dissenting-in-part.
I agree with the majority’s conclusions with respect to validity and the affirmance of the district court’s refusal to award sanctions under 28 U.S.C. § 1927. How*1120ever, I respectfully dissent from the majority’s construction of the terms “board” and “manufactured to have.”
Claim construction normally involves consideration of the claims, the specification, and the prosecution history. Markman v. Westview Instruments, Inc., 52 F.3d 967, 979 (Fed.Cir.1995) (en banc) (citations omitted). The claim language first and foremost defines the scope of the invention. Bell Communications Research, Inc. v. Vitalink Communications Corp., 55 F.3d 615, 619-20 (Fed.Cir.1995). Thus, in construing claim terms, we begin with the language of the claim itself. Johnson Worldwide Assocs. v. Zebco Corp., 175 F.3d 985, 989, (Fed.Cir.1999). As a general rule, claim terms are given the full range of their plain, ordinary, and accustomed meaning to one of ordinary skill in the art. See id. After revising the claim language, we consider the balance of the intrinsic evidence, including the rest of the specification and the prosecution history, if in evidence. Vitronics Corp. v. Conceptronic, Inc., 90 F.3d 1576, 1582 (Fed.Cir.1996) (en banc); Markman, 52 F.3d at 979.
The majority gives heed to the general rules of construction but then proceeds to ascertain the “plain and ordinary meaning” of the term “board” from various dictionaries. (Maj. op. at 1111). It establishes a duel between dictionary definitions and then selects one of the various definitions to support its results. The majority fails to recognize that the written description and the prosecution history clearly prescribe that the decking board of the invention is derived from a wood log. Wang Labs. v. America Online, Inc., 197 F.3d 1377 (Fed.Cir.1999).
Moreover, the majority opinion is in direct conflict with our recently issued opinion in Novartis Pharmaceuticals Corp. v. Eon Labs Manufacturing, Inc., 363 F.3d 1306 (Fed.Cir.2004). In Novartis, we held that when the dictionary definition “yields a range of possible meanings consisting of two competing definitions,” id. at 1309-10, we look to the intrinsic evidence “to determine as a matter of claim interpretation which of the available relevant definitions should be applied to the claim term at issue.” Id. at 1310. We ultimately determined that, because the intrinsic evidence used the disputed term only the context of the narrower definition, that narrower definition was the meaning the term would obtain. Id. at 1310-11. Here we have two or more possible meanings of the term “board” and therefore we should look to the context in which the term is used in the intrinsic evidence to ascertain which of the relevant definitions should be applied. See id. at 1311 (“While none of the statements in the intrinsic record is an explicit disclaimer of subject matter sufficient to vary the scope of the claim from its ordinary meaning, these statements are helpful in guiding us to choose between competing dictionary definitions of a claim term.” (footnotes omitted)); see also Texas Digital Sys., Inc. v. Telegenix, Inc., 308 F.3d 1193, 1203 (Fed.Cir.2002) (“Because words often have multiple dictionary definitions, some having no relation to the claimed invention, the intrinsic record must always be consulted to identify which of the different possible dictionary meanings of the claim terms in issue is most consistent with the use of the words by the inventor.”).1
*1121On appeal, Nystrom argues that the district court erred in construing “board,” a term appearing in all claims of the '831 patent, to mean “a piece of elongated construction material made from wood cut from a log.” As the district court acknowledged, the ordinary meaning of the term “board” does not connote that it must be made of wood cut from a log. However, as the district court concluded, the written description and drawings, as well as the prosecution history of the '831 patent, clearly narrow “board” beyond its broad ordinary meaning.
First, the written description clearly limits the meaning of the term “board.” The entire background section of the written description repeatedly refers to wood flooring. '831 patent, col. 1,11.16-67. For example, the background section explains that, although construction materials and methods for exterior decks and porches “changed dramatically with the advent of chemically treated lumber,” which was more weather-resistant, “very little change has been made in the basic design of the wood building materials used in such exterior constructions.” Id. at col. 1, 11. 38-51. The background section also notes that “the process used to cut such lumber from logs can produce inferior product on the outermost boards_” Id. at col. 1,11. 65-67. The patent professes a need for an exterior decking board that has certain attributes and that can “result in better utilization of material as the boards are cut from a log.” Id. at col. 2, 11. 1-6. An object of the invention therefore is to provide a “superior product when cut from a log, reducing the amount of scrap in the outermost boards cut from a log.” Id. at col. 2, 11. 22-23. The shaped top surface results in a configuration that “enables more usable boards to be obtained from a log.” Id. at col. 2,1. 34.
Additionally, the figures confirm that the scope of the patent extends only to boards made of wood cut from logs. Figs. 1-5 are said to show that “the convex top surface 13 is curved in the same general direction as the curvature of the growth rings GR.” Id., col. 3, 11. 25-27. Figure 4 depicts the “transverse sectional view of a log, the relationship to the outer surface of the log of several outer boards to be cut from it.” Id., col. 2,11. 55-57.
Nystrom characterizes boards made from logs simply as a preferred embodiment, and not the invention itself. He therefore takes . issue with the district court’s reliance on this court’s Wang Laboratories decision. However, the current case is akin to Wang Laboratories, where this court limited “frame” of data used in online information systems to only character-based protocols, as opposed to a broader construction of “frame” encompassing bit-mapped display systems as well. See Wang Labs., 197 F.3d at 1383 (“The only embodiment described in the '669 patent specification is the character-based protocol, and the claims were correctly interpreted as limited thereto.”). In the factual context of the specification, a board made from wood cut from logs is the invention itself, rather than merely a preferred embodiment.
Moreover, the prosecution history supports the patent’s definition of a “board” as one made from wood cut from logs. In a preliminary amendment prior to the first *1122office action, the applicant represented that “the particular configuration and dimensions of the board result in a uniformly superior product and reduction in waste or rejects due to bark or other flaws along the edges of the board when it is cut from near the outer circumference of a log.” After the first office action, Nystrom stated in a September 30, 1993, amendment that “[t]he present invention represents a unique and significant advance in the art of exterior wood flooring.” Responding to the examiner’s rejection of certain claims as obvious in view of the Yoshida prior art reference, Nystrom stated that Yoshida “discloses a floor covering tile made from synthetic resin sheets,” and that Yoshida “is clearly not concerned with materials made from wood.” By this statement, Nystrom argues that the prior art disclosed resin sheets while his material was wood. He cannot now argue that his board also includes non-wood boards.
On appeal, Nystrom contends that claims 16 and 17, which claim a “wood decking board” as opposed to a board, militate in favor of interpreting these terms differently. He argues that the district court’s construction upsets the “basic legal presumption that the use of different terms in the claims connotes different meanings.” Although not raised by the majority, Nystrom also made a claim differentiation argument to distinguish the board definition. See Tandon Corp. v. U.S. Int’l Trade Comm’n, 831 F.2d 1017, 1023 (Fed.Cir.1987) (“There is presumed to be a difference in meaning and scope when different words or phrases are used in separate claims. To the extent that the absence of such difference in meaning and scope would make a claim superfluous, the doctrine of claim differentiation states the presumption that the difference between claims is significant.”). Claim differentiation only creates a presumption that each claim in a patent has a different scope; it is “not a hard and fast rule of construction.” Kraft Foods, Inc. v. Int’l Trading Co., 203 F.3d 1362, 1368 (Fed.Cir.2000) (quoting Comark Communications, Inc. v. Harris Corp., 156 F.3d 1182, 1186 (Fed.Cir.1998)). As discussed above, this presumption is overcome by the intrinsic evidence, which clearly shows that Nystrom limited the meaning of “board” to mean only those boards made from wood cut from logs. See id. (determining that “the written description and prosecution history overcome any presumption arising from the doctrine of claim differentiation”); Multiform Desiccants, Inc. v. Medzam Ltd., 133 F.3d 1473, 1480 (Fed.Cir.1998) (cautioning that “the doctrine of claim differentiation cannot broaden claims beyond their correct scope, determined in light of the specification and the prosecution history and any relevant extrinsic evidence”); Tandon Corp., 831 F.2d at 1024 (“Whether or not claims differ from each other, one can not interpret a claim to be broader than what is contained in the specification and claims as filed.”).
In sum, it is clear from the intrinsic evidence that the term “board” is limited to a wood board cut from logs.2 I would *1123therefore affirm the district court’s construction of “board” as “a piece of elongated construction material made from wood cut from a log.”
Nystrom also challenges the district court’s determination as to the claims’ disclosure that the board’s top surface is “manufactured to have” a slightly rounded or curved configuration. The district court construed this “manufactured to have” term as “a manufacturing process utilizing woodworking techniques.” Nys-trom contends that “manufactured to have” takes its ordinary meaning and should cover all manufacturing techniques.
In narrowing “manufactured to have,” the district court reasoned that a narrowed construction of “manufactured to have” comported with the court’s construction of “board” as being made from wood cut from logs. The district court further cited to the prosecution history, in particular the applicant’s September 30, 1993 amendment. The examiner had rejected claims 1, 2, 5, 6, 11, and 12 as being “anticipated by the scientific principle that a board will warp in the direction of the grain.” Nys-trom responded as follows:
the claims have been amended to specify that applicant’s boards are manufactured to have the convex curvature. This would eliminate application of the scientific principle stated by the Examiner, even for those boards that do warp so that a convex top surface results.
Interpreting this passage, the district court stated that the only type of boards that will warp in the direction of the grain are wood boards, and concluded that the manufacturing process disclosed must be a woodworking technique such as cutting or milling. As the district court noted, any construction of “manufactured to have” is defined and limited by the proper construction of “board” as one being made from wood cut from logs. Viewing the claim language, phrased as “board ... having a top surface ... manufactured to have ... a convex top surface,” in its entirety, it is clear that any board made from wood cut from logs that is manufactured to have a convex top surface must be manufactured with a woodworking technique. In this sense, construing “manufactured to have” is somewhat redundant of the construction of “board.” I would therefore affirm the district court’s construction of “manufactured to have” because this term derives meaning from and rests on the district court’s correct construction of “board.”
. The dissent sees no inconsistency between the Novartis decision and the cases cited as precedential by the majority, supra at 1112 n. 2. In fact, the Novartis decision relies on each of those cases for support. See Novartis, 363 F.3d at 1310. Moreover, while the majority appears to suggest that Novartis is not good law, the dissent posits that the Novartis decision simply adheres to the rule articulated in Newell Cos. v. Kenney Mfg. Co., 864 F.2d 757, *1121765 (Fed.Cir.1988), and cited by the majority, that prior panel decisions are binding on subsequent panels unless overturned en banc. Compare Vitronics, 90 F.3d at 1584 ("Extrinsic evidence is that evidence which is external to the patent and file history, such as expert testimony, inventor testimony, dictionaries, and technical treatises and articles.”), with Tex. Digital, 308 F.3d at 1203 ("[Cjategorizing [dictionaries] as 'extrinsic evidence’ or even a 'special form of extrinsic evidence' is misplaced and does not inform the analysis.”).
. The majority states that the dissent places "undue emphasis on the written description and prosecution history” in the proffered interpretation. The dissent, however, is only applying the interpretive framework dictated by this court's precedent. See Markman v. Westview Instruments, Inc., 52 F.3d 967, 979 (Fed.Cir.1995) (en banc), aff’d, 517 U.S. 370, 116 S.Ct. 1384, 134 L.Ed.2d 577 (1996) ("Claims must be read in view of the specification, of which they are a part. The specification contains a written description of the invention that must enable one of ordinary skill in the art to make and use the invention. For claim construction purposes, the description may act as a sort of dictionary, which explains the invention and may define terms used in the claims.”). Markman, the dissent believes, has not been overruled.