Full opinion text
Opinion for the court filed by Circuit Judge REYNA. Opinion dissenting in part filed by Chief Judge RADER. Opinion concurring in part and dissenting in part filed by Circuit Judge PROST. REYNA, Circuit Judge. Plaintiffs Apple Inc. and Next Software, Inc. (“Apple”) filed a complaint against Defendants Motorola, Inc. and Motorola Mobility, Inc. (“Motorola”) on October 29, 2010 in the United States District Court for the Western District of Wisconsin, asserting infringement of three patents. Motorola counterclaimed, asserting six of its own patents. Apple amended its complaint to include an additional twelve patents. Both parties also sought declaratory judgments of non-infringement and invalidity. After claim construction began in Wisconsin, the case was transferred to the United States District Court for the Northern District of Illinois, Judge Posner sitting by designation. The district court in Illinois completed claim construction. Based upon its claim construction decisions, the court granted summary judgment of non-infringement with respect to certain claims and excluded the vast majority of both parties’ damages expert evidence for the remaining claims. With little expert evidence deemed admissible, the court granted summary judgment that neither side was entitled to any damages or an injunction. Despite infringement being assumed, the district court dismissed all claims with prejudice before trial. Only six patents are at issue on appeal: Apple’s U.S. Patent Nos. 7,479,949; 6,343,-263; and 5,946,647; and Motorola’s U.S. Patent Nos. 6,359,898; 6,175,559; and 5,319,712. The parties contest the district court’s claim construction, admissibility, damages, and injunction decisions. As detailed below, we affirm the district court’s claim construction decisions, with the exception of its construction of Apple’s '949 patent. With a minor exception, the district court’s decision to exclude the damages evidence presented by both Apple and Motorola is reversed. We also reverse the district court’s grant of summary judgment of no damages for infringement of Apple’s patents. Based upon our reversal of the district court’s claim construction of the '949 patent, we vacate the court’s grant of summary judgment regarding Apple’s request for an injunction. The court’s decision that Motorola is not entitled to an injunction for infringement of the FRAND-committed '898 patent is affirmed. We address these, and all related issues, in turn. Claim Construction The parties raise claim construction issues regarding Apple’s '949, '263, and '647 patents and Motorola’s '559 and '712 patents. Claim construction is a question of law that we review de novo. Cybor Corp. v. FAS Techs., Inc., 138 F.3d 1448, 1454-55 (Fed.Cir.1998) (en banc). Apple’s '949 patent The district court construed claims 1, 2, 9, and 10 of the '949 patent and, based upon its construction, granted Motorola’s motion for summary judgment of nonin-fringement for the majority of the accused products. Because the district court mistakenly construed certain limitations as means-plus-function limitations, we reverse its claim construction and vacate the subsequent summary judgment decision. The '949 patent discloses the use of finger contacts to control a touchscreen computer. Claims 1, 2, 9, and 10 of the '949 patent are recited below, with the limitations at issue emphasized. Claim 1 recites: A computing device, comprising: a touch screen display; one or more processors; memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including: instructions for detecting one or more finger contacts with the touch screen display; instructions for applying one or more heuristics to the one or more finger contacts to determine a command for the device; and instructions for processing the command; wherein the one or more heuristics comprise: a vertical screen Scrolling heuristic for determining that the one or more finger contacts correspond to a one-dimensional vertical screen scrolling command rather than a two-dimensional screen translation command based on an angle of initial movement of a finger contact with respect to the touch screen display; a two-dimensional screen translation heuristic for determining that the one or more finger contacts correspond to the two-dimensional screen translation command rather than the one-dimensional vertical screen scrolling command based on the angle of initial movement of the finger contact with respect to the touch screen display; and a next item heuristic for determining that the one or more finger contacts correspond to a command to transition from displaying a respective item in a set of items to displaying a next item in the set of items. '949 patent at col. 122 1. 37-col. 123 1. 2 (emphases added). Claim 2 recites: The computing device of claim 1, wherein the one or more heuristics include a heuristic for determining that the one or more finger contacts correspond to a command to translate content within a frame rather than translating an entire page that includes the frame. Id. at col. 123, Ins 3-7 (emphasis added). Claim 10 recites: The computing device of claim 9, wherein the first set of heuristics comprises a heuristic for determining that the one or more first finger contacts correspond to a one-dimensional horizontal screen scrolling command rather than the two-dimensional screen translation command based on the angle of initial movement of the finger contact with respect to the touch screen display. Id. at col. 124, 11. 1-7. The district court first found that the claim term “heuristic” was not indefinite, instead construing it as “one or more rules to be applied to data to assist in drawing inferences from that data.” Next, the court found that the “heuristic” limitations in claims 1, 2, 9, and 10 described functions “without describing the structure necessary to perform the functions.” Accordingly, the court concluded that these claim limitations were means-plus-function limitations under 35 U.S.C. § 112, ¶ 6, despite not reciting the word “means.” The court next found that the specification contained sufficient “corresponding structure” capable of performing the claimed functions. 35 U.S.C. § 112. In doing so, the court limited the “next item heuristic” in claim 1 to “a heuristic that uses as one input a user’s finger tap on the right side of the device’s touch screen.” Based upon this construction, Motorola moved for summary judgment of noninfringement. The court concluded that the only accused products that use a “finger tap” in this manner are those that come pre-loaded with a specific program: the Amazon Kindle application. The court granted Motorola’s motion for summary judgment of non-infringement for the remaining accused products. On appeal, Motorola again argues that “heuristic” is indefinite. In the alternative, Motorola argues that the district court correctly concluded that the heuristic limitations were drafted in means-plus-function format and correctly limited the “next item heuristic” limitation to the finger tap gesture. Apple points out that the claims do not use the word “means” and that this creates a strong presumption against construing the limitations as means-plus-function limitations. Apple argues that the heuristic limitations connote sufficiently definite structure such that Motorola has not overcome this strong presumption. Whether claim language invokes Section 112, ¶ 6 is a question of law that we review de novo. Inventio AG v. ThyssenKrupp Elevator Americas Corp., 649 F.3d 1350, 1356 (Fed.Cir.2011); Lighting World, Inc. v. Birchwood Lighting, Inc., 382 F.3d 1354, 1358 (Fed.Cir.2004). Section 112, ¶ 6 states: An element in a claim for a combination may be expressed as a means or step for performing a specified function without the recital of structure, material, or acts in support thereof, and such claim shall be construed to cover the corresponding structure, material, or acts described in the specification and equivalents thereof. 35 U.S.C. § 112, ¶ 6. The overall means-plus-function analysis is a two-step process. Naturally, there is some analytical overlap between these two steps. In the first step, we must determine if the claim limitation is drafted in means-plus-function format. As part of this step, we must construe the claim limitation to decide if it connotes “sufficiently definite structure” to a person of ordinary skill in the art, which requires us to consider the specification (among other evidence). In the second step, if the limitation is in means-plus-function format, we must specifically review the specification for “corresponding structure.” Thus, while these two “structure” inquiries are inherently related, they are distinct. The Dissent is concerned that we have impermissibly looked for corresponding structure in the specification before deciding that the claim is in means-plus-function format thereby creating a new rule that renders “every means-plus-function claim term indefinite.” J. Prost Dissent at 1335-36 (emphasis in original) (“Dissent”). This is not our analysis. The Dissent correctly notes that the first step in the means-plus-function analysis requires us to determine whether the entire claim limitation at issue connotes “sufficiently definite structure” to a person of ordinary skill in the art. Dissent at 1334-35. In so doing, we naturally look to the specification, prosecution history, and relevant external evidence to construe the limitation. While this inquiry may be similar to looking for corresponding structure in the specification, our precedent requires it when deciding whether a claim limitation lacking means connotes sufficiently definite structure to a person of ordinary skill in the art. See, e.g., Inventio, 649 F.3d at 1357 (“It is proper to consult the intrinsic record, including the written description, when determining if a challenger has rebutted the presumption that a claim lacking the term ‘means’ recites sufficiently definite structure.”); Lighting World, 382 F.3d at 1360-64 (examining the written description and external evidence); Flo Healthcare Solutions, LLC v. Kappos, 697 F.3d 1367, 1374 (Fed.Cir.2012) (examining remaining claim language, written description, and external evidence); Linear Tech. Corp. v. Impala Linear Corp., 379 F.3d 1311, 1320 (Fed.Cir.2004) (examining remaining claim language and external evidence). Because these inquiries are distinct, it is possible to find that a claim limitation does not connote sufficiently definite structure despite the presence of some corresponding structure in the specification. See, e.g., Massachusetts Inst. of Tech. v. Abacus Software, 462 F.3d 1344, 1355 (Fed.Cir.2006) (“MIT’); Welker Bearing Co. v. PHD, Inc., 550 F.3d 1090, 1096-97 (Fed.Cir.2008). As such, not “every” mean-plus-function limitation is indefinite under our precedent; only those that lack the term means, do not connote sufficiently definite structure, and lack corresponding structure. We do not state or apply a different rule in this case. In this case, as we find that the claims connote sufficiently definite structure to a person of ordinary skill in the art, we do not reach the second step of the means-plus-function analysis. As the district court recognized, when a claim limitation lacks the term “means,” it creates a rebuttable presumption that Section 112, ¶ 6 does not apply. See, e.g., Lighting World, 382 F.3d at 1358; CCS Fitness, Inc. v. Brunswick Corp., 288 F.3d 1359, 1369 (Fed.Cir.2002). This presumption may be overcome if the claim fails to recite “sufficiently definite structure” or merely recites a “function without reciting sufficient structure for performing that function.” Linear, 379 F.3d at 1319 (quoting Watts v. XL Sys. Inc., 232 F.3d 877, 880 (Fed.Cir.2000)); see also Inventio, 649 F.3d at 1356. We have repeatedly characterized this presumption as “strong” and “not readily overcome” and, as such, have “seldom” held that a limitation without recitation of “means” is a means-plus-function limitation. Lighting World, 382 F.3d at 1358, 1362; Inventio, 649 F.3d at 1356; see also Flo Healthcare, 697 F.3d at 1374 (“When the claim drafter has not signaled his intent to invoke § 112, ¶ 6 by using the term ‘means,’ we are unwilling to apply that provision without a showing that the limitation essentially is devoid of anything that can be construed as structure.”). The Dissent suggests that choosing to include “means” in a claim limitation is a “minor drafting decision” that correspondingly merits little weight in a Section 112, ¶ 6 analysis. Dissent at 1337. We disagree. The strong presumption created by not including means in a claim limitation provides clarity and predictability for the public and the patentee alike. It helps the public determine when claim elements are expressly limited to structures disclosed in the specification (or their equivalents) and provides the patentee with the tools for reliably invoking or avoiding means-plus-function claiming. It also signals to the court that the patentee has chosen to avail, or avoid, the benefits of Section 112, ¶ 6. We recognize that the choice to draft a claim in “broad structural terms” rather than in a means-plus-function format may render the claim more vulnerable to an invalidity attack. Id. Whether to draft a claim in broad structural terms is the claim drafter’s choice, and any resulting risk that emanates from that choice is not a basis for the court to rewrite a claim in means-plus-function format. See id. By focusing on the claim terms the patentee chose, this presumption also reaffirms the primacy of the claim language during claim construction, as outlined in Phillips v. AWH Corp. 415 F.3d 1303 (Fed.Cir.2005). Here, as in all aspects of claim construction, “the name of the game is the claim.” In re Hiniker Co., 150 F.3d 1362, 1369 (Fed.Cir.1998) (quoting Giles Sutherland Rich, Extent of Protection and Interpretation of Claims-American Perspectives, 21 Int’l Rev. Indus. Prop. & Copyright L. 497, 499 (1990)). In this case, Motorola bears the burden of overcoming the presumption that Section 112, ¶ 6 does not apply by a preponderance of the evidence. See Apex Inc. v. Raritan Computer, Inc., 325 F.3d 1364, 1372 (Fed.Cir.2003). The district court made several erroneous findings that led it to incorrectly conclude that Motorola rebutted this strong presumption. The district court misapplied our precedent by requiring the claim limitations of the '949 patent themselves to disclose “a step-by-step algorithm as required by Aristocrat Technologies.” Aristocrat and related cases hold that, if a patentee has invoked computer-implemented means-plus-function claiming, the corresponding structure in the specification for the computer implemented function must be an algorithm unless a general purpose computer is sufficient for performing the function. Aristocrat Techs. Austl. Pty Ltd. v. Int’l Game Tech., 521 F.3d 1328, 1333 (Fed.Cir.2008) (requiring disclosure of an algorithm when it is not disputed that claims were drafted in means-plus-function format); WMS Gaming, Inc. v. Int’l Game Tech., 184 F.3d 1339, 1349 (Fed.Cir.1999) (same); ePlus, Inc. v. Lawson Software, Inc., 700 F.3d 509, 518 (Fed.Cir.2012); Typhoon Touch Techs., Inc. v. Dell, Inc., 659 F.3d 1376, 1385 (Fed.Cir.2011). But see In re Katz Interactive Call Processing Patent Litig., 639 F.3d 1303, 1316 (Fed.Cir.2011) (finding that disclosure of a general purpose computer is sufficient corresponding structure for means-plus-function claims). In all these cases, the claims recited the term “means,” thereby expressly invoking means-plus-function claiming. In addition, the parties in these cases did not dispute on appeal that these claims were drafted in means-plus-function format. Hence, where a claim is not drafted in means-plus-function format, the reasoning in the Aristocrat line of cases does not automatically apply, and an algorithm is therefore not necessarily required. The correct inquiry, when “means” is absent from a limitation, is whether the limitation, read in light of the remaining claim language, specification, prosecution history, and relevant extrinsic evidence, has sufficiently definite structure to a person of ordinary skill in the art. Here, the answer is yes. “Structure” to a person of ordinary skill in the art of computer-implemented inventions may differ from more traditional, mechanical structure. For example, looking for traditional “physical structure” in a computer software claim is fruitless because software does not contain physical structures. Indeed, the typical physical structure that implements software, a computer, cannot be relied upon to provide sufficiently definite structure for a software claim lacking “means.” Rather, to one of skill in the art, the “structure” of computer software is understood through, for example, an outline of an algorithm, a flowchart, or a specific set of instructions or rules. See, e.g., Typhoon Touch, 659 F.3d at 1385 (“[T]he patent need only disclose sufficient structure for a person of skill in the field to provide an operative software program for the specified function.”); Finisar Corp. v. DirecTV Grp., Inc., 523 F.3d 1323, 1340 (Fed.Cir.2008). Requiring traditional' physical structure in software limitations lacking the term means would result in all of these limitations being construed as means-plus-function limitations and subsequently being found indefinite. A limitation has sufficient structure when it recites a claim term with a structural definition that is either provided in the specification or generally known in the art. See, e.g., Flo Healthcare, 697 F.3d at 1374 (“We will not apply § 112, ¶ 6 if the limitation contains a term that ‘is used in common parlance or by persons of skill in the pertinent art to designate structure.’ ”) (quoting Lighting World, 382 F.3d at 1359); Personalized Media, 161 F.3d at 704-05. In Personalized Media, we found that the claim term “detector,” by itself, connoted sufficient structure to a person of ordinary skill in the art. 161 F.3d at 704-05 (agreeing with ALJ that “ ‘detector’ had a well-known meaning to those of skill in the electrical arts connotative of structure”). There, we contrasted the structural term “detector” with generic, non-structural, terms such as “means,” “element,” and “device.” Id. at 705; see also Apex, 325 F.3d at 1373 (finding that the term “circuit,” coupled with identifiers such as “interface,” “programming,” and “logic,” connoted sufficient structure to a person of ordinary skill in the art). Structure may also be provided by describing the claim limitation’s operation, such as its input, output, or connections. The limitation’s operation is more than just its function; it is how the function is achieved in the context of the invention. For example, in Linear, we found that the claim term “circuit” has a known structural definition and that the patent described the circuit’s operation, including its input, output, and objective. 379 F.3d at 1320-21. Similarly, in Lighting World, we found that “connector” had a known structural definition and that the specification described its operational requirements, including which claim elements it was connected to and how they were connected. 382 F.3d at 1361-63. In both cases, we found the presumption against means-plus-function claiming was unrebutted. Even if a patentee elects to use a “generic” claim term, such as “a nonce word or a verbal construct,” properly construing that term (in view of the specification, prosecution history, etc.) may still provide sufficient structure such that the presumption against means-plus-function claiming remains intact. Id. at 1360; see also Inventio, 649 F.3d at 1356-57 (“Claims are interpreted in light of the written description supporting them, and that is true whether or not the claim construction involves interpreting a ‘means’ clause.”); MIT, 462 F.3d at 1354 (“The generic terms ‘mechanism,’ ‘means,’ ‘element,’ and ‘device,’ typically do not connote sufficiently definite structure.”). For example, in Inventio, the claim included the generic term “device.” 649 F.3d at 1354 (reciting “at least one modernizing device and connecting the at least one modernizing device to said floor terminals and said at least one computing unit.”) (emphasis added). However, the specification described the modernizing device’s input, output, internal components, and how the internal components were interconnected. Id. at 1358-59. As such, the presumption against means-plus-function treatment was not overcome. See also Flo Healthcare, 697 F.3d at 1374-75 (noting that “mechanism” is a generic term, but then looking to remaining claim language and written description before finding that the full claim limitation connoted structure). These cases teach that, if a limitation recites a term with a known structural meaning, or recites either a known or generic term with a sufficient description of its operation, the presumption against means-plus-function claiming remains intact. The limitation need not connote a single, specific structure; rather, it may describe a class of structures. See, e.g., Personalized Media Communications, LLC v. Int’l Trade Comm’n, 161 F.3d 696, 705 (Fed.Cir.1998) (“Even though the term ‘detector’ does not specifically evoke a particular structure, it does convey to one knowledgeable in the art a variety of structures known as ‘detectors.’ ”); Flo Healthcare, 697 F.3d at 1374-75 (finding that claim term “height adjustment mechanism” designates “a class of structures that are generally understood to persons of skill in the art”). Indeed, even if the patent describes all structures that perform the recited function, this, by itself, does not overcome the strong presumption that means-plus-function claiming does not apply when the term “means” is not recited in the claim. Lighting World, 382 F.3d at 1361-62. By contrast, if the claim merely recites a generic nonce word and the remaining claim language, specification, prosecution history, and relevant external evidence provide no further structural description to a person of ordinary skill in the art, then the presumption against means-plus-function claiming is rebutted. In MIT, for example, the claims recited a “colorant selection mechanism.” 462 F.3d at 1353. As noted, “mechanism” by itself does not connote sufficient structure, and the term “colorant selection” was not defined in the specification or otherwise known to a person of ordinary skill in the art. Id. at 1353-55. Further, the paten-tee used the terms “mechanism” and “means” interchangeably in the specification. Id. at 1354; see also Mas-Hamilton Grp. v. LaGard Inc., 156 F.3d 1206, 1214-16 (Fed.Cir.1998) (claim recited “element” and “member” and patent provided no further structural description of these generic terms); Welker Bearing, 550 F.3d at 1096-97 (claim recited a “mechanism” without further structure described in specification). Thus, if a claim recites a generic term that, properly construed in light of the specification, lacks sufficiently definite structure to a person of ordinary skill in the art, the presumption is overcome and the patentee has invoked means-plus-function claiming. With this precedent in mind, we turn to the claim limitations at issue in the '949 patent. We find that “heuristic” has a known meaning and the '949 patent also describes the limitation’s operation, including its input, output, and how its output may be achieved. Accordingly, the heuristic claim limitations recited above have “sufficiently definite structure,” to a person of ordinary skill in the art, for performing the recited functions. Broadly speaking, the function of the recited limitations is to identify a command based upon particular finger contacts. To achieve this function, the patent describes “heuristics.” Depending upon the circumstances, heuristic is not necessarily a generic, structureless “nonce word or a verbal construct” without any meaning, such as “mechanism,” “means,” “element,” or “widget.” The district court correctly determined that a person of ordinary skill in the art would understand “heuristic” to mean “one or more rules to be applied to data to assist in drawing inferences from that data.” In this sense, “heuristic” is similar to words that define a class of structures, such as “connector,” “circuit,” and “detector,” and it does not include all means for performing the recited function. See, e.g., Flo Healthcare, 697 F.3d at 1374 (“We will not apply § 112, ¶ 6 if the limitation contains a term that ‘is used in common parlance or by persons of skill in the pertinent art to designate structure’ ”) (quoting Lighting World, 382 F.3d at 1359); Personalized Media, 161 F.3d at 704-05. The fact that heuristic is defined partly in terms of its function does not detract “from the definiteness of [the] structure” it may connote. Personalized Media, 161 F.3d at 703-05. Indeed, “many devices take their names from the functions they perform.” Greenberg v. Ethicon Endo-Surgery, Inc., 91 F.3d 1580, 1583 (Fed.Cir.1996); see also MIT, 462 F.3d at 1354; Lighting World, 382 F.3d at 1359-60. We need not decide here whether the term “heuristic,” by itself, connotes sufficient structure to maintain the presumption against means-plus-function claiming because, in this case, the claims do not nakedly recite heuristics without further description in the remaining claim language and specification. To the contrary, the claim language and specification disclose the heuristics’ operation within the context of the invention, including the inputs, outputs, and how certain outputs are achieved. In all cases, the claimed input is a finger contact. The specification explains that the finger contacts may be taps, swipes, double taps, or finger rolling, and may involve one or two fingers contacting the screen at different initial angles. See, e.g., '949 Patent at col. 19, 11. 30-46; col. 65, 11. 21-24; col. 66,11. 47-51. The claims recite heuristics with varying objectives, including vertical screen scrolling, two-dimensional screen translation, moving to the next item in a list, and translating content within a frame. The claims also explain that the invention differentiates between vertical scrolling and two-dimensional translation based upon the angle of initial movement of the finger contact. The written description provides further details regarding the heuristics’ inputs and outputs. Regarding one-dimensional vertical screen scrolling, the specification explains that “in response to an upward swipe gesture 3937 by the user that is within a predetermined angle (e.g., 27°) of being perfectly vertical, the web page may scroll one-dimensionally upward in the vertical direction.” '949 Patent at col. 64, 11. 21-25. Regarding two-dimensional translation, the specification discloses that “in response to an upward swipe gesture 3939 (FIG.39C) by the user that is not within a predetermined angle (e.g., 27°) of being perfectly vertical, the web page may scroll two-dimensionally along the direction of the swipe.” Id. at col. 64, 11. 30-34. The specification defines two-dimensional movement as “simultaneous movement in both the vertical and horizontal directions.” Id. The specification explains how a user can move to the next item in a list via a finger tap gesture on the right side of the screen, a right-to-left finger swipe, or by tapping a next image icon. Id. at col. 30, 11. 42-67. The specification also discusses the structure behind translating “content within a frame rather than translating the entire page that includes the frame.” Id. at col. 123, 11. 6-8. For performing this function, the specification describes an “M-finger translation gesture 4214,” where M is a number different from the number of fingers used to translate the entire page. Id. at col. 75, 11. 18-26. The specification also explains that the direction of translation may be the direction of the “movement of the M-finger translation gesture.” Id. at col. 75, 11. 34-35. Alternatively, the direction of translation may be determined by the angle of the movement of the M-finger gesture, according to a particular rule, i.e. a specific, identifying heuristic. Id. at col. 75,11. 39-44. The figures in the '949 patent provide further structural details. Figs. 12A, 39C, 42A, 42B, and 42C illustrate the finger contacts described in the specification that result in vertical scrolling (3937), two-dimensional translation (3939), turning to the next item (1218, 1220, and 1212), or translating within a frame (4214). In sum, the claims at issue differ from those that simply recite a generic means or mechanism, without further description in the remaining claim language or the specification. Instead, the claim language and specification outline the rules that the heuristics follow, based upon, for example, the initial angle of a finger contact, the number of fingers making contact, the direction of movement of a finger contact, a specific swiping gesture, taping a certain location on the screen, or the angle of movement of a finger on the screen. See Welker Bearing, 550 F.3d at 1096-97; Inventio, 649 F.3d at 1359 (“This is not a case where a claim nakedly recites a ‘device’ and the written description fails to place clear structural limitations on the ‘device.’ ”). Thus, the '949 patent recites a claim term with a known meaning and also describes its operation, including its input, output, and how its output may be achieved. Accordingly, the heuristic claim limitations provide “sufficiently definite structure,” to a person of ordinary skill in the art, for performing the recited function, and Motorola has not rebutted the strong presumption against means-plus-function claiming. We reverse the district court’s construction that the “heuristic” claim limitations were drafted in means-plus-function format and vacate its summary judgment of non-infringement. Apple’s '647 Patent Regarding Apple’s '647 patent, the parties dispute the meaning of the claim terms “analyzer server” and “linking actions to the detected structures.” The district court construed “analyzer server” as “a server routine separate from a client that receives data having structures from the client” and “linking actions to the detected structures” as “creating a specified connection between each detected structure and at least one computer subroutine that causes the CPU to perform a sequence of operations on that detected structure.” Apple argues that both constructions are erroneous. We disagree with Apple and affirm the district court’s claim construction. The '647 patent discloses a system for recognizing certain structures (such as a telephone number) on a touchscreen and then linking certain actions (such as calling the telephone number) to the structure. For example, a user may be able to call or save a phone number it has received via text message or email simply by touching the number on the screen of its device. Claim 1 of the '647 patent, with relevant claim limitations emphasized, recites: A computer-based system for detecting structures in data and performing actions on detected structures, comprising: an input device for receiving data; an output device for presenting the data; a memory storing information including program routines including an analyzer server for detecting structures in the data, and for linking actions to the detected structures; a user interface enabling the selection of a detected structure and a linked action; and an action processor for performing the selected action linked to the selected structure; and a processing unit coupled to the input device, the output device, and the memory for controlling the execution of the program routines. '647 patent at col. 7, 11. 9-24 (emphasis added). The district court agreed with Motorola that “analyzer server” should be construed as “a server routine separate from a client that receives data having structures from the client.” Apple argues that the analyzer server need not be “separate from a client.” Instead, Apple argues that “analyzer server” should be construed as “a program routine(s) that receives data, uses patterns to detect structures in the data, and links actions to the detected structures.” We agree with the district court’s construction of “analyzer server.” As the district court recognized, the plain meaning of “server,” when viewed from the perspective of a person of ordinary skill in the art, entails a client-server relationship. Consistent with this perspective, the specification discloses an analyzer server that is separate from the application it serves. The analyzer server is part of the “program 165 of the present invention.” '647 patent at col. 3,11. 38-39. Fig. 1 shows the program 165 and the application 167 as separate parts of a random-access memory (RAM): Id. at Fig. 1. Further, the specification states that “the program 165 of the present invention is stored in RAM 170 and causes CPU 120 to identify structures in data presented by the application 167.” Id. at col. 3,11. 37-41. Thus, the specification describes the analyzer server and the application, which it serves, as separate structures. Apple does not point to evidence suggesting a different ordinary meaning, nor do we discern such evidence in the record before this court. Indeed, Apple’s proposed construction contradicts the claim language because it reads “analyzer server” out of the claim. The claim recites “routines including an analyzer server for detecting structures in the data, and for linking actions to the detected structures.” Apple’s proposed construction recites program routines that detect structures and links actions to the detected structures, without any mention of “analyzer servers.” Apple’s construction essentially takes the claim text and removes the “analyzer server,” leaving the rest basically unchanged. Thus, Apple’s construction conflicts with the claim language by ignoring the claim term “server.” See, e.g., Merck & Co. v. Teva Pharms. USA, Inc., 395 F.3d 1364, 1372 (Fed.Cir.2005) (“A claim construction that gives meaning to all the terms of the claim is preferred over one that does not do so.”); Pause Tech., LLC v. TiVo, Inc., 419 F.3d 1326, 1334 (Fed.Cir.2005) (“In construing claims, however, we must give each claim term the respect that it is due.”); Strattec Sec. Corp. v. Gen. Auto. Specialty Co., 126 F.3d 1411, 1417 (Fed.Cir.1997) (holding that it was legal error for the district court to instruct the jury that the claim term “sheet” was not properly considered part of the claim); Exxon Chem. Patents, Inc. v. Lubrizol Corp., 64 F.3d 1553, 1557 (Fed. Cir.1995) (“We must give meaning to all the words in Exxon’s claims.”). By contrast, the district court’s construction comports with the ordinary meaning of “server” and is supported by the specification. Accordingly, we affirm the district court’s construction of “analyzer server.” The district court also agreed with Motorola that “linking actions to the detected structures” should be construed as “creating a specified connection between each detected structure and at least one computer subroutine that causes the CPU to perform a sequence of operations on that detected structure.” Apple argues that the district court’s construction is erroneous for two reasons. First, the district court incorrectly added the “specified connection” limitation. Second, the claims require linking multiple actions to each structure, rather than “at least one.” Apple contends that the correct construction is “associating detected structures to computer subroutines that cause the CPU to perform a sequence of operations on the particular structure to which they are associated.” We agree with the district court. Apple argues that the claims require only “associating” between the structure and the subroutines but ignores that the claims recite “linking.” From a general sense, the plain meaning of associating relates to a mere commonality, while linking infers a joining. Additionally, the specification here demonstrates that linking is more than just associating. The patent consistently differentiates between associating and linking and implies that linking is a more specific connection than merely associating. For example, the specification explains that actions are “associated” with specific “grammars” or “patterns,” and that “linking” occurs only after these grammars or patterns are “detected.” See, e.g., '647 patent at col. 5, 11. 59-61 (“upon detection of a structure based on a particular pattern, actions associated with the particular pattern are linked 825 to the detected structure”); col. 7, 11. 38-39 (“wherein the analyzer server links to a detected structure the actions associated with the grammar”); col. 3,11. 65-67 (“analyzer server 220 links actions associated with the responsible pattern to the detected structure, using conventional pointers”); col. 5, 11. 31-33 (“analyzer server 220 links the actions associated with grammars 410 and strings 420 to these identified structures”) (emphases added). Apple argues that requiring a “specified connection” limits the claims to the use of the “pointers” described in the specification. The district court explained that a pointer is “a term of art in computer engineering” that “stores a computer memory address.” The specification explains that pointers may be used to link the associated actions to the detected structures. '647 patent at col. 3, 11. 65-67 (“upon detection of a structure, analyzer server links actions associated with the responsible pattern to the detected structure, using conventional pointers”); col. 4, 1. 64 — col. 5, 1. 5 (“[U]pon identification of a structure in the text, parser links the actions associated with the grammar to the identified structure. More particularly, parser retrieves from grammar file pointers attached to the grammar and attaches the same pointers to the identified structure. These pointers direct the system to the associated actions contained in associated actions file. Thus, upon selection of the identified structure, user interface can locate the linked actions.”). Although the district court stated that the specification “makes clear that linking is accomplished through pointers,” it did not, as Apple argues, actually limit the claims to “pointers.” Rather, the court interpreted linking to require a “specified connection,” not just a connection established with the use of pointers. The specification explains that linking may be accomplished through the use of pointers but does not require their use and neither did the district court. Thus, the district court’s construction comports with the specification, including the repeated differentiation between linking and associating and the pointers embodiment described therein. Apple is also incorrect that the claims require each structure to be linked with multiple actions. Apple points to the claim’s recitation of the plural “actions.” See '647 patent at col. 7, 11. 17-18 (“an analyzer server for detecting structures in the data, and for linking actions to the detected structures”) (emphasis added). The plain language of the claims does not require multiple actions for each structure because the claim recites linking multiple actions to multiple structures. As such, the plural “actions” may be reasonably read as at least one action per structure. In fact, Fig. 4 displays an example of the invention with only one action linked to a specific structure. '647 patent at Fig. 4. In Fig. 4, the “date grammar” structure only has one corresponding action, “put in electronic calendar.” This directly contradicts Apple’s proposal to require the claims to link multiple actions to each structure. '647 patent at Fig. 4. Accordingly, we affirm the district court’s construction of “linking actions to the detected structures.” Apple’s '263 Patent The '263 patent discloses a system for processing data in “realtime.” The parties dispute whether the “realtime application program interface (API)” in claim 1 must itself function in realtime or whether it must just facilitate realtime processing by other subsystems. The district court concluded that the API need just facilitate realtime processing and construed “realtime API” as an “API that allows realtime interaction between two or more subsystems.” Motorola argues that this construction reads “realtime” out of the claim. We disagree and affirm the district court’s construction. Claim 1, with the relevant limitation emphasized, recites: A signal processing system for providing a plurality of realtime services to and from a number of independent client applications and devices, said system comprising: a subsystem comprising a host central processing unit (CPU) operating in accordance with at least one application program and a device handler program, said subsystem further comprising an adapter subsystem interoperating with said host CPU and said device; a realtime signal processing subsystem for performing a plurality of data transforms comprising a plurality of realtime signal processing operations; and at least one realtime application program interface (API) coupled between the subsystem and the realtime signal processing subsystem to allow the subsystem to interoperate with said real-time services. '263 patent at col. 11, 11. 28-43 (emphasis added). The district court noted that, generally, to be realtime, a system “must satisfy explicitly (bounded) response-time constraints or risk severe consequences,” such as degraded performance. Motorola contends that the district court’s construction reads “realtime” out of the claim because it does not require the API itself to function in realtime. Motorola points to independent claim 31, which recites an API without the “realtime” qualifier, and argues that, by including real-time in claim 1, the patentee intended that the API itself operate in realtime. See Phillips, 415 F.3d at 1314 (finding that use of the word “steel” in the term “steel baffles” “strongly implies” a difference between steel baffles and non-steel baffles). We agree with the district court. The specification describes the API as an interface that sends commands and parameters to the “real-time engine,” which actually performs the realtime data processing. The specification does not describe the API itself as meeting any specific response-time constraints or otherwise needing realtime functionality. Instead, the API’s role is to send commands and parameters to the real-time engine. See, e.g., '263 patent at col. 6, 11. 33-38 (“each interface receives commands from an application program, through the handler 44, and instructs the real-time engine to carry out the necessary transforms”); col. 5, 11. 22-25 (“the particular transforms to be performed are sent as commands to the real-time engine from the adapter handler 44 via suitable application programming interfaces 48”); col. 10, 11. 40-44 (“in response thereto, the API 48 which receives these commands supplies the real-time engine with the appropriate parameters for performing the transforms in the required format”). By contrast, the real-time engine is described as performing the actual processing, such as “text-to-speech conversion” or “video processing.” '263 patent at abst. (“a data transmission system having a real-time engine for processing isochro-nous streams of data”); col. 10, 11. 16-18 (“the actual modulation and demodulation of the hardware interface adapter’s isoch-ronous PCM data stream is accomplished entirely by the real-time engine”); col. 9, 11. 65-67 (“the handler has no involvement with the isochronous data stream created by the real-time engine”). Thus, although the API interacts with the real-time engine, it is the latter that actually performs the time-constrained processing. Contrary to Motorola’s argument, the district court did not read “realtime” out of the claim. The API is an “interface.” As such, it communicates and interacts with other subsystems that process data in real-time without necessarily processing any data itself. This is what the specification describes and what the district court correctly understood. Further, the district court’s construction does not prevent the API from functioning in realtime, it just does not require the API to function in realtime. This is consistent with the claims and written description, which only require the API to facilitate the functionality of the real-time engine. We affirm the district court’s claim construction. Motorola’s '559 Patent Turning to Motorola’s asserted patents, the '559 patent discloses a method for generating “preamble sequences,” which are used in communications between cell phones and base stations. The district court construed claim 5 of the '559 patent to require that the third step (“multiplying the outer code by the inner code”) take place only after the first two steps (“forming an outer code” and “forming an inner code”) are completed. Based upon this construction, the court granted summary judgment of non-infringement to Apple. Motorola appeals only the claim construction decision. We affirm. Claim 5, with relevant limitations emphasized, recites: A method for generating preamble sequences in a CDMA system, the method comprising the steps of: forming an outer code in a mobile station; forming an inner code in the mobile station utilizing the following equation: where Sj; j=0, 1, ..., M-l are a set of orthogonal codewords of length P, where M and P are positive integers; and multiplying the outer code by the inner code to generate a preamble sequence. '559 patent at col. 5, 11. 20-35 (emphasis added). Steps in a method claim need not necessarily be performed in the order they are written. Altiris, Inc. v. Symantec Corp., 318 F.3d 1363, 1369 (Fed.Cir.2003). On the other hand, if grammar, logic, the specification, or the prosecution history require the steps to be performed sequentially, then the claims are so limited. Id.; Loral Fairchild Corp. v. Sony Corp., 181 F.3d 1313, 1322 (Fed.Cir.1999) (“Although not every process claim is limited to the performance of its steps in the order written, the language of the claim, the specification and the prosecution history support a limiting construction in this case.”). The district court correctly noted that, while in a preferred embodiment the inner and outer codes are formed before the multiplication step begins, this alone does not limit the claims. See, e.g., '559 patent at Fig. 4. The district court also acknowledged that the invention would likely function even if the multiplication step began before the full inner and outer codes were formed. Notwithstanding the foregoing, the court construed the claims to require that step three occur only after steps one and two are completed. We agree with the district court. The claims recite multiplying “the” inner code with “the” outer code to create a preamble “sequence.” Both the inner code and outer code are sequences of numbers. The plain meaning of multiplying “the” codes together is that the entire sequences are multiplied together after they have been formed. If claim 17 was directed to a method that, multiplied only parts of the inner and outer code together, it would not recite multiplying “the” codes together to form the preamble sequence. The more natural reading of the claim language supports the district court’s finding that the inner and outer codes must be fully formed before they are multiplied together. The specification supports this reading. When discussing multiplying the inner and outer codes, the specification describes forming an inner and outer code and then multiplying the codes together: The present invention provides a method for generating preamble sequences in a CDMA communication system. The method comprises forming an outer code and an inner code at a mobile station. The mobile station then multiplies the outer code by the inner code to generate a preamble sequence. '559 patent at col. 2, 11. 52-57 (emphasis added). Because it is supported by the plain meaning of the claim language and the specification, we affirm the district court’s claim construction. Because this construction was the basis for the district court’s grant of summary judgment of noninfringement, we need not reach the court’s construction of “a set of orthogonal codewords.” Accordingly, we affirm the district court’s grant of Apple’s motion for summary judgment of non-infringement of the '559 patent based upon its construction that the steps of claim 17 be performed in the sequence described above. Motorola’s '712 patent The '712 patent discloses a system for encrypting data communications. The district court found that the claimed “transmit overflow sequence number,” or TOSN, is “never transmitted” to the receiver in the claimed system. Because the counterpart to the TOSN in the accused products is transmitted to the receiver, the district court granted Apple’s motion for summary judgment of non-infringement. We agree with the district court’s construction and therefore affirm its grant of summary judgment. Claim 17 of the '712 patent, with relevant limitations emphasized, recites: In a communication system having a physical layer, data link layer, and a network layer, a method for providing cryptographic protection of a data stream, comprising: (a) assigning a packet sequence number to a packet derived from a data stream received from the network layer; (b) updating a transmit overflow sequence number as a function of the packet sequence number; and (c) encrypting, prior to communicating the packet and the packet sequence number on the physical layer, the packet as a function of the packet sequence number and the transmit overflow sequence number. '712 patent at col. 8,1. 65-col. 9,1. 12. The specification explains that, in order to encrypt and decrypt the data being transmitted, the invention assigns each “packet” of data both a “packet sequence number” and an overflow number. See id. at abst; col. 5,11. 13-15. The “key” used to encrypt and decrypt the data includes both the packet sequence number and the overflow number. The packet sequence numbers are assigned sequentially up to a maximum. Once the maximum is reached, the packet sequence number count “rolls over” and begins again from number one. For example, if the maximum is 128, after 128 is assigned to a packet of data, the next packet is assigned a packet sequence number of one. Id. at col. 4,11. 15-17. The overflow counter counts the number of times the packet sequence number rolls over. Id. at col. 3, 11. 54-55, 65-68. Thus, as demonstrated in the above illustration, each time the packet sequence number rolls over, the overflow sequence number increases by one. Id. On the transmission side, the overflow number is called the TOSN. On the receiving side, it is called the receiving overflow sequence number (“ROSN”). In this manner, the key to encrypt and decrypt the data includes both a packet sequence number and an overflow number. When the packets of data are transmitted, the claimed system also transmits the packet sequence number. Id. at col. 5, 11. 29-32; col. 9, 11. 7-9. The specification does not describe transmitting the TOSN or any other overflow number. Id. Instead, the receiver couples the packet sequence number it receives with a ROSN it generates and uses these two numbers to decrypt the data packet. The specification supports the district court’s construction that the TOSN is never transmitted. First, the claims recite a TOSN, not a generic overflow sequence number. The specification plainly describes transmitting certain aspects of the system (the data packet, the packet sequence number) but never suggests that the TOSN is part of that transmission. Nor is it necessary for the TOSN to be transmitted. If it were, there would be no need for the receiver to generate a ROSN. Further, if both the TOSN and packet sequence number were transmitted, an eavesdropper could intercept the entire key needed to decrypt the message. This would defeat the invention’s purpose of providing increased security by not transmitting the entire key. Statements made by Motorola during prosecution of a related Japanese patent further support this construction. Before the Japan Patent Office, Motorola distinguished the prior art by explaining, with the included emphasis, that “the overflow sequence number is never transmitted.” Motorola further explained that, because the TOSN is never transmitted, “there is no chance to intercept the overflow sequence number; thus, [the invention] provides a higher level of security.” Motorola made this argument on more than one occasion. The Japanese application claims priority to the PCT application that issued in the United States as the '712 patent. Both the Japanese application and the '712 patent have the same specification, and the Japanese application included an identical claim to claim 17 at the time of Motorola’s statements. Our precedent does not precisely address the impact of statements such as Motorola’s here. Motorola’s description of the TOSN came after the '712 patent issued and was made in front of a foreign patent office. This court has previously found that statements made in related, later-prosecuted U.S. patents may inform the meaning of earlier issued claims. See, e.g., Microsoft Corp. v. Multi-Tech Sys., Inc., 357 F.3d 1340, 1350 (Fed.Cir.2004). In Microsoft, the court noted that “any statement of the patentee in the prosecution of a related application as to the scope of the invention would be relevant to claim construction, and the relevance of the statement made in this instance is enhanced by the fact that it was made in an official proceeding in which the patentee had every incentive to exercise care in characterizing the scope of its invention.” Id. (emphasis added). Based upon this reasoning, the Microsoft court concluded that the patentee’s statements made during the prosecution of a later patent were relevant to an earlier issued patent that shared a common specification. Id. Of course, statements made in unrelated applications are not relevant to claim construction. See, e.g., Goldenberg v. Cytogen, Inc., 373 F.3d 1158, 1167 (Fed.Cir.2004) (explaining that an unrelated patent or application is one that does not “have a familial relationship” with the patent at issue); Abbott Labs. v. Dey, L.P., 287 F.3d 1097, 1105 (Fed.Cir.2002) (finding applications unrelated when they had “no formal relationship and were presented to the patent office as patentably distinct inventions”). This court has also considered statements made before a foreign patent office when construing claims if they are relevant and not related to unique aspects of foreign patent law. See, e.g., Abbott Labs. v. Sandoz, Inc., 566 F.3d 1282, 1290 (Fed.Cir.2009) (“While statements made during prosecution of a foreign counterpart to a U.S. patent application have a narrow application to U.S. claim construction ... in this case the JP '199 application is part of the prosecution history of the '507 patent itself’) (internal citations omitted); Gillette Co. v. Energizer Holdings, Inc., 405 F.3d 1367, 1374 (Fed.Cir.2005) (considering the patentee’s arguments before the European Patent Office (EPO) and concluding that a “blatant admission by this same defendant before the EPO clearly supported]” the court’s construction); Tanabe Seiyaku Co., v. U.S. Int’l Trade Comm’n, 109 F.3d 726, 733 (Fed.Cir.1997); Caterpillar Tractor Co. v. Berco, S.p.A., 714 F.2d 1110, 1116 (Fed.Cir.1983). But see ALA Eng’g Ltd. v. Magotteaux Int’l S/A, 657 F.3d 1264, 1279 (Fed.Cir.2011) (“[O]ur precedent cautions against indiscriminate reliance on the prosecution of corresponding foreign applications in the claim construction analysis.”); Pfizer, Inc. v. Ranbaxy Labs. Ltd., 457 F.3d 1284, 1290 (Fed.Cir.2006) (“[Statements made during prosecution of foreign counterparts to the '893 patent are irrelevant to claim construction because they were made in response to patentability requirements unique to Danish and European law.”). The principles illustrated in these decisions provide ample support for holding Motorola to the statements made during Japanese prosecution. Motorola’s statements that the TOSN “is never transmitted” to the receiver could not be clearer. See Gillette, 405 F.3d at 1374 (holding party to “blatant admission” in argument made to EPO). Motorola also explained, more than once, that not transmitting the TOSN improves the security of the claimed system. This explanation is consistent with the claims and the invention described in the specification. See AIA, 657 F.3d at 1279 (noting concerns regarding differences in foreign patent law). Thus, as in Microsoft, the statements were “made in an official proceeding in which the patentee had every incentive to exercise care in characterizing the scope of its invention.” 357 F.3d at 1350. Further, the two patents are related and share a familial relationship. They both claim priority to the same PCT application. Both specifications are the same. And, at the time Motorola made its statements to the Japanese patent office, the Japanese application contained a claim identical to claim 17. Thus, the construction supported by the specification is also supported by Motorola’s statements before the Japan patent office. We hold that it was not error for the district court to rely upon Motorola’s statements to the Japan Patent Office. As such, we affirm the district court’s construction. The district court granted Apple’s motion for summary judgment of non-infringement based upon its construction of claim 17. On appeal, Motorola does not dispute that Apple does not infringe under this construction. Accordingly, we affirm the district court’s grant of summary judgment of non-infringement of the '712 patent. Admissibility Of Damages ExpeRT Evidence In preparation for trial on the patent infringement claims still at issue, the parties presented expert testimony supporting their damages calculations. Both parties challenged the admissibility of the other side’s expert evidence. The district court held a Daubert hearing at which the experts testified and the parties presented oral arguments regarding admissibility. Subsequently, the district court excluded nearly all of both parties’ expert evidence relating to damages for infringement. Not all of these decisions are before us. On appeal, Apple argues that the proposed testimony of its damages expert regarding the '949 and '263 patents is admissible. Motorola argues that the proposed testimony of its damages expert for the '898 patent is admissible. Because the district court erred in its decision to exclude the parties’ expert evidence, we reverse. The legal framework for admission of expert testimony is provided by the Federal Rules of Evidence, along with Daubert v. Merrell Dow Pharm., Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993) and its progeny. See Fed.R.Evid. 702, 703. In Daubert, the Court addressed the proper standard for admitting expert testimony and emphasized that the focus “must be solely on principles and methodology, not on the conclusions that they generate.” 509 U.S. at 595, 113 S.Ct. 2786. In Kumho Tire Co. v. Carmichael, 526 U.S. 137, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999), the Court clarified that the district court’s “ga-tekeeping obligation” applied to all expert testimony. 526 U.S. at 147, 119 S.Ct. 1167. Subsequently, “Rule 702 was amended in response to Daubert and cases applying it, including Kumho Tire.” Micro Chem., Inc. v. Lextron, Inc., 317 F.3d 1387, 1391 (Fed.Cir.2003). Rule 702 states: A witness who is qualified as an expert by knowledge, skill, experience, training, or education may testi