In order to see how the patent claims are worded so as to lock down the
coding, it is instructive to focus on claims 46 to 51 of the '348
patent:
46. A method for encoding a Transport Format
Combination Indicator (TFCI) in a CDMA mobile communication system,
comprising:
- inputting TFCI information bits in a 10 bit
unit;
- generating, by a codeword generator, a codeword containing
32 bits based on the TFCI information bits; and
- outputting the
generated codeword,
wherein the codeword is generated by combining
at least two basis sequences selected by the TFCI information bits from among
all basis sequences available for encoding and all the basis sequences are
mapped with the TFCI information bits in a 10 bit
unit.
47. The method of claim 46, wherein all the
basis sequences are “01010101010101010101010101010101”,
“00110011001100110011001100110011”, “00001111000011110000111100001111”,
“00000000111111110000000011111111”, “00000000000000001111111111111111”,
“11111111111111111111111111111111”, “00101000011000111111000001110111”,
“00000001110011010110110111000111”, “00001010111110010001101100101011” and
“00011100001101110010111101010001”.
48. The method of
claim 47, wherein if the TFCI information bits are less than 10 bits, 0 is added
to the TFCI information bits to represent the TFCI information bits in a 10 bit
unit.
49. The method of claim 48, wherein the at least
one basis sequence is selected by multiplying the TFCI information bits by their
corresponding basis sequences.
50. The method of claim
49, further comprising the step of puncturing two bits from the 32 bits of
codeword and outputting 30 bits of codeword.
51. The
method of claim 50, wherein the output 30 bit codeword is at least one of
“101010101010101101010101010101”, “011001100110011011001100110011”,
“000111100001111000111100001111”, “000000011111111000000011111111”,
“000000000000000111111111111111”, “111111111111111111111111111111”,
“010100001100011111000001110111”, “000000111001101110110111000111”,
“000101011111001001101100101011” and
“001110000110111010111101010001”.
I assume for the
purpose of this discussion that the claims are valid. This requires that
the claims be definite (so that it is possible to construe the claims),
that the inventor has demonstrated enablement (requiring that the claims
read on an embodiment disclosed in the specification, and that the specification
discloses sufficient detail to enable the skilled artisan to make and use the
invention), and that the claims do not read on prior art.
In
order to determine whether at least one of these patent claims has been
infringed by some method, whether implemented in hardware or software, it is
necessary to focus on the language of the claims alone. It is an error to
compare the alleged infringing method with the embodiment in the specification
and then argue that the methods are sufficiently dissimilar that no infringement
has occurred. Infringement is decided through construing the claims to
determine the scope of the claims, and then determining whether or not the
alleged infringing method falls within the scope of one of these specific claims
46 to 51.
In litigation, the parties will brief the judge on their
interpretation of the terms of the claims, a Markman hearing will be held
to discuss any issues or disputes that arise, and the judge will rule on the
claim construction. The claims must be construed as broadly as is consistent
with the wording of the claims and with the specification. This does not
require that the claims should be limited to the properties of any particular
embodiment (i.e., device or method) described within the specification.
Indeed it is improper to attempt to construe claims by limiting them to
embodiments of the invention specifically disclosed in the specification (except
in the special case of means plus function claims).
One should
start with claim 46. We have a method for encoding TFCI information bits in 10
bit units, and outputting a 32 bit codewords. It is surely appropriate here to
construe TFCI information as consisting of a list or array of bit sequences,
where each member of the list is an input word consisting of exactly 10 binary
bits, and where each such input word is converted to an output codeword
consisting of exactly 32 binary bits. This is surely a task that one might
reasonably program a digital computer to perform, and therefore the claims would
read on any hardware or software implemention of the coding scheme that
satisfies the claim limitations.
One can presume that neither Samsung
nor the PTO discovered any prior art that reads on claim 46, and, in particular,
implementing the method in accordance with all the claim limitations
listed in the wherein clause of claim 46. (The fact that
error-correcting codes are well-known in the art, and that there are
well-understood families of such codes is irrelevant for the purposes of
determining infringement under claim 46 unless there is some specific prior art
reference or combination of prior art references that teaches the limitations of
the specific claim.) In particular, any invalidating prior art must involve
combining at least two of the basis sequences, and ensuring that all the base
sequences are mapped in a 10-bit unit in accordance with the claim
construction.
Maybe some prior art will be discovered after the patent
has issued that will invalidate claim 46. If so the scope of these claims will
be determined by at least one of claims 47 to 51, unless there is prior art that
invalidates all of 46 to 51. Thus the claims will survive in some form unless
and until claim 51 has itself been invalidated.
Well, maybe claim 46
does read on prior art not known to the PTO when the patent issued. If so, the
scope of the claim will narrow to encodings using the specific 10 basis
sequences listed in claim 47.
If there is prior art for which the input
words are not padded, then the scope of the claims will narrow to claim 48, that
covers all codings of the required where input words with fewer than 10 bits are
padded with zeros.
To guard against the situation where a lawyer or
judge would seek to restrict the claims to methods that output 32-bit words
alone, claim 50 ensures that a method that subsequently discards two bits and
outputs 30-bit words also infringes.
Finally there is claim 51, which
reads on any method that satisfies the limitations of claim 50, which translates
10-bit TFCIs to 30-bit codewords. Now the 1024 possible TFCI words must map to
distinct 30-bit codewords. There are 1073741824 such codewords to choose from.
Any method within the scope of claim 50 which includes the 10 specific codewords
stated in claim 51 will infringe.
Note that, in all of this discussion,
there is no discussion of inventiveness, and no discussion as to whether
or not the invention described in the claims is a straightforward application of
basic methods described in standard textbooks on coding theory and taught
at university. It is the specifics of the claims and any prior art that
determine whether or not the claims are valid, and whether or not they are
infringed.
[ Reply to This | Parent | # ]
|