Obviousness

Analysis of Obviousness for U.S. Patent 5,243,655 under 35 U.S.C. § 103

This analysis evaluates whether the claimed invention in U.S. Patent 5,243,655 ("the '655 patent") would have been obvious to a Person Having Ordinary Skill in the Art (PHOSITA) at the time of the invention, based on the prior art references cited during prosecution.

Under 35 U.S.C. § 103, an invention is unpatentable if the differences between the claimed invention and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art. This analysis requires not only identifying individual elements of the claims in the prior art but also establishing a motivation for a PHOSITA to combine those elements to arrive at the claimed invention.

A PHOSITA in early 1992 would be an engineer or computer scientist familiar with barcode symbologies, data encoding techniques, error correction codes, and the design and operation of optical scanning systems, including the emerging field of two-dimensional barcodes and the challenges associated with hand-held laser scanners.

Analysis of Core Decoding Method (Claims 16 & 26)

Claims 16 and 26 are directed at the method and apparatus for decoding a two-dimensional bar code symbol. The key features of this method are:

1. Scanning a 2D symbol to produce a scan line of data.
2. Decoding the scan line into codeword values, including a row indicator codeword.
3. Assigning a row number to each codeword based on the row indicator and a "cluster" to which the codeword belongs.
4. Assembling, or "stitching," the codewords into a final matrix according to their assigned row numbers, even if the scan line crosses multiple rows.

Primary Reference: U.S. Patent 4,794,239 (Allais)

The Allais patent is the most relevant prior art. It establishes the foundational concept of a stacked two-dimensional bar code (Code 49) where individual rows contain encoded data. Crucially, Allais teaches the inclusion of a row number within the data of each row to identify its position within the stack. A PHOSITA starting with Allais would have a clear blueprint for a multi-row barcode symbol designed for increased data capacity.

However, a PHOSITA would also recognize a significant limitation in the Allais system when used with a hand-held rastering laser scanner, a known technology at the time. As illustrated in FIG. 1A of the '655 patent, a skewed scan line can easily cross from one row to another in the middle of a scan. While the row number encoded in each row of Allais's symbol helps identify which rows have been scanned, it does not provide a mechanism for determining, on a codeword-by-codeword basis, to which row a specific codeword belongs if it is scanned far from the explicit row number identifier. This ambiguity presents a clear problem that a PHOSITA would be motivated to solve to improve the decoding robustness and reliability of such a symbol.

The Inventive Step and the "Gap" in the Prior Art

The '655 patent solves this ambiguity problem with its novel concept of partitioning the entire set of codewords into "at least three mutually exclusive clusters." The system then assigns codewords from a different cluster to each adjacent row in a repeating sequence (e.g., Row 0 uses Cluster 0, Row 1 uses Cluster 3, Row 2 uses Cluster 6, and so on). This provides a local row discriminator for every single codeword. When a skewed scan reads a series of codewords, the decoder can identify the cluster of each codeword and, knowing the cluster-to-row assignment pattern, correctly "stitch" the data into the appropriate rows of the final data matrix.

Combination with Secondary Art and Motivation

No single piece of cited prior art discloses or suggests this "clustering" method for row discrimination.

• Chandler et al. ('936) and Priddy et al. ('354) teach away from the '655 patent's approach. Their matrix codes rely on global finder and orientation patterns that require the entire symbol to be viewed and decoded as a whole, using a coordinate system. This is fundamentally different from the line-by-line, partial-scan "stitching" method of the '655 patent, which is specifically designed to accommodate the limitations of a sweeping laser scanner.
• Stearns et al. ('525) provides a motivation to improve decoding robustness in general but does so at the character level (using candidate lists for damaged symbols). It offers no suggestion for solving the structural, multi-row ambiguity problem.

A hypothetical argument for obviousness would be that a PHOSITA, faced with the row-crossing problem in Allais, would have found it obvious to add row-identifying information to each codeword. While this is a plausible goal, the specific implementation in the '655 patent—partitioning the entire symbology into distinct, alternating sets of bar-space patterns—is a non-trivial and elegant solution. It is not a simple or obvious extension of existing techniques. There is no teaching in the cited art that would have suggested this particular path. Therefore, a key element of claims 16 and 26, the use of codeword clusters for row assignment and stitching, appears to be non-obvious over the cited prior art.

Analysis of the Overall System (Claim 1)

Claim 1 describes the complete system, including an encoder for creating and transferring the symbol to a carrier and a recognizer for reading and decoding it. The systems disclosed in Allais, Chandler, and Priddy all teach the general combination of an encoder and a decoder for 2D symbols. However, the novelty of the '655 system resides in the specific structure of the symbol it creates (PDF417 with its cluster-based architecture) and the corresponding unique decoding logic it employs. Since the core decoding method is considered non-obvious, the claim for a system that implements this non-obvious method would likewise be considered non-obvious.

Analysis of the Data Carrier (Claim 35)

Claim 35 covers a data carrier that includes both a machine-readable 2D pattern and human-readable data, with both being generated by the claimed system. The practice of printing both human-readable text and barcodes on a single medium (e.g., a shipping label) was well-established and would have been obvious. The motivation is self-evident: to allow for both human and machine interpretation of the information.

However, the claim specifies that the machine-readable pattern is the one generated by the system's encoding means. The patentability of this claim is therefore tied to the patentability of the symbol itself. An argument for obviousness would fail because it would not have been obvious to create the specific, non-obvious 2D pattern described in the '655 patent in the first place. The claim is not merely for placing any 2D barcode next to text, but for a carrier bearing the specific, novel, and non-obvious symbol of the invention. Therefore, Claim 35 would also be considered non-obvious.

Conclusion

While the prior art, particularly U.S. Patent 4,794,239 (Allais), establishes the concept of stacked 2D barcodes with row identifiers, it does not teach or suggest the key innovation of the '655 patent. The inventive step lies in the use of mutually exclusive codeword clusters assigned to adjacent rows in a repeating pattern, which enables a decoder to determine the correct row for each individual codeword in a skewed scan line. This "stitching" capability allows for robust decoding with hand-held laser scanners, a problem not adequately solved by the cited art.

There is no clear teaching, suggestion, or motivation in the combination of the provided references to lead a PHOSITA to this specific solution. The alternative methods in the art (e.g., full-frame matrix codes) teach away from the row-by-row, partial-scan decoding approach of the '655 patent. Therefore, the central claims of U.S. Patent 5,243,655 are likely not obvious under 35 U.S.C. § 103 in view of the cited prior art.