Academia Sinica v. Microsoft Corp

I. Executive Summary and Procedural Information

• Parties & Counsel:
  • Plaintiff: Academia Sinica (Taiwan)
  • Defendant: Microsoft Corporation (Washington)
  • Plaintiff's Counsel: Groombridge, Wu, Baughman & Stone LLP
• Case Identification: 2:26-cv-00397, E.D. Tex., 05/14/2026
• Venue Allegations: Plaintiff alleges venue is proper in the Eastern District of Texas because Defendant maintains a regular and established place of business in the district, specifically through its operation of "Microsoft Windows Stores" within Best Buy locations, and has committed acts of infringement there.
• Core Dispute: Plaintiff alleges that Defendant's code-completion features (e.g., IntelliSense, IntelliCode) in its software development products, such as Visual Studio, infringe two patents related to methods for generating typing candidates from user input.
• Technical Context: The technology at issue concerns methods for improving text-entry efficiency by suggesting word or code completions based on abbreviated user inputs, including non-contiguous character subsequences, rather than just initial prefixes.
• Key Procedural History: The complaint alleges that Plaintiff notified Defendant of the asserted patents in October 2023. It also alleges that one of Defendant's own patents (U.S. Pat. No. 9,009,591) cited the Taiwan counterpart to the '468 Patent, which may be relevant to willfulness. The '999 Patent is a continuation of the application that led to the '468 Patent. The complaint also notes that Defendant has previously consented to personal jurisdiction in the Eastern District of Texas in other patent cases.

Case Timeline

Date	Event
2006-09-27	Priority Date for '468 Patent and '999 Patent
2013-01-29	'468 Patent Issued
2013-11-26	'999 Patent Issued
2023-10-XX	Plaintiff allegedly notified Defendant of infringement
2026-05-14	Complaint Filed

II. Technology and Patent(s)-in-Suit Analysis

U.S. Patent No. 8,364,468

• Patent Identification: U.S. Patent No. 8,364,468, titled "Typing Candidate Generating Method for Enhancing Typing Efficiency," issued on January 29, 2013. Compl. ¶25

The Invention Explained

• Problem Addressed: The patent's background section states that typing efficiency on keyboards is hindered by factors such as incorrect spellings, typing errors, and the difficulty of inputting characters with diacritical (accent) marks. Compl. ¶29 '468 Patent, col. 1:15-27
• The Patented Solution: The invention proposes a method to generate candidate words for text completion based on a user-inputted "subsequence" of characters. Unlike traditional autocomplete that relies on matching the beginning of a word, this method allows the subsequence to be a non-contiguous set of characters or to match a string found anywhere within the target word. The system searches a text storage (such as a dictionary or text previously typed by the user) for potential matches and presents them to the user. Compl. ¶¶30-31 '468 Patent, abstract '468 Patent, col. 6:58-62
• Technical Importance: The technology provided a more flexible and powerful alternative to prefix-based text completion, improving input speed and efficiency, particularly for long or complex words. Compl. ¶34

Key Claims at a Glance

• The complaint asserts at least independent Claim 1. Compl. ¶43
• The essential elements of independent Claim 1 include:
  • maintaining a text storage representing words and word sequences accepted from the user;
  • accepting a text input from the user that is a subsequence of less than all the letters of a desired word sequence;
  • searching the text storage for candidate words where the input sequence is a subsequence, without requiring the input to be an initial portion of the candidate;
  • determining the desired word sequence from the candidates; and
  • performing the search without requiring prior specification of the candidate as a possible result. Compl. ¶44 '468 Patent, col. 7:52-col. 9:4
• The complaint does not explicitly reserve the right to assert dependent claims, but this is standard litigation practice.

U.S. Patent No. 8,594,999

• Patent Identification: U.S. Patent No. 8,594,999, titled "Typing Candidate Generating Method for Enhancing Typing Efficiency," issued on November 26, 2013. Compl. ¶26

The Invention Explained

• Problem Addressed: As a continuation of the '468 Patent application, the '999 Patent addresses the same fundamental problem of overcoming hindrances to typing efficiency, such as incorrect spellings and typing errors. Compl. ¶29 '999 Patent, col. 1:21-32
• The Patented Solution: The '999 Patent refines the inventive concept by framing it in terms of "lexical items." The method involves accepting "further input" as a sequence of "input lexical items" that represents a subsequence of the lexical items making up a desired word sequence. The system then searches a text storage of previously accepted words to find candidates containing that subsequence, without needing prior specification of what the candidates might be. Compl. ¶¶30-31 '999 Patent, abstract '999 Patent, col. 8:61-col. 9:20
• Technical Importance: This approach formalizes the subsequence matching concept using "lexical items," making it adaptable to contexts beyond natural language, such as programming languages or character-based input methods for Asian languages. Compl. ¶32 '999 Patent, col. 8:48-54

Key Claims at a Glance

• The complaint asserts at least independent Claim 1. Compl. ¶56
• The essential elements of independent Claim 1 include:
  • maintaining a text storage of previously accepted words and word sequences;
  • accepting further input as a sequence of "input lexical items" that is a subsequence of the lexical items representing a desired word sequence;
  • searching the text storage (without prior specification of candidates) to locate candidates where the input lexical item sequence is a subsequence; and
  • determining the desired word sequence from the located candidates. Compl. ¶57 '999 Patent, col. 8:61-col. 9:20
• The complaint does not explicitly reserve the right to assert dependent claims.

III. The Accused Instrumentality

Product Identification

The accused instrumentalities are Microsoft's software products and services that include code-completion, suggestion, or selection functionality, such as "IntelliSense," "IntelliCode," and "whole-line completions." Compl. ¶2 The complaint focuses on Microsoft's Visual Studio and Visual Studio Code integrated development environments (IDEs). Compl. ¶2 Compl. ¶35

Functionality and Market Context

• The accused products are software development tools that assist programmers in writing code. Compl. ¶35 The accused "IntelliSense" and "IntelliCode" features automatically suggest and complete text as a developer types. Compl. ¶37 The complaint alleges this functionality is based on subsequence-based searches. For example, it alleges a user can type the non-initial letters "pc" to retrieve the candidate "MyParentClass." Compl. ¶37 This screenshot from the complaint's exhibits illustrates the accused functionality of retrieving candidates based on a non-initial subsequence input. Compl. Ex. C, p. 9
• The complaint alleges that Visual Studio is used extensively in the software industry and holds the largest market share for IDEs in the United States. Compl. ¶35

IV. Analysis of Infringement Allegations

'468 Patent Infringement Allegations

Claim Element (from Independent Claim 1)	Alleged Infringing Functionality	Complaint Citation	Patent Citation
maintaining a text storage representing words and word sequences accepted as sentence text input from the user;	Microsoft's Visual Studio allegedly maintains a text storage by, for example, keeping track of parameters the user types and allocating memory for the IntelliSense process to parse and analyze code in real-time.	¶44; Ex. C, pp. 5-6	col. 7:53-55
accepting a text input from the user...the text input consisting of an input sequence of letters that represents a subsequence of less than all the letters from a desired sequence...	Visual Studio allegedly accepts a text input, such as "pc," which represents a subsequence of the desired word sequence "MyParentClass."	¶44; Ex. C, pp. 7, 9	col. 7:56-62
searching the text storage for one or more candidates in which the input sequence represents a subsequence...without requiring that the subsequence form an initial portion of the candidate; and	Visual Studio's IntelliSense allegedly uses "fuzzy matching" to search for and identify candidates, like "MyParentClass," even when the input "pc" does not form the initial portion of the candidate.	¶44; Ex. C, pp. 10-12	col. 8:1-6
determining the desired sequence of one or more words from the candidates;	Visual Studio allegedly determines the desired sequence by allowing the user to select a candidate from a list or by automatically completing the word when the user presses a key like Tab.	¶44; Ex. C, pp. 13, 15	col. 8:7-8
wherein searching the text storage is performed without requiring prior specification of the candidate as a possible result of the searching.	Visual Studio allegedly performs the search without prior specification of candidates by automatically parsing the user's code in real-time to generate a list of potential completions as the user types.	¶44; Ex. C, pp. 16-17	col. 9:2-4

• Identified Points of Contention:
  • Scope Question: A potential issue is whether "sentence text input," as used in the patent, can be interpreted to cover computer source code. The patent's discussion of URLs and email addresses may support a broader construction beyond natural language sentences. '468 Patent, col. 8:39-45
  • Technical Question: The analysis may focus on whether the "text storage" maintained by the accused products is functionally equivalent to the storage described in the patent, which includes capabilities for training and storing user-specific text. '468 Patent, col. 3:7-11 '468 Patent, col. 6:33-35

'999 Patent Infringement Allegations

This screenshot from the complaint's exhibits shows the accused functionality of retrieving a candidate ("third_friend_member") based on a non-contiguous subsequence input ("frimem"). Compl. Ex. D, p. 10

Claim Element (from Independent Claim 1)	Alleged Infringing Functionality	Complaint Citation	Patent Citation
maintaining a text storage representing words and word sequences previously accepted as sentence text input via the entry interface from the user;	Microsoft's Visual Studio allegedly maintains a text storage of previously accepted code by keeping track of parameters, methods, and other elements the user has typed.	¶57; Ex. D, pp. 5-6	col. 8:62-65
accepting further input comprising a sequence of input lexical items...the sequence of input lexical items comprises a subsequence of the lexical items of the representation of the desired sequence...	Visual Studio allegedly accepts user input such as "frimem," where each letter is treated as an "input lexical item" forming a subsequence of the lexical items that represent the desired sequence "third_friend_member."	¶57; Ex. D, pp. 7, 10	col. 9:1-9
searching the text storage...without requiring prior specification of candidates...to locate one or more candidates in which the sequence of input lexical items...forms a subsequence...	Visual Studio allegedly searches its text storage of previously typed code without prior specification to locate candidates like "third_friend_member" where the input "frimem" forms a subsequence.	¶57; Ex. D, pp. 11-14	col. 9:10-17
determining the desired sequence of one or more words from the candidates located in the text storage.	Visual Studio allegedly determines the desired sequence when the user selects a suggested candidate from the completion list or when the system automatically inserts the suggested completion.	¶57; Ex. D, pp. 15, 17-18	col. 9:18-20

• Identified Points of Contention:
  • Scope Question: The central dispute will likely involve the construction of the term "lexical item." The plaintiff's case appears to depend on this term being broad enough to cover individual typed characters, while a defendant may argue it has a more specific meaning in computer science (e.g., a complete token like an identifier or keyword).
  • Technical Question: It raises the question of whether the simple act of typing characters into a code editor constitutes "accepting further input comprising a sequence of input lexical items" as required by the claim, or if this mischaracterizes the technical operation.

V. Key Claim Terms for Construction

'468 Patent: "subsequence" (Claim 1)

• The Term: "subsequence"
• Context and Importance: This term is the core of the inventive concept. Its scope-specifically whether it must be contiguous-is critical to the infringement analysis. The complaint alleges infringement based on both non-initial and non-contiguous inputs.
• Intrinsic Evidence for Interpretation:
  • Evidence for a Broader Interpretation: The claim language requires that the subsequence is not necessarily an "initial portion," which suggests flexibility. '468 Patent, col. 8:4-6 The specification further states that an input word can be "found anywhere within the at least one candidate word." '468 Patent, col. 6:60-62 Furthermore, independent claim 17 explicitly claims a "non-contiguous subsequence," which could be argued to clarify, rather than limit, the general meaning of "subsequence" in claim 1. '468 Patent, col. 10:29-31
  • Evidence for a Narrower Interpretation: Parties may argue that under the doctrine of claim differentiation, the explicit inclusion of "non-contiguous" in claim 17 implies that the term "subsequence" in independent claim 1 does not inherently include non-contiguous sequences.

'999 Patent: "lexical item" (Claim 1)

• The Term: "lexical item"
• Context and Importance: The infringement theory for the '999 Patent hinges on interpreting individual typed characters (e.g., 'p', 'c') as "lexical items." Practitioners may focus on this term because its definition will likely determine whether typing source code in an IDE falls within the claim's scope.
• Intrinsic Evidence for Interpretation:
  • Evidence for a Broader Interpretation: The patent does not provide an explicit definition, which may suggest the term should be given its plain and ordinary meaning in the context of the invention. The specification's reference to input methods for Asian languages, where a "string of alphabets forms a code that retrieves a collection of candidate Asian characters," may support a flexible interpretation where basic input units can be considered "lexical items." '999 Patent, col. 8:48-52
  • Evidence for a Narrower Interpretation: A defendant may argue that in the field of computer science and compilers, a "lexical item" or "token" is a recognized unit of meaning (e.g., a keyword, an identifier, an operator) and is distinct from a single, constituent character. The patent's own language distinguishes between "letters of an alphabet" in a related patent and "lexical items" in the '999 patent, which may suggest a deliberate change in meaning. '999 Patent, col. 9:22-23

VI. Other Allegations

• Indirect Infringement: The complaint alleges both induced and contributory infringement. The inducement allegation is based on Microsoft's marketing and provision of user guides, training materials, and technical support that instruct customers on how to use the accused code-completion features. Compl. ¶47 Compl. ¶60 The contributory infringement allegation posits that the subsequence-based code completion functionality is a material component of the patented method, specially made for this purpose, with no substantial non-infringing uses. Compl. ¶48 Compl. ¶61
• Willful Infringement: The complaint alleges willful infringement based on both pre-suit and post-suit knowledge. Pre-suit knowledge is alleged based on communications between the parties starting in October 2023 and, notably, on the fact that a Microsoft patent (U.S. Pat. No. 9,009,591) allegedly cited the Taiwanese counterpart to the asserted '468 Patent in its own prosecution history. Compl. ¶45 Compl. ¶58 Willfulness is also alleged based on knowledge gained from the filing of the complaint itself. Compl. ¶49 Compl. ¶62

VII. Analyst's Conclusion: Key Questions for the Case

1. A central issue will be one of claim construction: can the term "lexical item" from the '999 patent, which has a specific meaning in computer science, be construed broadly enough to encompass individual characters typed by a user into a code editor, as the plaintiff's infringement theory appears to require?
2. A key evidentiary question will be one of technical implementation: does Microsoft's "IntelliSense" feature function by "searching a text storage" of previously entered text for "subsequences" in the specific manner claimed by the patents, or is there a fundamental mismatch in the underlying technical operation of the accused algorithms?
3. The willfulness claim will likely turn on a question of pre-suit knowledge: did Microsoft's citation to the plaintiff's related technology in its own patent application provide actual, subjective knowledge of the asserted patents' technology, and if so, did its continued use of the accused features constitute objectively reckless behavior?