The federal Credit for Increasing Research Activities, codified in Internal Revenue Code (IRC) Section 41, provides a tax credit calculated from eligible research expenditures under statutory formulas. The credit amount reduces tax liability dollar for dollar; it does not reimburse every dollar of research spending. Since its introduction in 1981, disputes have focused on the meaning of qualified research. Tax & Accounting Software Corp. v. United States illustrates an important historical disagreement over discovery and experimentation in commercial software development. Its reasoning must be distinguished from the regulations that govern later tax years.

The Historical Context and the 1986 Legislative Pivot

Congress introduced the research credit as Section 44F through the Economic Recovery Tax Act of 1981. The original framework drew on Section 174 research and experimental expenditure concepts. During the years involved in TAASC, Section 174 permitted current deduction of qualifying research costs. That historical rule should not be confused with the domestic and foreign expenditure rules applicable after subsequent statutory amendments, discussed below.

The Tax Reform Act of 1986 narrowed the definition of qualified research by adding requirements now associated with the four-part test in Section 41(d). Those requirements distinguish qualifying technological research from excluded activities. They do not impose a universal requirement for pioneering scientific breakthroughs.

The Four-Part Test of Section 41(d)

The following table describes the four-part framework, while distinguishing the historical expenditure provision from the current statutory cross-reference. Meeting these requirements does not override separate exclusions or the rules defining eligible expenses.

The primary point of contention in the Tax & Accounting Software Corp. case—and several contemporaneous cases like United Stationers, Inc. v. United States—was the interpretation of the phrase “discovering information” within the second test and the definition of a “process of experimentation” in the fourth test.

The Fact Pattern and Technical Disputes in TAASC v. United States

The taxpayer in this pivotal case, Tax & Accounting Software Corporation (TAASC), was an Oklahoma-based Subchapter S corporation specialized in creating software for tax and accounting professionals. During the 1993 and 1994 tax years, TAASC claimed significant research credits for the development of four specific software suites: EasyACCT, Professional Tax System, EasyMICR, and EasyTEL.

Technical Innovations at Issue

The software products developed by TAASC represented significant functional integrations for the accounting industry at the time. EasyACCT was an integrated program that synchronized transaction recording with the accumulation of historical data for financial statements, allowing seamless transfer to tax software—a feature the parties agreed was unique upon its introduction. The Professional Tax System similarly broke new ground by allowing the preparation of state and federal returns from a single dataset and facilitating electronic filing, while also being optimized to run on low-memory hardware.

Despite these commercial and functional advancements, the IRS disallowed the credits. While the government conceded that the expenditures met the Section 174 test and the business component test, it argued that the research did not “discover” information in the sense intended by the 1986 amendments and that the development did not involve a true “process of experimentation”.

The District Court’s Pro-Taxpayer Ruling

In 2000, the U.S. District Court for the Northern District of Oklahoma granted summary judgment for the taxpayers. It accepted a broader interpretation of discovery and experimentation than the government advocated. The Tenth Circuit later reversed and remanded for further proceedings under its interpretation of the statute.

The Tenth Circuit’s Reversal and Its Historical Discovery Standard

The Tenth Circuit’s 2002 decision required discovery of new information separate from the product. It rejected the government’s demand that the information advance underlying scientific principles. The decision therefore differed from the Seventh Circuit’s more restrictive formulation.

The “Separate from Product” Requirement

The distinction was between a new product and newly discovered information. Product novelty alone did not establish the separate discovery requirement under the court’s historical interpretation.

It is inaccurate to describe TAASC as requiring discovery of a new principle of computer science with independent commercial value. The court expressly rejected a requirement to expand or refine scientific principles.

Experimentation and Technical Feasibility

The Tenth Circuit accepted that generally known methods could be used in experimentation. Its concern was uncertainty about whether the desired result could be achieved, rather than novelty of the research technique.

The appellate court did not conclusively classify all TAASC activities as routine debugging. It reversed summary judgment and remanded. Its historical discussion should not be substituted for the later regulatory treatment of capability, method, and design uncertainty.

Judicial Divergence and the Multi-Circuit Conflict

United Stationers and Norwest were decided in 1998, before TAASC. Eustace followed in 2002. These decisions demonstrate restrictive historical interpretations, but they did not adopt one identical discovery standard. The separate high-threshold-of-innovation rules for internal-use software should not be conflated with the general four-part test.

The following table compares the judicial interpretations of these core requirements across the different jurisdictions during the height of the Discovery Test era.

Eustace recognized the disagreement between the Seventh and Tenth Circuits but concluded that the taxpayer failed under either approach. These historical outcomes do not establish a current rule that only pioneering technologies qualify.

Regulatory Correction: Treasury Decision 9104 and the Uncertainty Test

Treasury and the IRS finalized TD 9104 in January 2004, following earlier proposals and comments. The regulatory reconsideration was already underway before the 2002 TAASC decision; it should not be portrayed as solely a response to that case.

TD 9104 rejected a requirement to advance the common knowledge of a skilled professional. It focused the discovery inquiry on resolving uncertainty in development or improvement.

The Modern Definition of “Discovering Information”

Under the regulations, discovering information concerns eliminating uncertainty about a business component. Although this overlaps with research-expenditure concepts, the credit still requires satisfaction of additional statutory tests and exclusions.

Uncertainty is defined as existing if the information available to the taxpayer does not establish:

• The capability of developing or improving the business component;
• The method for developing or improving it; or
• The appropriate design of the component.

An unresolved technical question can support the uncertainty requirement, but it is not enough by itself to establish credit eligibility. The taxpayer must also substantiate the other requirements.

The Reconfigured Process of Experimentation

TD 9104 also redefined the “process of experimentation” test to be more consistent with commercial R&D practices. The regulations clarify that a process of experimentation involves three core steps:

• The identification of uncertainty concerning the development or improvement of a business component;
• The identification of one or more alternatives intended to eliminate that uncertainty; and
• The evaluation of those alternatives through modeling, simulation, or a systematic trial and error methodology.

Appropriate-design uncertainty can support experimentation even when capability or method is established. Thus, technical feasibility alone does not disqualify research under the later regulatory framework.

The Enduring Legacy of TAASC in Modern Software R&D

TAASC remains useful historical context. Modern software claims should be evaluated under the statutes and regulations applicable to the claim year, with separate attention to general research requirements and any internal-use-software restrictions.

Internal Use Software (IUS) and the High Threshold of Innovation

TAASC concerned software developed for customers; internal-use classification was not the central dispute in the opinion. Separately, qualifying internal-use software generally must satisfy the high threshold of innovation as well as the ordinary research-credit requirements, unless an exception applies.

Current internal-use rules focus on general and administrative functions. Software marketed to third parties, or enabling specified third-party interactions, is treated differently. Dual-function software has additional rules. Merely supporting a customer service does not automatically establish an exception. The high threshold of innovation is a separate test, not a revival of the former general discovery standard.

The Exclusion of “Research After Commercial Production“

Section 41 excludes research after commercial production begins. Under the regulations, this point is reached when the component meets the taxpayer’s basic functional and economic requirements. This exclusion arises from the statute and regulations, not from TAASC alone.

The regulations identify the following activities as outside qualified research in this context:

• Preproduction planning and tooling up;
• Trial production runs and troubleshooting production equipment;
• Debugging flaws in a business component once it is ready for use;
• Accumulating data relating to production processes.

Testing should be classified by its actual purpose and timing. Routine quality-control testing is excluded, while research on a new improvement may qualify separately if the requirements are met. Labels such as debugging, prototype, or development do not determine the answer.

Contemporaneous Documentation: The Modern “Failure of Proof” Trap

Little Sandy Coal Company v. Commissioner, decided by the Seventh Circuit in 2023, and Phoenix Design Group, Inc. v. Commissioner, decided by the U.S. Tax Court in 2024, illustrate the need to prove qualifying activities. Phoenix Design is not a 2024 Eighth Circuit decision. Neither case establishes that the legal burden of proof has universally increased since TAASC.

Little Sandy Coal and Activity-Based Substantiation

Little Sandy Coal involved claims for eleven vessels, with two projects selected for trial. The Seventh Circuit affirmed denial because the evidence did not establish that substantially all relevant research activities constituted elements of experimentation.

The court specifically noted that:

• Novelty of a vessel does not establish that the activities used to construct it satisfy the experimentation test.
• Broad estimates need a reasoned evidentiary basis; the decision does not impose a categorical ban on all estimates.
• The substantially-all threshold is at least 80% of research activities, measured by cost or another consistently applied reasonable basis. This activity test differs from the separate employee-wage and Form 6765 thresholds.

The practical lesson is to connect evidence to activities, rather than infer qualification from the finished product. This is not the same issue as TAASC’s historical separate-information requirement. Records should support both the nature of the work and the allocation of expenses.

Phoenix Design and the Engineering Design Problem

In Phoenix Design Group, Inc. v. Commissioner, T.C. Memo. 2024-113, the Tax Court held that none of three sampled engineering projects involved qualified research. The firm designed mechanical, electrical, plumbing, and fire-protection systems. The result depended on the evidence for those projects, not a categorical exclusion of engineering design.

The decision illustrates several distinctions relevant to engineering claims:

• Technical uncertainty: Ordinary project unknowns and calculations using established information do not automatically show qualifying uncertainty.
• Supporting records: Project documentation should identify the uncertainty being addressed and the activities undertaken to resolve it.
• Experimental activities: Describing an overall design workflow does not by itself prove a qualifying evaluation of alternatives. The actual work must satisfy the applicable requirements.

Disclosure and Expense Treatment: Form 6765 and Sections 174 and 174A

Changes to Form 6765 should be applied using the instructions for the relevant tax year. The December 2025 instructions, issued in February 2026, distinguish the timing of Section G requirements from separate amended-return claim requirements.

The Reconfigured Form 6765 Requirements

Under those instructions, the principal distinctions include:

• Section G business-component information is optional for tax years beginning before 2026 and required for years beginning after 2025, subject to exceptions.
• When Section G applies, its 80%/Top 50 rules govern component detail; remaining components are aggregated as directed. It is incorrect to state that every filer must individually describe every component.
• Amended-return research-credit claims have separate information requirements. Filers should follow the relevant instructions for component identification, activities, and expense categories.

These filing requirements do not replace the underlying duty to substantiate eligible research and expenses. Nor should the form changes be attributed directly to TAASC without evidence of that causal connection.

The Impact of TCJA and the Restoration of Domestic Expensing

The Tax Cuts and Jobs Act generally required five-year amortization of domestic research expenditures and fifteen-year amortization of foreign expenditures for tax years beginning in 2022 through 2024. Public Law 119-21 added Section 174A, allowing current deduction of domestic research or experimental expenditures for tax years beginning after December 31, 2024, with an alternative capitalization election. Foreign research expenditures remain subje