Sandisk Corp. v. Kingston Technology Co., Inc.

United States Court of Appeals
for the Federal Circuit
__________________________
SANDISK CORPORATION,
Plaintiff-Appellant,
v.
KINGSTON TECHNOLOGY CO., INC.
AND KINGSTON TECHNOLOGY CORP.,
Defendants-Appellees.
__________________________
2011-1346
__________________________
Appeal from the United States District Court for the
Western District of Wisconsin in consolidated case nos.
07-CV-0605 and 07-CV-0607, Senior Judge Barbara B.
Crabb.
__________________________
Decided: October 9, 2012
__________________________
GREGORY A. CASTANIAS, Jones Day, of Washington,
DC, argued for plaintiff-appellant. With him on the brief
was DOUGLAS R. COLE, of Columbus, Ohio. Of counsel
were VICTORIA DORFMAN, of New York, New York, and
THARAN G. LANIER, of Palo Alto, California.
DAVID M. BARKAN, Fish & Richardson P.C. of Red-
wood City, California, argued for defendants-appellees.
SANDISK CORP v. KINGSTON TECH 2
With him on the brief was DAVID HOFFMAN, of Austin,
Texas. Of counsel on the brief was CHRISTINE YANG, Law
Offices of S. J. CHRISTINE YANG, of Fountain Valley,
California. Of counsel was ALAN D. SMITH, of Boston,
Massachusetts.
__________________________
Before PROST, REYNA, and WALLACH, Circuit Judges.
Opinion for the court filed by Circuit Judge PROST.
Opinion concurring-in-part and dissenting-in-part filed by
Circuit Judge REYNA.
PROST, Circuit Judge.
SanDisk Corporation (“Sandisk”) sued Kingston
Technology Co., Inc. and Kingston Technology Corp.
(collectively “Kingston”) for infringement of U.S. Patent
Nos. 5,719,808 (“’808 patent”), 6,149,316 (“’316 Patent”),
6,426,893 (“’893 patent”), 6,757,842 (“’842 patent”), and
6,763,424 (“’424 patent”). After the district court issued
its claim construction opinion, SanDisk withdrew its
infringement claims with respect to the ’808 and ’893
patents and claims 1, 6, 7, 10, 12, 15, 18, and 20 of the
’842 patent. The district court granted Kingston’s motion
for summary judgment of non-infringement with respect
to certain asserted claims of the ’842, ’316, and ’424
patents. SanDisk dismissed its remaining infringement
claims and has appealed the district court’s judgment.
For the reasons set forth below, we affirm in part, vacate
in part, and remand for further proceedings.
I. BACKGROUND
The patents-in-suit all relate to various aspects of
flash memory, which is the type of Electrically Erasable
Programmable Read-Only Memory (“EEPROM”) used, for
3 SANDISK CORP v. KINGSTON TECH
example, in USB “thumb drives,” computers, smart
phones, and mp3 players. A benefit of flash memory is
that it is “non-volatile”; in other words, it continues to
store data even after the power source is removed. A
thumb drive, for example, retains its saved data after
being removed from a host computer and can be used to
transport data from one computer to another.
A typical flash memory device includes one or more
flash memory integrated circuit chips and a controller.
Each flash memory chip contains memory cells for storing
data. The cells are arranged as “pages” with multiple
pages comprising a “block” of cells.
The flash memory device’s controller accepts com-
mands from the “host” device (e.g., the computer to which
the USB thumb drive is attached) and then writes data
to, or retrieves data from, the memory chip depending on
the host device’s command. To write and retrieve data,
the controller must be able to identify where each piece of
data is located in the system. This tracking of data is
accomplished through “addressing.” In general, the
system uses two types of addresses: “physical” and “logi-
cal.” The physical address refers to the physical location
in the system where particular data is stored. The logical
address is the identifier for a specific piece of data; it
describes the data without regard to the data’s physical
location. Because a particular piece of data can change its
physical location, a logical address may be associated with
one physical address at one time and another physical
address at a different time. The controller maps the
logical address to the correct physical address, allowing
the flash memory system to provide the correct data to
the host device.
SANDISK CORP v. KINGSTON TECH 4
Unlike typical computer memory, the old data on the
flash memory cell must be erased every time new data is
written to the cell. The memory cells are erased an entire
block at a time, while data is written to the cell one page
at a time; the erasure of data, therefore, occurs in larger
segments than the writing of data. These erase/write
cycles wear down the memory cell until the cell eventually
no longer reliably stores information. The patents-in-suit
relate to various methods and systems for managing the
data in the flash memory system, including methods for
reducing the wear and tear on the flash memory cells.
SanDisk filed two complaints in the U.S. District
Court for the Western District of Wisconsin against
Kingston for patent infringement, and the district court
consolidated the two actions on January 28, 2008. After
the district court issued its claim construction order,
SanDisk withdrew its infringement allegations with
respect to claims 1, 6, 7, 10, 12, 15, 18, and 20 of the ’842
patent and all asserted claims of the ’893 and ’808 pat-
ents. Both parties moved for summary judgment. The
district court granted SanDisk’s motion as to SanDisk’s
claim that Kingston was contributorily infringing claims
20, 24, 28, and 30 of the ’424 patent by selling products
containing a Phison PS3006 controller. With respect to
all remaining asserted claims, the court found that King-
ston was not infringing as a matter of law and ultimately
entered judgment in favor of Kingston on those claims.
After the district court’s summary judgment order, the
parties entered into a Stipulation and Order Dismissing
Remaining Claims for Relief (“Stipulation”), whereby
SanDisk dismissed without prejudice its remaining in-
fringement claims involving the ’424 patent, and Kingston
dismissed without prejudice its related invalidity and
enforceability counterclaims. J.A. 17754-55. SanDisk
5 SANDISK CORP v. KINGSTON TECH
timely appealed, and we have jurisdiction under 28 U.S.C.
§ 1295(a)(1).
II. DISCUSSION
On appeal, SanDisk argues that the district court
erred in its construction of the following claim terms:
“recording a relative time of programming . . .” in claims 1
and 3 of the ’424 patent; “user data portion” and “over-
head portion” in claims 1, 10, and 61 of the ’842 patent
and claim 67 of the ’316 patent; “block characteristic
information” in claims 1 and 65 of the ’893 patent; and
“designating a combination[] . . .” in claim 16 of the ’808
patent. In addition to its claim construction arguments,
SanDisk further contends that the district court legally
erred in entering summary judgment of non-infringement
in favor of Kingston as to claim 20 of the ’424 patent and
claim 79 of the ’316 patent.
Claim construction is a question of law reviewed de
novo. Cybor Corp v. FAS Techs., Inc., 138 F.3d 1448,
1454-55 (Fed. Cir. 1998) (en banc). Claim terms generally
are construed in accordance with the ordinary and cus-
tomary meaning they would have to one of ordinary skill
in the art in light of the specification and the prosecution
history. Phillips v. AWH Corp., 415 F.3d 1303, 1312-14
(Fed. Cir. 2005) (en banc).
We review the district court’s grant of summary
judgment de novo, drawing all reasonable inferences in
favor of the nonmovant. Anderson v. Liberty Lobby, Inc.,
477 U.S. 242, 255 (1986). Summary judgment is appro-
priate “if the movant shows that there is no genuine
dispute as to any material fact and the movant is entitled
to judgment as a matter of law.” Fed. R. Civ. P. 56(a).
SANDISK CORP v. KINGSTON TECH 6
A. Claim Construction
As an initial matter, Kingston argues that we should
not address certain claim construction issues presented
by SanDisk because either we lack jurisdiction to review
the district court’s construction or SanDisk waived its
right to advance its proposed construction on appeal.
First, Kingston contends that we lack jurisdiction to
review the district court’s claim constructions related to
claims 1 and 65 of the ’893 patent, claim 16 of the ’808
patent, and claims 1 and 10 of the ’842 patent because
SanDisk voluntarily withdrew those claims from the
litigation without the parties stipulating to a judgment of
non-infringement; thus, Kingston argues, there is no
judgment for this court to review. We agree.
After the district court entered its claim construction
order, SanDisk informed Kingston that it was no longer
pursuing these claims. The district court never entered a
separate order dismissing these claims, but it acknowl-
edged in its summary judgment opinion that SanDisk had
withdrawn them. J.A. 40. We, therefore, treat SanDisk’s
withdrawal of the claims as being akin to either a Federal
Rule of Civil Procedure 15 amendment to the complaint,
see Gronholz v. Sears, Roebuck & Co., 836 F.2d 515, 518
(Fed. Cir. 1987) (“[A] plaintiff’s motion to dismiss a single
claim of a multi-count complaint is properly treated as an
amendment under . . . [Rule] 15.”), or a Rule 41(a) volun-
tary dismissal of claims without prejudice, see Nilssen v.
Motorola, Inc., 203 F.3d 782, 784 (Fed. Cir. 2000) (“In
many instances the procedure for, and effect of, an
amendment will be the same as a voluntary dismissal
because of the similarities between the governing rules.”)
(internal quotations and citations omitted). Regardless of
how we characterize the withdrawal, these claims are no
7 SANDISK CORP v. KINGSTON TECH
longer at issue, and we accordingly have no final judg-
ment before us with respect to these claims to review.
Our jurisdiction, however, is generally limited to final
judgments: “Under the ‘final judgment rule,’ parties may
only appeal a ‘final decision of a district court.’” Spread
Spectrum Screening L.L.C. v. Eastman Kodak Co., 657
F.3d 1349, 1354 (Fed. Cir. 2011) (quoting 28 U.S.C.
§ 1295(a)(1)). Here, SanDisk voluntarily withdrew the
’893 and ’808 patents and claims 1 and 10 of the ’842
patent from this action, and it does not dispute that the
district court never entered a stipulated judgment of non-
infringement with respect to these claims. Thus, these
claims do not present a current infringement controversy
before this court. Without such a controversy, we lack
Article III jurisdiction to decide these issues. See Streck,
Inc. v. Research & Diagnostics Sys., Inc., 665 F.3d 1269,
1281 (Fed. Cir. 2012) (“It is well-established that, in
patent cases, the existence of a ‘case or controversy must
be evaluated on a claim-by-claim basis.’”); Jang v. Boston
Sci. Corp., 532 F.3d 1330, 1336 (Fed. Cir. 2008) (resolving
claim construction issues “that do not actually affect the
infringement controversy between the parties” would
result in impermissible advisory opinion because “[t]he
Supreme Court has explicitly held that Article III does
not permit the courts to resolve issues when it is not clear
that the resolution of the question will resolve a concrete
controversy between interested parties”).
As a result, we reject SanDisk’s contention that the
district court’s ultimate entry of a final judgment confers
appellate jurisdiction over these withdrawn claims. To
the contrary, where, as here, a party’s claim construction
arguments do not affect the final judgment entered by the
court, they are not reviewable. See Mass. Inst. of Tech. v.
Abacus Software, 462 F.3d 1344, 1350 (Fed. Cir. 2006)
SANDISK CORP v. KINGSTON TECH 8
(refusing to address claim construction arguments “perti-
nent only to dismissed claims of invalidity” because “[a]n
appeal is not an opportunity to bring before the appellate
court every ruling with which one or more of the parties
disagrees without regard to whether the ruling has in any
way impacted the final judgment”).
Nor does the parties’ Stipulation, relied upon by San-
Disk, establish our jurisdiction. After the district court
ruled on the summary judgment motions (and after
SanDisk withdrew these claims), the parties entered into
the Stipulation, whereby SanDisk dismissed its “remain-
ing” infringement claims and Kingston dismissed its
related validity and unenforceability counterclaims.
According to SanDisk, the Stipulation’s recognition that
this court might “reverse[] remand[], or vacate[], in whole
or in part, the Court’s September 22, 2010 Claim Con-
struction Order” and that the parties agreed that the
“Stipulation shall not in any way prejudice any parties’
[sic] right to appeal this matter in whole or in part, in-
cluding, but not limited, to an appeal of the Court’s Sep-
tember 22, 2010 Claim Construction Order,” J.A. 17754-
55, evinces SanDisk’s “intent to pursue, on appeal, argu-
ments that those claim-construction rulings were legally
incorrect,” including any arguments related to the with-
drawn claims, SanDisk’s Reply Br. 2.
We are not persuaded. First, the Stipulation does not
change the fact that there is no final judgment with
respect to the withdrawn claims for us to review. Without
a final judgment as to the infringement or validity of
these claims, the court’s claim constructions that impact
only these withdrawn claims are not properly before us.
Second, the parties’ agreement that the Stipulation would
not affect their right to appeal the entire claim construc-
tion order cannot create a right to appeal where one
9 SANDISK CORP v. KINGSTON TECH
otherwise does not exist. See Bender v. Williamsport Area
School Dist., 475 U.S. 534, 541 (1986) (“[E]very federal
appellate court has a special obligation to satisfy itself . . .
of its own jurisdiction . . . even though the parties are
prepared to concede it.”) (internal quotation marks and
citations omitted). Consequently, we conclude that we
lack jurisdiction to resolve SanDisk’s claim construction
arguments that impact only the withdrawn claims.
Second, Kingston argues that SanDisk waived its
right to challenge the district court’s construction of “user
data portion” and “overhead data portion” in claim 61 of
the ’842 patent and claim 67 of the ’316 patent because (1)
the parties never presented these terms to the court for a
construction, and (2) SanDisk never disputed the court’s
construction of these terms during summary judgment.
SanDisk, however, maintains that it did not waive these
arguments because the district court construed the re-
lated terms “user data” and “overhead information” in
claims 1 and 10 of the ’842 patent, and the parties ac-
cepted that this construction also applied to claim 61 of
the ’842 patent and claim 67 of the ’316 patent. According
to Sandisk, Kingston confirmed this understanding when
it made the following representation to the district court
in its memorandum in support of its motion for summary
judgment:
Neither SanDisk nor Defendants presented the is-
sue of whether claims 61 and 67 were limited to
only one user data portion and one overhead data
portion during the claim construction process.
But as explained below, both parties understood
those claims to have the same scope as claims 1
and 10 of the ’842 patent. Indeed, Defendants
were quite surprised when SanDisk indicated it
would continue to assert claims 61 and 67, despite
SANDISK CORP v. KINGSTON TECH 10
the Court’s clear guidance as to the limited scope
of these claims imposed by the claim language.
J.A. 7341 n.4 (emphasis added). Moreover, according to
SanDisk, because it had already presented its claim
construction positions to the district court during the
Markman proceedings, it was not required to continue
challenging the court’s construction during summary
judgment to preserve its arguments for appeal.
We agree with SanDisk. Based on Kingston’s own
representations to the district court, the parties assumed
that the court’s constructions for claims 1 and 10 of the
’842 patent would also apply to the related terms in claim
61 of the ’842 patent and claim 67 of the ’316 patent. We,
therefore, are not persuaded by Kingston’s argument that
SanDisk failed to present its claim construction position
to the district court. Nor was SanDisk required to repeat
its unsuccessful construction arguments to the district
court during summary judgment to preserve the issue for
appeal. See O2 Micro Int’l Ltd. v. Beyond Innovation
Tech. Co., 521 F.3d 1351, 1358-59 (Fed. Cir. 2008) (finding
that party did not waive right to challenge claim con-
struction on appeal when it had advanced the argument
during Markman proceedings but did not object to the
district court’s jury instruction on that claim construc-
tion).
Accordingly, because we conclude that we lack juris-
diction over the ’808 and ’893 patents and claims 1 and 10
of the ’842 patent, we limit our review of the district
court’s claim constructions to (1) the “recording a relative
time of programming . . .” limitation in claims 1 and 3 of
the ’424 patent, and (2) the “at least a user data portion
and an overhead portion” limitation in claim 61 of the
’842 patent and claim 67 of the ’316 patent.
11 SANDISK CORP v. KINGSTON TECH
1. “recording a relative time of programming that at least
one page of new data and the at least one page of super-
ceded data” (’424 patent, claims 1 & 3)
In the conventional flash EEPROM system, an entire
block of data is copied to a new block, with the updated
data replacing the superceded data. The entire old data
block is then erased. This method caused re-writing of
the non-updated data with the resulting wear and tear on
the flash memory cells.
The ’424 patent covers a method for performing “par-
tial block” updates in flash memory devices. When the
flash memory system makes minor updates to already
stored data, such as by changing a few words in a docu-
ment, it performs a “partial block” update; in other words,
only part of the data block is updated. The controller
writes only the pages with the updated data into the new
block as opposed to rewriting the entire block of data.
The new data shares a logical address with the super-
ceded data. The controller reads the data from the blocks,
identifying those pages that have been superceded by a
more recently updated page sharing the logical address.
When the controller reports the data to the host system, it
substitutes the superceded pages with this updated data.
Claim 1 of the ’424 patent is representative:
1. In a non-volatile memory system having a plu-
rality of blocks of memory storage elements that
are individually erasable as a unit and which are
individually organized into a plurality of pages of
memory storage elements that are individually
programmable together, a method of substituting
new data for superceded data within at least one
page of one of the plurality of blocks while data in
SANDISK CORP v. KINGSTON TECH 12
at least another page of said one block is not re-
placed, comprising:
programming the new data into at least one
page of said one or another of the plural-
ity of blocks,
identifying the at least one page of superceded
data and the at least one page of new
data by a common logical address,
recording a relative time of programming the
at least one page of new data and the at
least one page of superceded data; and
wherein the at least one page of superceded
data is less than all the data contained in
said one block.
’424 patent col.12 l.60-col.13 l.10 (emphasis added).
The specification teaches two methods for identifying
the physical page containing the most recent version of
data with the same logical address. First, the specifica-
tion discloses writing a time stamp onto each individual
page that “provides an indication of its time of program-
ming, at least relative to the time that other pages with
the same logical address are programmed.” Id. at col.8
ll.26-40. Second, the specification teaches recording the
programming time for an entire block, referred to by the
parties as the “Block Recording Method.” See id. at col.9
l.40-col.10 l.43. In this method, “the time stamp . . . does
not need to be stored as part of each page” but “[r]ather, a
single time stamp can be recorded for each block, either
as part of the block or elsewhere within the non-volatile
memory, and is updated each time a page of data is
written into the block.” Id. at col.9 ll.42-50. Within the
13 SANDISK CORP v. KINGSTON TECH
block, the new data is physically stored after the old data
such that the most recent page with a particular logical
address is determined by the relative physical order of
those pages within the block: “Data is then read from
pages in an order of descending physical address, starting
from the last page of the most recently updated block
containing data pages having the same LBN [Logical
Block Number].” Id. at col.9 ll.50-53.
On appeal, SanDisk challenges the district court’s
construction of the “recording a relative time of program-
ming . . .” limitation in claims 1 and 3 of the ’424 patent.
According to SanDisk, the district court improperly con-
strued the claims to exclude the “Block Recording
Method” and further limited the claims during summary
judgment to require the recording of an actual time.
Specifically, SanDisk argues, inter alia, that Figures 8
and 11, along with the specification’s teaching of the
Block Recording Method, indicate that claims 1 and 3
encompass this method. In response, Kingston contends
that the claims need not cover all embodiments in the
specification particularly where none of the embodiments
is described as being “preferred.” Kingston maintains
that the district court’s construction was correct because
claims 1 and 3 of the ’424 patent explicitly require re-
cording the time the pages were programmed, as opposed
to only recording a single time value for the entire block
as in the Block Recording Method.
We agree with SanDisk that the district court im-
properly excluded the Block Recording Method from
claims 1 and 3 of the ’424 patent. First, the claim lan-
guage supports SanDisk’s broader construction. The
claims only require “recording a relative time of pro-
gramming,” not “a time of programming.” The use of
“relative” is significant: whereas “recording the time of
SANDISK CORP v. KINGSTON TECH 14
programming” would suggest that a time of programming
must be recorded for each page, “recording a relative time
of programming” merely requires recording some indica-
tion of the order of programming for those pages sharing
a logical address. The claims place no limitation on how
the claimed “recording” occurs.
Turning to the specification, its teachings are consis-
tent with this interpretation. The specification unambi-
guously discloses two distinct techniques for
“distinguish[ing] the pages containing the superceded
data from those containing the new, updated version [of
the data].” ’424 patent col.7 ll.59-60. In the first method,
a time stamp indicator is associated with each page of
data; in the second method—the Block Recording
Method—the time stamp indicator is associated with the
block, and the physical order of the pages in the block
identifies the most recently updated data with a particu-
lar logical address. As quoted above, in this “second
specification implementation of the inventive technique,”
“[t]he time stamp . . . does not need to be stored as part of
each page. Rather, a single time stamp can be recorded
for each block.” Id. at col.9 ll.40-53.
Figures 8 and 11 of the ’424 patent provide further il-
lustrations of this second implementation. In Figure 8,
the relative time of programming the updated and origi-
nal logical pages 3, 4, and 5 is determined by reading the
pages in the most recent physical block (PBN1) in reverse
order, “followed by reading the pages of the original block
(PBN0) in the same reverse order.” Id. at col.9 ll.56-57.
After the data in the updated pages—here, logical pages
3, 4, and 5 on physical pages 0, 1, and 2 of block PBN1—is
read, “the superceded data in those pages of the original
block PBN0 that are identified by the same logical page
15 SANDISK CORP v. KINGSTON TECH
numbers can be skipped during the reading process.” Id.
at col.9 ll.58-61.
In describing Figure 8, the specification additionally
instructs that “[o]nly an identity of those physical blocks
containing data of a common logical block and the relative
times that the physical blocks were programmed need to
be known in order to carry out this efficient reading
process.” Id. at col.10 ll.3-7.
Similarly, Figure 11 discloses a second update to logi-
cal page 5 within the same physical block (PBN1) where
the physical location of the two pages within the block
identifies the most recently updated page:
SANDISK CORP v. KINGSTON TECH 16
Specifically, in Figure 11, the new physical block
(PBN1) contains two versions of original page 5 located at
PBN1 physical page 2 and PBN1 physical page 3. Be-
cause the data will be read backwards starting from the
last page of the new block, the user data on PBN1 physi-
cal page 3 will be read, but the data on PBN1 physical
page 2 will not: “It will be noted that this example of
reading pages in a reverse order efficiently sorts out the
new data pages from the superceded data pages because
data are written in physical page locations of an erased
block in order from page 0 on.” Id. at col.10 ll.29-33.
Reading the claims in light of the specification, we
conclude that the district court erred in finding that
claims 1 and 3 of the ’424 patent excluded the Block
Recording Method. Further, to the extent that the dis-
trict court, during summary judgment, interpreted this
limitation as requiring the recording of an actual time of
programming, we agree with SanDisk that such an inter-
pretation is incorrect. Consistent with our conclusion
that the claims encompass the Block Recording Method, it
necessarily follows that the “relative time of program-
17 SANDISK CORP v. KINGSTON TECH
ming” can be recorded through non-temporal means such
as the location of the physical pages within the block or
the use of a “modulo-N counter” which also is disclosed in
the specification. 1
2. “at least a user data portion and an overhead portion”
(’842 patent, claim 61 & ’316 patent, claim 67)
The district court found that the “user data and over-
head information” limitation in claims 1 and 10 of the
’842 patent was “limited to a single user data and a single
overhead portion.” J.A. 35. During summary judgment,
the court applied this construction to the related limita-
tion, “a user data portion and an overhead portion” in
claim 61 of the ’842 patent and claim 67 of the ’316 pat-
ent, which, as we concluded above, are before us on ap-
peal.
Claim 61 of the ’842 patent is representative:
61. A method of operating a memory system with
a host system that includes a processor, wherein
the memory system includes one or more inte-
grated circuit chips individually including an ar-
ray of non-volatile floating gate memory cells
partitioned into a plurality of sectors that indi-
vidually include a distinct group of memory cells
that are erasable together as a unit, comprising:
1 The specification discloses storing the output
of a “modulo-N counter” to identify the most recently
updated page of a specific logical address. As the specifi-
cation explains, “[w]hen updating the data of a particular
page . . . the controller first reads the count stored in the
field 43 of the page whose data are being updated, incre-
ments the count by some amount, such as one, and then
writes that incremented count in the new block . . . .” ’424
patent col.8 ll.45-51.
SANDISK CORP v. KINGSTON TECH 18
providing said one or more of the memory in-
tegrated circuit chips and a memory con-
troller within a card that is removably
connectable to the host system said con-
troller being connectable to said proces-
sor for controlling operation of the
memory system when the card is con-
nected to the host system,
operating memory cells within individual sec-
tors with at least a user data portion and
an overhead portion,
causing the controller, in response to receipt
from the processor of an address in a
format designating at least one mass
memory storage block, to designate an
address of at least one non-volatile mem-
ory sector that corresponds with said at
least one mass memory storage block;
either writing user data to, or reading from,
the user data portion of said at least one
non-volatile memory sector; and
either writing to, or reading from, said over-
head portion of said at least one non-
volatile memory sector, overhead data re-
lated either to said at least one non-
volatile memory sector or to data stored
in the user data portion of said at least
one non-volatile memory sector.
’842 patent col.22 l.51-col.23 l.12 (emphases added).
SanDisk argues that the district court improperly fo-
cused on the claims’ use of the definite articles “the” and
19 SANDISK CORP v. KINGSTON TECH
“said” in connection with the user data portion and over-
head portion without looking at the language of the claim
as a whole. According to SanDisk, because the claims’
earlier references to the user data and overhead portions
use the indefinite articles “a” and “an,” under traditional
claim construction rules, those terms cover “one or more,”
not only one. This position, SanDisk contends, is sup-
ported by several dependent clams that expressly include
an “only one” user data portion and overhead portion
limitation. Lastly, SanDisk argues that the specification
suggests the possibility of multiple user data and over-
head data portions. In response, Kingston argues that
the claims’ use of indefinite articles does not assist San-
Disk because the specification only discloses a single user
data portion and a single overhead data portion. With
respect to SanDisk’s claim differentiation argument,
Kingston contends that the court’s construction does not
violate the doctrine of claim differentiation because the
independent claims “allow[] the system to include other
potential portions of information beside overhead and
user data.” Kingston’s Resp. Br. 36.
We agree with SanDisk that the court improperly lim-
ited the claims to only one user data portion and only one
overhead data portion. In its claim construction opinion,
the district court determined that “[t]he claim’s reference
to ‘the user data portion’ and ‘said overhead portion’
supports the proposed limitation.” J.A. 7. The court also
emphasized this use of “the” and “said” in concluding that
“the claim leaves no doubt that it covers a method involv-
ing only one user data portion and one overhead portion.”
J.A. 8.
This conclusion is contrary to Baldwin Graphics Sys-
tems, Inc. v. Siebert, 512 F.3d 1338 (Fed. Cir. 2008). In
Baldwin, we explained that the later use of “the” and
SANDISK CORP v. KINGSTON TECH 20
“said” to refer back to an earlier claim term does not limit
that claim term to the singular, and we also articulated
the general rule that the use of the indefinite articles “a”
or “an” means “one or more”:
[T]his court has repeatedly emphasized that an
indefinite article “a” or “an” in patent parlance
carries the meaning of “one or more” in open-
ended claims containing the transitional phrase
“comprising.” That “a” or “an” can mean “one or
more” is best described as a rule, rather than
merely as a presumption or even a convention.
The exceptions to this rule are extremely limited: a
patentee must “evince[] a clear intent” to limit “a”
or “an” to “one.” The subsequent use of definite ar-
ticles “the” or “said” in a claim to refer back to the
same claim term does not change the general plu-
ral rule, but simply reinvokes that non-singular
meaning.
Id. at 1342 (second alteration in original) (internal quota-
tion marks and citations omitted) (emphasis added).
Further, this general rule applies unless “the language of
the claims themselves, the specification, or the prosecu-
tion history necessitate[s] a departure from the rule.” Id.
at 1342-43. In this case, the intrinsic evidence does not
demonstrate an intention to exclude multiple user data
portions or overhead portions from the claims’ scope.
First, the claims recite “at least a user data and an
overhead data portion.” ’316 patent col.21 ll.40-41; ’842
patent col.22 ll.64-65. The phrase “at least” suggests that
the claim covers more than one user data portion and
overhead portion. See Biagro W. Sales, Inc. v. Grow More,
Inc., 423 F.3d 1296, 1304 (Fed. Cir. 2005) (“The phrase ‘at
least one’ in patent claims typically is construed to mean
21 SANDISK CORP v. KINGSTON TECH
‘one or more.’”). This interpretation further comports
with the general rule set forth in Baldwin against limit-
ing claim terms using the indefinite articles “a” and “an”
to mean “one.” 512 F.3d at 1342-43.
Dependent claims 16 and 67 of the ’842 patent and
claim 73 of the ’316 patent additionally bolster SanDisk’s
interpretation. Those claims add the limitation that
either the individual blocks (claim 16 of the ’842 patent)
or sectors (claim 67 of the ’842 patent and claim 73 of the
’316 patent) “include only one user data portion and only
one overhead portion.” ’842 patent col.18 ll.7-9, col.24 ll.7-
9; ’316 patent col.22 ll.14-16. Where, as here, the sole
difference between the independent claim and the de-
pendent claims is the limitation that one party is trying to
read into the independent claim, “the doctrine of claim
differentiation is at its strongest.” Liebel-Flarsheim Co. v.
Medrad, Inc., 358 F.3d 898, 910 (Fed. Cir. 2004); see also
Phillips, 415 F.3d at 1315 (“[T]he presence of a dependent
claim that adds a particular limitation gives rise to a
presumption that the limitation in question is not in the
independent claim.”).
Although the doctrine creates only a rebuttable pre-
sumption, see, e.g., Liebel-Flarsheim, 358 F.3d at 910,
here Kingston has not identified any intrinsic evidence
that overcomes this presumption and justifies its narrow
construction. To the contrary, we agree with SanDisk
that the specification suggests that the precise configura-
tion of the user data and overhead portions is not fixed:
It is to be understood that the partitioning be-
tween the user data portion 403 and the spare
[i.e., overhead] portion 405 need not be rigid. The
relative size of the various partitioned areas may
be logically reassigned. Also the grouping of the
SANDISK CORP v. KINGSTON TECH 22
various areas is largely for the purpose of discus-
sion and not necessarily physically so.
’842 patent col.8 ll.52-57 (emphasis added).
Thus, for the above reasons, we conclude that the dis-
trict court erred in its construction of the phrase “at least
a user data portion and an overhead portion” in claim 61
of the ’842 patent and claim 67 of the ’316 patent. Under
the correct construction, this claim limitation covers “one
or more” user data portion and overhead portion.
B. Summary Judgment of Non-Infringement
In addition to SanDisk’s claim construction argu-
ments, SanDisk raises two challenges to the district
court’s summary judgment decision. First, SanDisk
argues that the district court erred in applying the disclo-
sure-dedication rule from Johnson & Johnston Associates,
Inc. v. R.E. Service Co., 285 F.3d 1046 (Fed. Cir. 2002) (en
banc), in finding that Kingston’s accused products did not
infringe claim 20 of the ’424 patent under the doctrine of
equivalents. Second, SanDisk contends that the district
court erred in finding that the accused products did not
infringe claim 79 of the ’316 patent either literally or
under the doctrine of equivalents because they lacked a
“controller.” This second finding of non-infringement also
implicates Johnson & Johnston’s disclosure-dedication
rule. 2
2 On appeal, Kingston has maintained the con-
fidentiality of the specific design implementations of its
accused products. To preserve that confidentiality, we
limit our discussion of the products to counsels’ state-
ments during oral argument, see Oral Argument 12:20-
12:27, available at http://www.cafc.uscourts.gov/oral-
argument-recordings/2011-1346/all (“No kind of block
23 SANDISK CORP v. KINGSTON TECH
1. ’424 Patent, Claim 20
Claim 20 of the ’424 patent recites:
20. In a re-programmable non-volatile memory
system having a plurality of blocks of memory
storage elements that are erasable together as a
unit, the plurality of blocks individually being di-
vided into a plurality of a given number of pages
of memory storage elements that are programma-
ble together, a method of operating the memory
system, comprising:
programming individual ones of a first plural-
ity of said given number of pages in each
of at least a first block with original data
and a logical page address associated
with the original data,
thereafter programming individual ones of a
second plurality of a total number of
pages less than said given number in a
second block with updated data and a
logical page address associated with the
updated data, wherein the logical page
addresses associated with the updated
data programmed into the second plural-
ity of pages are the same as those associ-
ated with the original data programmed
into the first plurality of pages, and
thereafter reading and assembling data from
the first and second plurality of pages in-
marker alone or in combination with a logical block
address is in any way disclosed . . . .”), and the non-
confidential information in the parties’ briefs.
SANDISK CORP v. KINGSTON TECH 24
cluding, for pages having the same logi-
cal addresses, selecting the updated data
from the pages most recently pro-
grammed and omitting use of the origi-
nal data from the pages earlier
programmed.
’424 patent col.15 ll.40-64 (emphases added).
Claim 20 claims a method for performing partial block
updates. In a partial block update, data retains the same
logical address throughout each update. The controller
identifies the data sharing a logical page address and
provides the most updated data with that address to the
host device. To accomplish this, the logical page address
must identify a specific logical page within a block. It is
undisputed that a logical block number, by itself, does not
identify a logical page address. Instead, to provide a
logical page address, additional information must be
added to the logical block number, such as a logical page
offset.
The parties’ dispute centers on whether the accused
products contain an equivalent of the “programming . . . a
logical page address associated with the original data”
limitation in claim 20. The district court construed a
“logical page address” as not being limited to a “logical
block number plus logical offset.” J.A. 37.
During summary judgment, SanDisk argued that an
indication that the pages within a block are stored se-
quentially—referred to by SanDisk’s expert as a “Sequen-
tial Block Marker”—coupled with programming a logical
block address was equivalent to the “programming . . . a
logical page address” limitation in claim 20. According to
SanDisk, when the data is programmed sequentially, the
25 SANDISK CORP v. KINGSTON TECH
logical page matches the physical page. SanDisk’s Br. 53.
For data programmed into physical page 3 of logical block
6, for example, “[t]he logical page address is necessarily
logical block 6, logical page 3 because the logical page
address within the block (i.e., 3) matches the physical
address of the block.” Id. at 53 n.4. Therefore, in San-
Disk’s view, knowing both the logical block number and
that the pages are sequentially stored identifies the
logical page address. The district court, however, deter-
mined that SanDisk could not pursue this infringement
theory, finding that Figure 9 of the ’424 patent disclosed
this alleged equivalent and thus dedicated it to the public
under the disclosure-dedication rule set forth in Johnson
& Johnston.
On appeal, SanDisk argues that the disclosure in the
specification relied upon by the district court does not
amount to a dedication under Johnson & Johnston be-
cause neither Figure 9 relied on by the district court nor
the remainder of the specification discloses using a Se-
quential Block Marker in combination with a logical block
address to identify the logical page address. Instead,
according to SanDisk, all the disclosed embodiments,
including Figure 9 reproduced below, show that a logical
page address includes only (1) a logical block number
(LBN) and (2) a logical page offset:
SANDISK CORP v. KINGSTON TECH 26
In response, Kingston argues that Figures 4, 8, and 9
in the ’424 patent depict systems in which the logical
block number is programmed and the blocks are written
in sequential order, such that the logical page address
and the physical page address are the same. Kingston
contends that this information discloses to one of ordinary
skill SanDisk’s proposed equivalent to the “programming .
. . a logical page address” limitation.
We agree with Sandisk that its proposed equivalent
was not dedicated to the public. Under the doctrine of
equivalents, “a product or process that does not literally
infringe upon the express terms of a patent claim may
nonetheless be found to infringe if there is ‘equivalence’
between the elements of the accused product or process
and the claimed elements of the patented invention.”
Warner-Jenkinson Co. v. Hilton Davis Chem. Co., 520
U.S. 17, 21 (1997). A patentee, however, can disclaim an
equivalent by disclosing it in the specification. As we held
in Johnson & Johnston, “[W]hen a patent drafter dis-
closes but declines to claim subject matter, . . . this action
dedicates that unclaimed subject matter to the public.”
285 F.3d at 1054.
Johnson & Johnston’s disclosure-dedication rule is
not without restriction. In PSC Computer Products v.
Foxconn International, Inc., 355 F.3d 1353 (Fed. Cir.
2004), we explained that the rule “does not mean that any
generic reference in a written specification necessarily
dedicates all members of that particular genus to the
public.” Id. at 1360. Rather, “the disclosure must be of
such specificity that one of ordinary skill in the art could
identify the subject matter that had been disclosed and
not claimed.” Id. Additionally, in Pfizer, Inc. v. Teva
Pharmaceuticals USA, Inc., 429 F.3d 1364 (Fed. Cir.
2005), this court further clarified that “before unclaimed
27 SANDISK CORP v. KINGSTON TECH
subject matter is deemed to have been dedicated to the
public, that unclaimed subject matter must have been
identified by the patentee as an alternative to a claim
limitation.” Id. at 1379. Whether the disclosure-
dedication rule prevents a patentee from pursuing a
doctrine of equivalents infringement theory is a question
of law we review de novo. Id. at 1378.
Here, the disclosures in Figure 9 relied upon by the
district court and the other teachings in the specification
cited by Kingston do not satisfy the disclosure-dedication
rule’s requirements. Figure 9 and the accompanying
description in the specification teach a logical page ad-
dress composed of a logical block number and a logical
page offset. Neither Figure 9 nor any other portion of the
specification identified by Kingston refers to using an
indication in the address that the block is sequentially
programmed combined with programming a logical block
address as an alternative to “programming . . . a logical
page address.” Whether a person of ordinary skill ulti-
mately could employ the disclosures of the patent to
implement a purported equivalent does not amount to
actually disclosing to one of ordinary skill that equivalent
“as an alternative to a claim limitation.” See Pfizer, 429
F.3d at 1379. We have considered Kingston’s remaining
arguments and find they lack merit. Accordingly, we
conclude that the district court erred in finding that
SanDisk’s proposed equivalent to the “programming . . . a
logical page address limitation” was dedicated to the
public.
2. ’316 Patent, Claim 79
Claim 79 of the ’316 patent provides in pertinent part:
SANDISK CORP v. KINGSTON TECH 28
79. A memory system connectable to a host proc-
essor to enable the exchange of data therebe-
tween, and memory system comprising:
an array of non-volatile floating gate memory
cells partitioned into a plurality of blocks
of cells that individually store a given
amount of user data and overhead data,
wherein the memory cells are individu-
ally programmable into one of more than
two distinct threshold level ranges corre-
sponding to more than one bit of data per
cell,
a controller connected to the array and re-
movably connectable to the host through
an electrical connector, said controller
including:
an address generator that is respon-
sive to receipt of a mass memory
storage block address from the
host to address a corresponding
at least one of the plurality of
memory blocks, and
a data transfer control that responds
to an instruction from the host to
perform a designated one of read-
ing user data from, or writing
user data to, said at least one ad-
dress block, including a data writ-
ing circuit that generates at least
some of the overhead data associ-
ated with at least one of at least
one addressed block or user data
29 SANDISK CORP v. KINGSTON TECH
being written therein, and a data
reading circuit that reads the
overhead data from said at least
one addressed block, wherein the
data writing circuit programs the
individual memory cells into said
one or more than two distinct
threshold level ranges and the
data reading circuit reads one of
more than two distinct threshold
level ranges form the individual
memory cells.
’316 patent col.22 ll.35-79 (emphases added).
The memory system in claim 79 includes two main
limitations: (1) an array of non-volatile floating gate
memory cells and (2) a controller. The controller has both
an “address generator” and a “data transfer control.”
Within the data transfer control are a “data writing
circuit” and a “data reading circuit.” It is undisputed that
the accused devices contain data writing and data reading
circuits, but those circuits are located on the flash mem-
ory chip, not the controller chip. The key issue is whether
the read and write circuitry in the accused products are
part of the claimed “controller,” either literally or under
the doctrine of equivalents, even though the circuitry is
located on the flash memory chip.
The district court recognized that the circuitry for the
controller could be located on more than one chip but
found that the accused products did not literally infringe
because “[t]here is no evidence that the data writing and
reading circuitry on the flash memory chips of the accused
products are otherwise part of the ‘controller.’” J.A. 60.
With respect to infringement under the doctrine of
SANDISK CORP v. KINGSTON TECH 30
equivalents, the district court again applied the disclo-
sure-dedication rule of Johnson & Johnston, finding that
the ’316 patent, through the incorporated by reference
U.S. Patent No. 5,172,338 (“’338 patent”), 3 disclosed but
did not claim programming and reading that is that is
regulated on the separate flash memory chip. J.A. 61.
On appeal, SanDisk argues that the district court’s
literal infringement analysis is irreconcilable with its
doctrine of equivalents analysis. According to SanDisk, if,
as the district court found, the ’316 patent discloses
controller circuitry on the flash memory chip, claim 79
should be read to cover that embodiment; on the other
hand, if such an embodiment is not disclosed, the disclo-
sure-dedication rule does not apply to bar infringement
under the doctrine of equivalents. SanDisk further
contends that because the claims define the controller as
having a read circuit and a write circuit, the accused
products—which include read and write circuitry—
necessarily infringe.
In response, Kingston challenges SanDisk’s framing
of the question, arguing that the correct inquiry is not
whether the claimed controller must reside on one chip,
as SanDisk contends, but rather whether the writing and
reading circuits in the accused products are a part of the
controller. Kingston maintains that, contrary to San-
Disk’s position, the presence of read and write circuits
somewhere in the accused system does not mean that
3 The ’316 patent expressly incorporates U.S.
Patent Application No. 07/337,579 (“’579 application”) by
reference. See ’316 patent col.6 ll.3-9, col.11 ll.4-11. The
’338 patent is a continuation-in-part of the ’579 applica-
tion. On appeal, the parties do not dispute that the ’338
patent is incorporated by reference into the ’316 patent.
31 SANDISK CORP v. KINGSTON TECH
those circuits necessarily are part of the controller as
required by the claim. Kingston also disputes SanDisk’s
position that the court’s doctrine of equivalents analysis is
incongruous with the literal infringement analysis.
According to Kingston, the specification discloses two
embodiments, and only one of those embodiments is
claimed. In the first embodiment, the read and write
circuits are part of the controller; in the second unclaimed
embodiment, disclosed in Figure 5 of the incorporated
’338 patent, the read and write circuits are separate from
the controller.
With respect to literal infringement, we agree with
Kingston that the district court correctly found that
SanDisk failed to present evidence that the read and
write circuitry in the accused devices were part of the
controller such that claim 79 of the ’316 patent was liter-
ally infringed. Although the district court concluded that
the controller circuitry need not be located on the same
chip, that construction does not resolve the infringement
issue. The read and write circuits are not only part of the
claimed system; the claims explicitly require that these
circuits be part of the claimed controller. Thus, the
presence of a read circuit and a write circuit somewhere
in the system does not establish that those circuits are
part of the controller for purposes of establishing in-
fringement.
The district court correctly treated this infringement
issue as a question of fact and found that SanDisk failed
to present any evidence on summary judgment that “the
data writing and reading circuit on the flash memory
chips of the accused products are otherwise part of the
‘controller.’” J.A. 60. On appeal, SanDisk has not identi-
fied any evidence that the district court overlooked in
reaching this conclusion. Consequently, we affirm the
SANDISK CORP v. KINGSTON TECH 32
district court’s judgment in favor of Kingston as to literal
infringement of claim 79.
We reach a different conclusion, however, with re-
spect to the district court’s application of the disclosure-
dedication rule in its doctrine of equivalents analysis.
The district court found that Kingston’s accused products
did not infringe under the doctrine of equivalents because
the ’316 patent disclosed the use of read and write cir-
cuitry located on the flash memory chip rather than the
controller chip. The disclosure relied upon by the district
court does not appear within the four corners of the ’316
patent but instead is in Figure 5 of the incorporated ’338
patent.
We have yet to address the circumstances in which
the disclosure of subject matter in a document incorpo-
rated by reference amounts to a dedication of that subject
matter under Johnson & Johnston. Because a document
incorporated by reference “becomes effectively part of the
host document as if it were explicitly contained therein,”
Telemac Cellular Corp. v. Topp Telecom, Inc., 247 F.3d
1316, 1329 (Fed. Cir. 2001), the disclosure of subject
matter in an incorporated document can dedicate that
subject matter to the public for purposes of the host
patent. Incorporation by reference, however, “does not
convert the invention of the incorporated patent into the
invention of the host patent.” Modine Mfg. Co. v. Int’l
Trade Comm’n, 75 F.3d 1545, 1552 (Fed. Cir. 1996).
Thus, in determining whether incorporated subject mat-
ter satisfies the disclosure-dedication rule standards set
forth in Johnson & Johnston and its progeny, we must
look first to the teachings of the host patent. Consistent
with Pfizer, the host patent must sufficiently inform one
of ordinary skill that the incorporated document contains
subject matter that is an alternative to a claim limitation.
33 SANDISK CORP v. KINGSTON TECH
429 F.3d at 1379. If it does, the inquiry then shifts to the
incorporated document to assess whether the disclosure of
that subject matter is “of such specificity that one of
ordinary skill in the art could identify the subject matter
that had been disclosed and not claimed.” PSC Computer
Prods., 355 F.3d at 1360.
Here, the ’316 patent’s discussion of the incorporated
’338 patent does not sufficiently identify to one of ordi-
nary skill that the incorporated patent contains subject
matter that is an alternative to the claimed controller. To
the contrary, in discussing the parent ’579 application,
the ’316 patent speaks only in general terms: “Optimized
erase implementations have been disclosed in two copend-
ing U.S. patent applications,” ’316 patent col.6 ll.3-4; and
“Optimized implementations of write operation for Flash
EEprom device have been disclosed in two previously
cited co-pending U.S. applications . . . ,” id. at col.11 ll.4-6.
Such cursory discussion does not sufficiently provide
notice to one of ordinary skill that the incorporated ’338
patent contains subject matter that is an alternative to
the claimed controller, as required by Pfizer. Conse-
quently, we conclude that the district court erred as a
matter of law in finding that the ’316 patent dedicated
SanDisk’s proposed equivalent to the public.
III. CONCLUSION
We lack jurisdiction to address SanDisk’s claim con-
struction arguments that are solely related to the ’893
and ’808 patents and claims 1 and 10 of the ’842 patent.
As to the claim construction issues that are properly
before us on appeal, we conclude that the district court
erred in its constructions of the “recording a relative time
of programming . . .” limitation in claims 1 and 3 of the
’424 patent, and the “at least a user data portion and an
SANDISK CORP v. KINGSTON TECH 34
overhead portion” limitation in claim 61 of the ’842 patent
and claim 67 of the ’316 patent. Those constructions
accordingly are reversed and the judgment of non-
infringement in favor of Kingston as to those claims is
vacated. Lastly, the district court’s judgment of no literal
infringement of claim 79 of the ’316 patent is affirmed,
but the court’s judgment that Kingston did not infringe
claim 20 of the ’424 patent and claim 79 of the ’316 patent
under the doctrine of equivalents is vacated. The case is
remanded to the district court for further proceedings
consistent with this opinion.
AFFIRMED-IN-PART, VACATED-IN-PART, AND
REMANDED
United States Court of Appeals
for the Federal Circuit
__________________________
SANDISK CORPORATION,
Plaintiff-Appellant,
v.
KINGSTON TECHNOLOGY CO., INC.
AND KINGSTON TECHNOLOGY CORP.,
Defendants-Appellees.
__________________________
2011-1346
__________________________
Appeal from the United States District Court for the
Western District of Wisconsin in consolidated case nos.
07-CV-0605 and 07-CV-0607, Senior Judge Barbara B.
Crabb.
__________________________
REYNA, Circuit Judge, concurring-in-part and dissenting-
in-part.
I concur with the majority opinion except for its con-
struction of the term “recording a relative time of pro-
gramming” in claims 1 and 3 of the ’424 patent. The
majority construes this term to include an embodiment in
which no indication, temporal or otherwise, is recorded.
From this conclusion I respectfully dissent.
Claim 1 is representative of claims 1 and 3 of the ’424
patent:
SANDISK CORP v. KINGSTON TECH 2
1. In a non-volatile memory system having a plu-
rality of blocks of memory storage elements that
are individually erasable as a unit and which are
individually organized into a plurality of pages of
memory storage elements that are individually
programmable together, a method of substituting
new data for superceded data within at least one
page of one of the plurality of blocks while data in
at least another page of said one block is not re-
placed, comprising:
programming the new data into at least one
page of said one or another of the plural-
ity of blocks,
identifying the at least one page of superceded
data and the at least one page of new data
by a common logical address,
recording a relative time of programming the
at least one page of new data and the at
least one page of superceded data; and
wherein the at least one page of superceded
data is less than all the data contained in
said one block.
’424 patent col.12 l.60-col.13 l.10 (emphasis added).
The specification describes two embodiments. The
first is faithful to the claim language, recording a time
stamp in each individual page. The second, called the
“Block Recording Method,” is not. According to the speci-
fication, in the block recording method, “the time stamp
. . . does not need to be stored as part of each page” and “is
used only to determine the relative age of the data stored
in blocks.” Id. at col.9 ll.41-47. The block time stamp is
3 SANDISK CORP v. KINGSTON TECH
“recorded for each block, either as part of the block or
elsewhere within the non-volatile memory, and is updated
each time a page of data is written into the block.” 1 Id. at
col.9 ll.42-50. In the block recording method, pages are
written in order within the block, and the most current
data for a logical page within a block will always be the
last physical page in the block containing data for that
logical page. However, since the physical pages do not
contain any timestamp data, relative or otherwise, all
that can be said is that the pages were written at the
same time or after the time stamp in the physical block.
In other words, although the relative order can be in-
ferred from the physical number of the page, nothing is
known about the relative times.
The majority avoids the distinction between “order”
and “relative time” by focusing on the “relative” limitation
in the claim and ignoring the limitations that the time
must be recorded and that it must be a time. Thus, the
majority states that “‘recording a relative time of pro-
gramming’ merely requires recording some indication of
the order of programming for those pages sharing a
logical address.” Majority Op. at 14. Because recording
1 Updating the timestamp for a block once it is
written appears to be impossible. A cell must be erased
before it can be re-written. See ’424 patent col.1 ll.23-24.
Cells can only be erased by erasing an entire block. See
id. Therefore, to update the timestamp in a block when a
page in that block is written, it would be necessary to
erase the entire block, including the newly written data.
It is possible, as the specification indicates, to store the
relative time elsewhere in non-volatile memory. That is
exactly what the first embodiment does, and it would
appear that storage of the timestamps in any location
other than the page being written would be much less
efficient and would result in additional wear on the flash
memory.
SANDISK CORP v. KINGSTON TECH 4
new updates in the next available page in the most recent
block implicitly encodes the order in which the updates
were received, the majority concludes that the claim was
met. This construction improperly ignores express limita-
tions of the claims and uses the specification to broaden
the patent. See, e.g., Maxwell v. J. Baker, Inc., 86 F.3d
1098, 1105 (Fed. Cir. 1996) (rejecting a claim interpreta-
tion that would ignore explicit limitations in the claim);
Unique Concepts, Inc. v. Brown, 939 F.2d 1558, 1562 (Fed.
Cir. 1991) (“All the limitations of a claim must be consid-
ered meaningful.”); see also United States v. Adams, 383
U.S. 39 (1966) (stating that claims limit the inventions
and “specifications cannot be used to expand the patent
monopoly”).
The majority then attempts to explain how, under this
reading, the block recording method satisfies the lan-
guage of the claim. In particular, the majority discusses
Figure 11, which depicts a second update to logical page 5
within the same physical block (PBN1) where the physical
location of the two pages within the block identifies the
most recently updated page:
5 SANDISK CORP v. KINGSTON TECH
Specifically, in Figure 11, the new physical block
(PBN1) contains two versions of original page 5 located at
PBN1 physical page 2 and PBN1 physical page 3. In this
example, page 5 on the left has been modified twice. The
first modification is stored at page 2 on the right. The
second is stored at page 3 on the right. Because the data
will be read backwards starting from the last page of the
new block, the user data on PBN1 physical page 3 will be
read, but the data on PBN1 physical page 2 will not.
If anything, the majority’s analysis of this embodi-
ment proves that the claim does not cover the block
recording method. In PBN1, pages 2 and 3 were written
at two different times. When page 2 was written, a
relative time was recorded for the new physical block.
However, when page 3 was written, no relative time was
recorded. The relative times of the updates of pages 2 and
3 are unknown and unknowable, because no relative time
SANDISK CORP v. KINGSTON TECH 6
information is recorded at the page level. Instead, the
system infers the order—as opposed to the relative
times—in which the data was written. This is undoubt-
edly more elegant than recording relative times, but it
does not satisfy the terms of the claim, which require that
a relative time be recorded.
Even if it is possible to update the time for the block,
this does not save the embodiment. If the block time is
not updated, time information is only available for the
first physical page in the block. If the block time is up-
dated, time information is only available for the last page
that has been written in the block. In either case, all that
can be determined is the order of the pages, not their
relative times. Indeed, in describing the block recording
method, the specification notes that “[o]nly . . . the rela-
tive times that the physical blocks were programmed need
to be known.” ’424 Patent col.10 ll.3-7 (emphasis added).
Thus, in the block recording method, relative times are
recorded for physical blocks, not for pages. If a page
update does not require a new block, no relative time is
recorded, and the terms of the claim are not met.
The majority focuses only on the term “relative” and
ignores that the claims explicitly require a time to be
recorded. In doing so, it improperly expands SanDisk’s
patent monopoly beyond what was claimed. I respect-
fully dissent.