minnesota, Missouri, Wisconsin) keep their methods confi-
i some of these cases we were able to determine the coding scheme by
; data. A question mark after the letter X indicates the corresponding item
the coding, but we do not know the method involved. The expression (A)
X indicates that the corresponding item is part of a scheme that is an
5 to the social security number,
O
O
M
O
JC
n *■
M
»
f
22 MATHEMATICS MAGAZINE
TABLE 8. Summary of Schemes for Assigning Driver's License Numbers.
Stale
Socud
Security
.Vernier
Computer or
Sequential
Number
Cheek
Dipt
Last
Verne
Coded
First
iVa me
CM
THuT
Usm
CM
Ytar of
Birth
CM
Month of
Birth
Coded
Birth
Coded
Sex
Coded
Alabama X
Alula. X
Ariwoi X
Arkansas X Xf AI
California X
Colorado X
Connecticut X X
Deiaware X X
Florida X X X X X X X
Georgia X XiA.
Hawaii X
Idaho X X