Executive Summary

This brief provides a structural analysis of the expert opinion, focusing on how the report is constructed rather than on the ultimate merits of the underlying dispute.

The analysis identifies the report’s core reasoning chain, the assumptions on which that reasoning depends, and early signals where the opinion may rely on unsupported inference, uneven evidentiary support, or underdeveloped methodology.

Key Analytical Vulnerabilities

The comparative technology framework asserts equivalence without demonstrating it.

Conclusions 1 and 2 position Napster alongside cassette decks, VCRs, CD burners, FTP, and email to establish that music sharing is a normalized consumer activity. The report enumerates these comparators but does not analyze whether Napster’s centralized index function, user scale, or network searchability distinguish it from those technologies in legally or technically material ways. Commercial availability of prior technologies is treated as evidence of legal permissibility, and functional equivalence to Napster is assumed rather than argued. The entire normalization framework that precedes and supports the downstream technical conclusions depends on a bridging inference the report never supplies.

The technical infeasibility conclusions move from specific failure modes to categorical impossibility without closing the analytical gap.

The file name argument identifies real ambiguity problems but reaches a universal conclusion from illustrative examples. The checksum argument shows that encoding variability produces multiple checksums per recording — supported by the Metallica letter data — but does not analyze whether partial matching, reference library approaches, or probabilistic thresholds could achieve functional screening short of perfect accuracy. The pre-authorization argument conflates the difficulty of cryptographic identity verification with the claim that the Internet’s architecture cannot support any authorization requirement. Each conclusion overstates what the underlying analysis establishes.

Conclusion 6 characterizes the recording industry’s engineering decisions as failures without establishing that technically adequate alternatives were available at the times those decisions were made.

The report attributes the absence of widespread watermarking to poor RIAA engineering management, then uses SDMI’s 1999 Musicode selection as retroactive evidence that a workable standard could have been adopted a decade earlier. That inference compresses a twenty-year technical development timeline without analyzing whether any specific technology was deployable at scale in the late 1980s or early 1990s. The normative characterization — “bungles,” poor management — is not supported by independent technical benchmarking and rests primarily on the expert’s own judgment layered over trade press accounts.

Conclusion 7’s legitimate use argument is asserted without empirical support.

The report presents space-shifting, preview use, promotional distribution, and instructional access as evidence that Napster supports meaningful legitimate uses it cannot distinguish from infringing ones. No usage data, user survey, or documented activity pattern supports these examples. They are constructed hypothetically. The technical claim that Napster cannot distinguish use-purpose is adequately supported. The factual claim that legitimate uses constitute a non-trivial share of actual Napster activity is not.

The bot detection concession in Conclusion 9 sits in tension with the broader infeasibility argument.

The report acknowledges that Napster can detect automated access in real time and block IP addresses during active sessions. The distinction drawn — that bot detection operates during a session while user blocking must occur at login — is described at a general level without technical specification. The report does not document detection signals, reliability rates, or error tolerances. The more structurally significant issue is that the demonstrated capacity for real-time behavioral monitoring creates a question the report does not fully resolve: whether that monitoring architecture carries implications for other content-level filtering capabilities the report characterizes as infeasible.

Primary Scrutiny Targets

The basis for functional equivalence between Napster and prior consumer technologies.

The expert should be examined on what analytical steps support treating a networked, centrally indexed file-sharing system as legally and technically equivalent to a cassette deck or CD burner. The answer will expose whether the comparison is grounded in technical analysis or structural assertion.

The gap between demonstrated failure modes and categorical infeasibility.

On each of the three identification mechanisms — file names, checksums, pre-authorization — the expert should be pressed on whether the analysis established that the mechanism cannot function at any useful threshold, or only that it cannot function perfectly. The distinction matters for both admissibility and factual weight.

The methodological foundation for the watermarking chronology.

The expert’s opinion that RIAA’s engineering decisions were deficient and that adequate technology was available earlier rests largely on trade press sources and personal attendance at a 1993 conference. Deposition should establish what independent technical analysis, if any, supports the deployability assessments underlying that conclusion.

The empirical basis for legitimate use prevalence.

The expert should be examined on what, if any, data supports the claim that space-shifting, preview use, and similar activities represent actual Napster usage rather than hypothetical possibilities constructed for the opinion.

The technical specifics of bot detection.

xamination on how Napster’s bot detection mechanism actually operates — its signals, reliability, and error rates — will test whether the distinction between bot detection and user filtering is as clean as the report presents, and whether that demonstrated monitoring capacity has broader implications the report declines to address.

Section 1: Core Opinion Summary

The Expert’s Central Claim

Tygar’s primary opinion is that Napster cannot be held responsible for copyright infringement on its platform because the technical architecture of the system, and of the broader internet, makes identification, screening, and blocking of infringing content technically infeasible. That infeasibility, as the report constructs it, is not a temporary engineering limitation. It is presented as structural and inherent to how distributed networks and file-sharing systems operate. The report advances a secondary claim alongside this: whatever technical tools might have addressed the problem were available to the recording industry years earlier and were not adopted due to poor engineering management by the RIAA, not any failure by Napster.

These two claims work together. The first neutralizes Napster’s capacity to act. The second redistributes the locus of responsibility for the absence of technical controls.

The Sequencing of Supporting Conclusions

Tygar structures nine numbered conclusions, but they are not equal in argumentative weight. The load-bearing conclusions are 3, 4, 5, 6, and 7. Conclusions 1 and 2 perform a different function: they position Napster within a broader category of existing, legally tolerated technologies. Conclusions 8 and 9 are narrower defensive positions addressing specific technical mechanisms Napster does or does not use.

The sequence matters. Conclusions 1 and 2 establish that Napster is not categorically different from other consumer technologies that reproduce or distribute music — cassette decks, VCRs, CD burners, ripping software, FTP, email, and web search engines, all presented as functional analogues. This framing precedes and supports the downstream technical conclusions by establishing that Napster exists within a recognized category of dual-use technology rather than as a novel and uniquely infringing tool.

Conclusions 3 and 4 carry the core technical argument. Napster cannot identify copyrighted content because no reliable mechanism for doing so exists within the system. File names are ambiguous and easily manipulated. Checksums are unreliable because a single source recording produces different digital files depending on encoding variables. Pre-authorization is infeasible because Napster has no way to verify that an entity claiming rights actually holds them. Together, these two conclusions establish that the technical capacity to screen for infringing content is absent and cannot practicably be introduced.

Conclusion 5 extends this to the structural level. Requiring pre-authorization for shared content would fundamentally alter the architecture of the internet, converting a decentralized, peer-driven system into a centrally controlled distribution channel — shifting the infeasibility argument from “Napster cannot do this” to “the internet as designed cannot do this without becoming something else.”

Conclusion 6 turns the argument toward the recording industry. The RIAA has engaged in technical efforts to mark rights information in digital recordings since at least 1980. Despite two decades of effort and significant resources, no workable, widely deployed watermarking standard exists. The report attributes this to poor technical decision-making by the RIAA, specifically the failed Copycode system and the subsequent BBN watermarking proposal that SDMI ultimately rejected. The conclusion follows that if workable watermarking technology had been adopted in the late 1980s or early 1990s, Napster and similar systems would have had a technical basis for screening content. The report attributes the absence of that technology to the recording industry’s engineering failures, not to any inherent impossibility.

Conclusion 7 adds a use-context dimension. Even if Napster could identify a file as containing copyrighted material, it could not determine whether a particular use of that file was infringing. Space-shifting, format-shifting, personal archiving, promotional distribution by artists, and sampling for purchase decisions are all presented as potentially legitimate uses of copyrighted material that Napster cannot distinguish from infringing use.

Conclusions 8 and 9 address two specific technical mechanisms. Conclusion 8 defends Napster’s use of ID and password authentication as reasonable and superior to alternatives including IP address blocking, biometrics, smart cards, and credit card authentication. Conclusion 9 explains why Napster blocks automated bots, citing performance maintenance, while distinguishing that capacity from the more complex problem of blocking users named on an infringement list. These conclusions appear to respond to anticipated arguments that Napster’s existing technical controls demonstrate a capacity it declines to exercise against infringers.

The Reasoning Path from Technology Comparison to Infeasibility

The report builds its argument in two broad phases.

The first phase, covering Conclusions 1 and 2, is primarily comparative and contextual. Napster is placed alongside consumer recording technology that has existed for decades and internet file-sharing infrastructure that predates Napster by years. This is a recognized argumentative pattern in technology infringement cases: the defendant’s system is positioned within a class of technologies that courts have legally tolerated or explicitly protected under frameworks like the Betamax doctrine. The report does not invoke Sony v. Universal City Studios by name in the portion of the text provided, but the structural logic of the comparison follows that pattern closely. By the time the report reaches Conclusion 1’s discussion of dual dubbing cassette decks, CD burners, and ripping software — including the observation that Circuit City advertises computers with integrated CD recording capability — the framing is clear: music reproduction and sharing are normalized consumer activities supported by mainstream commercial products.

The second phase, covering Conclusions 3 through 7, builds the technical infeasibility argument in layers. Each conclusion addresses a different level at which screening might theoretically be implemented and explains why that level cannot function. File name filtering fails because names are ambiguous and easily circumvented. Checksum matching fails because the same source material produces variable files. Pre-authorization fails because identity and rights verification cannot be reliably performed. Structural alteration of the internet to require pre-authorization is presented as technically incompatible with how distributed networks operate. Watermarking — the one mechanism that might have worked — is unavailable because the recording industry failed to standardize and deploy it.

Where Inference Supports the Argument

Several steps in the argument rest on inference or assertion rather than demonstrated technical analysis.

The claim that file name filtering is infeasible relies on hypothetical examples — the ambiguity of “BS” in a file name, the misspelling “Metalica” — rather than empirical data about the proportion of files that would be correctly identified through name-based screening. The report acknowledges that some users actively use misleading names and cites a website advocating this practice, but does not quantify the actual prevalence of ambiguous or misleading file names in Napster’s index at the relevant time.

The checksum analysis is more technically grounded. The report explains compression variable effects and references the Metallica letter data as a concrete illustration. The conclusion that checksum matching is infeasible as a general screening tool, however, is stated rather than demonstrated through systematic analysis of matching rates or error tolerances.

The RIAA watermarking history is extensively documented through news articles and trade press coverage. The conclusion that watermarking technology was available and deployable in the late 1980s or early 1990s rests primarily on Tygar’s own assessment and his 1993 conference attendance, not on independent technical benchmarking.

The argument that SDMI’s quick selection of a transitional watermarking standard demonstrates that a workable standard could have been adopted a decade earlier is inferential. It treats the 1999 selection process as retroactively establishing what was achievable in 1989.

Known Expert Argumentation Patterns

Two recognizable patterns appear in the structure of this report.

The first is contextual normalization through technological comparison. Tygar places Napster within a broadly accepted class of prior technologies to argue that the challenged conduct is not categorically different from what courts have already tolerated. This pattern is common in technology cases where the central question involves a dual-use system. The risk is that the analogy may not hold at the relevant level of analysis. The question is often not whether music sharing is generally legal but whether the specific mechanism and scale of the defendant’s system creates liability that prior analogues do not.

The second is responsibility displacement through prior party failure. Conclusion 6 directs attention toward the recording industry’s engineering decisions as the proximate cause of the absence of content identification technology — the argument being that Napster could have screened content if the RIAA had acted differently in the 1980s and 1990s. Two analytical dependencies underlie this argument: that a workable, widely deployed watermarking standard would have been technically effective, and that Napster would have adopted it. The report demonstrates neither. The displacement argument treats the RIAA’s engineering history as establishing that the tools were available and deliberately not deployed, but does not address whether Napster would have implemented watermark screening even if a standard had existed.

The Overall Argumentative Architecture

Napster cannot screen for infringing content because no reliable technical mechanism for doing so exists within the system’s architecture. The broader internet cannot accommodate mandatory pre-authorization without architectural transformation. The one mechanism that might have made screening possible — watermarking — does not exist in usable form because the recording industry failed to standardize it. Napster’s existing access controls are reasonable and appropriate for their purpose. And even if content could be identified, Napster could not determine whether a given use was infringing.

The opinion does not directly address whether Napster had knowledge of infringing activity or financial benefit from it. It addresses technical capacity and technical responsibility. That framing is consistent with a defense posture focused on the secondary liability elements of contributory and vicarious infringement analysis — specifically the knowledge and control elements — but the report approaches those elements through technical description rather than direct legal analysis.

Section 2: Reasoning Structure Map

The Tygar report does not advance a single linear argument. It operates as a layered defense across nine conclusions, each addressing a discrete legal or technical issue, but collectively assembling a unified position: Napster is a neutral technical intermediary that lacks both the capability and the information necessary to police copyright infringement, and the recording industry’s own engineering failures preclude placing that burden on Napster now.

Tracing the argument requires following not just individual conclusion chains, but the sequencing logic that connects them.

The Foundational Move: Normalizing Napster

Before reaching its technical conclusions, the report establishes a framing premise through Conclusions 1 and 2. These conclusions do not make a technical argument about Napster’s architecture. They make a normative equivalence argument, positioning Napster within an existing ecosystem of consumer technologies and Internet protocols that have long enabled music copying and file sharing.

Conclusion 1 proceeds by enumeration. The report lists consumer hardware — cassette decks, VCRs, CD burners, MiniDisc recorders, TiVo devices, MP3 players — and describes their copying functionality in varying detail. The analytical step is implicit rather than demonstrated: because these devices exist commercially and are sold to consumers, their copying functions carry some form of social or legal acceptance. The report does not establish what legal standard governs that inference. Commercial availability is treated as functional equivalence to legitimacy.

The reasoning chain within Conclusion 1 runs as follows: consumer devices have long enabled music copying and sharing; they are commercially available and widely used; Napster enables music sharing through the same fundamental mechanism; therefore Napster’s function is not categorically different from these accepted technologies. Each step depends on the prior one, but the transition from “commercially available” to “legally analogous” is stated rather than argued. That transition does not appear as a demonstrated analytical step.

Conclusion 2 extends the same structure to the Internet context, placing Napster alongside email, FTP, the World Wide Web, search engines, Gnutella, and Freenet. The analytical move is identical: file sharing is foundational to Internet architecture, it predates Napster by decades, and Napster represents one instance of a universally recognized technical function. The report then adds an argument about controllability, noting that Gnutella and Freenet operate without central control points and are difficult to regulate — implicitly positioning Napster’s centralized architecture as less problematic by comparison. That inference is not stated directly.

Together, Conclusions 1 and 2 establish the platform from which all subsequent technical conclusions operate. They are not independent arguments. They are load-bearing premises that shape how the later incapability arguments are received. If Napster is a continuation of established, socially accepted technology, the burden of providing copyright filtering reads differently than if Napster is a novel, purpose-built infringement tool.

The Technical Incapability Chain: Conclusions 3, 4, and 5

Conclusions 3, 4, and 5 build a sequential incapability argument. Each conclusion forecloses a category of technical intervention, ordered so that each subsequent conclusion addresses the logical follow-up question raised by the prior one.

Conclusion 3 begins with a definitional assertion: Napster has no access to copyright information because no reliable mechanism exists for identifying whether a given audio file is protected. The report evaluates three candidate identification methods — file names, checksums, and pre-authorization — and explains why each fails.

The file name analysis proceeds by describing the ambiguity inherent in user-assigned file names. The report provides concrete illustrations: initials that could correspond to multiple artists, the space-shifting user with idiosyncratically named personal files, deliberate mislabeling by third parties. The Metallica checksum data (Bates NAP008871-90) is introduced to quantify the checksum problem — 2,280,474 items with 470,846 distinct checksums, a figure substantially exceeding Metallica’s actual catalog. The report uses that data to demonstrate that even within a single artist’s catalog, checksum-based identification produces an unmanageable number of variants.

The checksum analysis rests on a technical premise: that multiple MP3 encodings of the same source material will produce different checksums because compression variables, analog stages, mastering differences, and software variation all affect the output file. The report treats this as established without citing a technical source directly. It is supported inferentially through the Metallica data and the expert’s general expertise.

The pre-authorization analysis identifies a different failure mode: even if Napster attempted to collect authorizations from rights holders, it would have no reliable way to verify that the submitting party actually held the rights. This concern bridges directly to Conclusion 4.

Conclusion 4 is shorter and functions as a logical corollary to Conclusion 3’s third approach. Authorization requires identity verification; identity verification on the Internet is unreliable because copyright registration documents are public and email addresses can be fabricated; therefore pre-authorization cannot serve as a functional substitute for copyright identification. The report supports this with a specific scenario — someone using a falsified Hotmail address in the name of a Metallica member — that illustrates the identity verification gap concretely.

Conclusion 5 addresses a broader systemic argument. Even setting aside the identity verification problem, requiring pre-authorization for file sharing would impose a centralized gatekeeping function incompatible with the Internet’s decentralized architecture. The report makes a structural claim: the Internet is designed as a ground-up, decentralized information network governed by the IETF, and a mandatory authorization layer would alter that technical model in ways that create performance problems and may be technically infeasible.

Across Conclusions 3, 4, and 5, the foreclosure is sequential: copyright status cannot be identified from the file itself; the identity of someone claiming to authorize a file cannot be reliably verified; and even if both were possible, the systemic cost would be incompatible with Internet architecture. The chain is constructed so that overcoming one objection surfaces the next, and each conclusion depends on the assumptions embedded in the prior one.

The Industry Fault Insertion: Conclusion 6

Conclusion 6 operates differently from the incapability chain. Rather than arguing that a technical solution is impossible, it argues that a workable technical solution was not adopted, and places responsibility for that failure on the recording industry.

Watermarking technology capable of carrying rights information has existed since at least the early 1990s. The RIAA made a series of poor engineering decisions across two decades — Copycode, the BBN system, and the slow path to SDMI — that prevented any marking standard from being adopted. SDMI, once formed, selected a transitional watermarking technology (Musicode) within three months. The report’s inference is direct: had the RIAA applied comparable engineering management in the late 1980s or early 1990s, recordings in circulation today would carry copyright information that Napster could read.

This conclusion does not argue that Napster should have implemented watermark screening. It argues that Napster cannot implement what does not exist, and that the absence of a deployed watermarking standard is attributable to the industry’s own engineering decisions.

The report assembles this argument through a documented chronology — the 1980 RIAA letter to universities, the Copycode failure, the NBS test results, the BBN system and its rejection by SDMI, and the SDMI Musicode selection — sourced primarily through trade press and news articles cited by publication, date, and author. The expert also incorporates direct personal experience, including attendance at the 1993 Harvard workshop on digital watermarking and awareness of Digimarc’s commercial watermarking deployment for image protection.

The structural significance of Conclusion 6 is that it repositions the plaintiff’s implicit claim — that Napster should filter on copyright — as a demand that Napster implement a capability the industry itself failed to build.

The reasoning depends on a temporal assumption: that had the RIAA adopted a working standard in the late 1980s or early 1990s, that standard would be sufficiently widespread today to give Napster reliable screening capability. The report asserts this as a reasonable inference but does not model what adoption rates or deployment timelines would have looked like. That gap is a point of potential vulnerability under examination.

The Legitimate Use Argument: Conclusion 7

Conclusion 7 introduces a separate analytical strand: even if Napster could identify copyrighted material, it could not determine whether a particular use of that material was infringing. The report presents a series of use scenarios — personal space-shifting, format conversion, preview before purchase, promotional distribution by artists, classroom access — and asserts that these appear to be legitimate uses even when they involve copyrighted recordings.

The reasoning structure is enumeration without quantification. The report does not assess how frequently these use cases occur relative to infringing uses. It presents them as sufficient to establish that Napster cannot distinguish legitimate from infringing use, and therefore cannot function as an infringement filter even if it could identify copyright status.

This adds a second layer to the incapability chain. Conclusion 3 argues Napster cannot identify copyright status. Conclusion 7 argues that even successful copyright identification would be insufficient, because use-purpose cannot be assessed from the technical transaction alone.

The Operational Capacity Arguments: Conclusions 8 and 9

Conclusions 8 and 9 address specific technical capabilities Napster does possess — ID/password authentication and bot detection — and explain why neither extends to copyright enforcement.

Conclusion 8 argues that Napster’s ID/password system is reasonable and technically superior to alternatives. The report evaluates IP address blocking, name-based identification, biometrics, smart cards, public key cryptography, and credit card authentication, describing the practical limitations of each. ID/password represents the appropriate standard of care for user authentication in this context.

Conclusion 9 is worth reading carefully. Bot detection does not demonstrate a general capacity to filter by content. Bots are detected in real time based on behavioral signatures during an active session; user authentication occurs at login. A user who has been banned and returns with new credentials cannot be identified through bot-detection methods. The report uses an analogy to anchor the distinction: a shopkeeper who can catch a shoplifter in the act cannot screen incoming customers against a list of known shoplifters. These are technically distinct capabilities, not variations of the same one.

Both conclusions function as a preemptive response to an anticipated argument: that Napster’s demonstrated ability to perform certain filtering shows it could implement copyright filtering if it chose to. The report addresses that inference directly, distinguishing the mechanism and timing of each capability and explaining why they do not generalize.

How the Reasoning Chain Assembles

The architecture of the full argument runs in sequence. Napster is normalized within existing technology and Internet infrastructure. The incapability chain forecloses identification, verification, and systemic authorization. The industry fault argument explains why the tools that would enable screening do not exist. The legitimate use argument establishes that even capable screening would be insufficient. The operational arguments rebut the inference that Napster’s existing filtering capabilities demonstrate a broader filtering capacity it has declined to use.

Each conclusion is designed to be self-contained, but they are sequenced so that later conclusions address objections that would naturally arise from accepting the earlier ones. The reasoning moves from external normalization, to technical incapability, to industry responsibility, to use-purpose indeterminacy, to operational scope limitation.

The transitions between conclusions are generally stated rather than demonstrated. The report does not establish an explicit logical connection showing that each conclusion supports the next. The sequencing is visible structurally, but it is not argued as a chain.

Section 3: Evidence and Data Sources

The evidentiary foundation of the Tygar report is heterogeneous. It draws from news articles, industry documents, government reports, product advertisements, web-accessible software directories, litigation correspondence, patent filings, press releases, and the expert’s own professional experience. These sources are not uniformly integrated into the argument. Some conclusions rest on multiple independent sources. Others rest on a single article, a product description, or a personal observation. The distribution of evidentiary support across the nine conclusions is uneven in ways that matter for how the report functions as an argumentative document.

Industry and Consumer Product Sources

The largest category of cited materials supports Conclusion 1, which holds that Napster allows users to share recorded music in ways comparable to existing consumer technologies. The report cites product descriptions from Circuit City’s commercial website for two HP computer models (HP 8655C and HP 8665C), including verbatim advertising copy describing CD-RW capability and the bundled MusicMatch Jukebox software. It also references the HP 8670 product description from an HP-affiliated URL. The Adaptec Easy CD Creator 4 Deluxe datasheet is cited, and the report reproduces the product’s copyright warning verbatim.

These sources function as documentary evidence that consumer-facing products were commercially available and openly advertised as enabling music copying and burning. The report does not use them to demonstrate that the products were widely used or that they were the functional equivalent of Napster in terms of scale or copying scope. They establish that the products existed and were sold through major retail channels.

The inference that Napster is comparably situated to these products depends on that comparison holding at the level of capability, which these sources support in a literal sense.

The report also cites Sony product pages for the Memory Stick Walkman, two specific Sony URLs, and a Sony Vaio Music Clip user manual distributed over the web. These sources illustrate that MP3 playback and file download capabilities were standard features in commercially available consumer devices, and that Sony, itself a major recording industry stakeholder, distributed products and documentation enabling MP3 use. The report identifies Sony’s own MP3-related distribution activities as contextually significant, though the analytical step connecting Sony’s product distribution to the legal or factual questions in the litigation is not spelled out.

Ripping and encoding software is documented through a set of mp3.com software directory URLs covering Windows, Macintosh, Unix, and other platforms. AOL’s corporate website is cited for the claim that Winamp had 25 million registered users as of June 1999, and AOL’s web center page is referenced as a pointer to further MP3 software resources. These sources establish the volume and ubiquity of MP3-related software rather than any specific technical characteristic of that software. The report does not analyze the functionality of these tools beyond what the cited web pages describe. Dual dubbing CD recorders are noted through a brief reference to two Philips models (CDR765 and CDR870), without a specific source document; the report treats their existence as self-evident.

News and Trade Press Sources

The Conclusion 6 section on watermarking relies heavily on trade press and journalism. Ten sources anchor this discussion: a 30 June 1987 Wall Street Journal article by Stephen Kreider Yoder on DAT recorders and RIAA lobbying for Copycode; a 2 March 1988 Wall Street Journal article by Jeffrey A. Tannenbaum reporting that the National Bureau of Standards found Copycode unreliable, audio-quality-degrading, and easily bypassed; a 3 September 1988 Daily Telegraph article by Barry Fox on Copycode’s early origins; a 24 August 1987 Wall Street Journal article by Gregory Sandow on audio industry dynamics; a 19 January 1996 Audio Week item reporting that Copycode’s inventor was appointed to a senior RIAA post; a 2 March 1996 New Scientist article by Barry Fox describing the BBN watermarking system and its technical vulnerabilities; a 23 September 1999 Daily Telegraph article by Fox on SDMI’s rejection of the BBN system; a 13 September 1999 Audio Week item on watermarking questions and delays; a 16 August 1999 Audio Week item on the selection of Musicode for DVD Audio and SDMI; and a 27 July 2000 New York Times article by Amy Harmon on AOL and InterTrust Technologies.

Together these sources construct a historical narrative of RIAA’s watermarking efforts from 1980 through 2000. The report uses them to support the claim that the recording industry had multiple technically viable opportunities to implement rights-marking technologies and declined to do so effectively. The NBS findings on Copycode are cited to establish that government testing confirmed technical failure. The Barry Fox articles, appearing across multiple publications over more than a decade, form the primary evidentiary thread for the claim that the BBN system was technically flawed and subsequently rejected by SDMI.

No primary technical documents support these technical conclusions. The NBS test results appear through newspaper reporting rather than through the NBS report itself, which the report does not cite directly. The technical findings attributed to the NBS — that Copycode was unreliable, degraded audio quality, and could be bypassed using five distinct methods — therefore rest on secondary journalism rather than the underlying government testing document.

Government and Institutional Sources

The October 1989 Office of Technology Assessment report, Copyright and Home Copying: Technology Challenges the Law, is cited for a block quotation describing the range of technically feasible copy-protection approaches as characterized by the RIAA Engineering Committee in April 1989. The report uses this source to establish that RIAA itself believed a marking scheme was within technical reach in 1989, and that the failure to implement one was therefore attributable to engineering decision-making rather than technical impossibility.

The OTA report is the only government document cited, and it functions as an industry-generated admission: the substance of the quotation comes from RIAA’s own Engineering Committee, which makes it potentially more durable than third-party assessments in terms of attribution.

SDMI Documents

Three URLs pointing to the SDMI members-only website are cited in support of the watermarking discussion: the SDMI Call for Proposals for screening digital content, dated 5 May 1999; a summary of responses dated 30 May 1999; and Sony’s proposal dated 22 May 1999.

The report reproduces a portion of Sony’s proposal stating that Sony had been engaged in watermarking technology development for several years, then notes that Sony’s system was not selected by SDMI. These sources establish that multiple commercially capable entities submitted watermarking proposals within a short timeframe, and that SDMI was able to select a transitional standard (Musicode/Aris) within months of formation. The report uses this timeline to support the inference that RIAA’s earlier failure to adopt comparable standards reflected poor engineering management rather than technical unavailability.

The report does not reproduce or analyze the technical content of the SDMI proposals themselves. The analytical connection between the proposals’ existence and the viability of earlier adoption is asserted rather than demonstrated through technical comparison.

Litigation Correspondence

Two letters from attorney Howard King to Napster’s Sean Parker are cited, identified by Bates numbers NAP008871-90. The first, dated 3 May 2000, lists 1,456,075 items claimed as improperly copied Metallica recordings. The second, dated 18 May 2000, lists 2,280,474 items with 470,846 distinct MD5 checksums.

The report uses these letters to demonstrate the impracticality of checksum-based copyright identification. The argument is that the number of distinct checksums vastly exceeds the number of actual Metallica recordings, making checksum matching an unreliable identification mechanism. The report also notes that two items in the Metallica list appear to be attributed to Chris Isaac, and that a number of listed items are identified as live recordings, which King’s own cover letter excludes from Metallica’s copyright claims.

These letters function as adverse-party-generated data points. The report uses them not for their legal content but as a concrete illustration of checksum proliferation under real-world conditions. The analytical inference — that a finite set of copyrighted recordings would generate an unmanageable number of distinct checksums — is supported by the letter data, but the report does not perform an independent technical analysis of the checksum methodology itself.

Patent and Academic Sources

The BBN watermarking patent (international patent application WO 93/12599) is cited and summarized through the New Scientist article rather than cited directly. Digimarc’s image watermarking technology is referenced alongside a 20 June 1997 Digimarc press release on Playboy’s adoption of the technology, cited to establish the commercial viability of watermarking outside the audio context. These sources support the claim that watermarking was technically well-understood and commercially deployed in adjacent fields by the mid-1990s.

The report also references academic work on digital steganography, specifically Kineo Matsui’s work presented at a 1993 Harvard Kennedy School workshop that Tygar personally attended, and two of Matsui’s earlier 1990 papers. The first-hand account of the workshop and the subsequent publication of the proceedings in Volume 1, Issue 1 of The Journal of the Interactive Multimedia Association Intellectual Property Project establish that audio watermarking had been discussed as a technical concept at least since the early 1990s.

This category of sources establishes a timeline for when watermarking technology was technically accessible. The analytical connection between the availability of these techniques and what the recording industry could or should have implemented is driven primarily by the expert’s own technical judgment rather than by the cited documents themselves.

Web-Sourced Materials and Self-Citation

The report cites web pages for stopnapster.com (including the “trojans” and “bombs” subpages), the ID3 field documentation at id3.org, the Gnutella and Freenet project pages, SpinFrenzy’s homepage, the bookfinder.com shopping bot example, and an eBay bot-detection news item from Netscape’s technology news service. These sources illustrate specific technical points — the practical futility of file name filtering, the structure of P2P alternatives, and the performance rationale for bot rejection — rather than establishing factual foundations for the expert’s central conclusions.

At several points the report incorporates personal observations and professional experience as evidentiary inputs: Tygar’s attendance at the 1993 watermarking workshop, his ownership of audio equipment including a Denon minisystem and MD recordings, his experience with ISP IP address reassignment through PacBell, and his personal Napster searches for variant spellings of “Metallica.” These are not cited documents but are treated within the report as corroborating evidence for specific technical claims. The Data Compression FAQ and the Microsoft Developers Network are listed in the introduction as background materials used but not explicitly cited in the text. No specific content from either source appears in the report’s cited analysis.

Section 4: Assumption Dependencies

Napster’s technical neutrality as the baseline condition

The entire opinion structure rests on a foundational assumption that Napster is technically neutral with respect to copyright status — that it cannot distinguish protected from unprotected material because no technically feasible mechanism exists to make that distinction at the system level. This assumption does not arise from a formal technical test of Napster’s architecture. The report states it as a premise and then illustrates it through a series of technical sub-arguments, treating the inability to distinguish copyright status as an inherent property of file-sharing systems generally, not as a design choice specific to Napster.

That framing carries structural weight throughout the report. Every conclusion touching Napster’s capacity to screen, block, or authorize content depends on this premise holding.

If technical neutrality is treated as a fixed characteristic rather than a design parameter, the question of whether Napster could have been designed differently does not arise within the report’s analytical frame.

The inadequacy of file-name identification is treated as conclusive

Conclusion 3 builds substantially on the premise that file names are too ambiguous and too easily manipulated to serve as a copyright identification mechanism. The report treats this inadequacy as both categorical and permanent — not a limitation of any particular implementation, but a structural feature of how file names work.

The supporting argument relies on examples: the ambiguity of abbreviations like “BS,” the difficulty of identifying classical recordings by title alone, and deliberate mislabeling by users seeking to evade identification. These are real phenomena, but the report treats them as sufficient to establish that file-name-based screening is categorically infeasible rather than impractical in its most basic form. Whether a more sophisticated or augmented file-name system — one incorporating metadata, genre information, or verified artist identifiers — would face the same categorical limitation is not examined.

The report also assumes that user workarounds would be widespread and successful enough to defeat any file-name-based screening. The evidence for this is the existence of a deliberate mislabeling website and a Napster search result showing hits on a misspelling of “Metallica.” The assumption that adversarial workarounds would neutralize screening at scale is embedded in the analysis without any quantitative framing.

Checksum analysis fails because of MP3 encoding variability

The checksum argument in Conclusion 3 depends on the assumption that different MP3 encodings of the same underlying recording will always produce different checksums. The report supports this by identifying encoding variables: compression settings, analog conversion steps, mixing and mastering differences, start and stop points, and encoding software.

This reasoning assumes that variability across different encodings is sufficiently large and sufficiently common to make checksum-based identification non-viable as a general approach. The report does not examine whether a reference database approach — checksum lists derived from the actual files circulating on Napster — could partially address this problem. The Metallica list example, showing 470,846 distinct checksums for a single artist, illustrates the scale of the problem, but the report treats that figure as evidence of infeasibility rather than as a bounded dataset with its own structure.

Whether partial matching, acoustic fingerprinting, or probabilistic identification could address encoding variability is not analyzed. The assumption that checksums are the only available technical identification mechanism is embedded in this section without examination of alternatives.

The RIAA’s failure to adopt watermarking forecloses Napster’s ability to screen content

Conclusion 6 depends on a sequenced set of assumptions. The first is that watermarking technology, had the recording industry adopted it in the 1980s or early 1990s, would today be embedded in a sufficient proportion of commercially released recordings to make screening practical. The second is that the industry’s failure to adopt such a standard is the proximate reason Napster cannot perform rights-based screening today. Both assumptions work together to locate the source of the screening gap in the recording industry’s technical decisions rather than in Napster’s architecture.

The report does not examine whether Napster could have incorporated any form of screening based on currently available metadata, existing partial watermarking systems, or SDMI’s transitional technology at the time the system was deployed. The operative assumption is that the absence of a universal, industry-adopted watermarking standard makes any Napster-side screening technically unavailable.

The report also assumes that even a transitional or imperfect watermarking standard, if adopted earlier, would have produced meaningful coverage of the relevant recordings. This assumption is not quantified. The argument is that “most recordings” would carry rights information, but the coverage threshold required to make screening operational is not established.

Authorization verification is assumed to require universal technical solutions

Conclusion 4 frames the authorization verification problem in terms of identity authentication at internet scale. The report assumes that checking whether a submitted authorization actually comes from a rights holder is technically infeasible because internet identity mechanisms — email addresses, usernames, return addresses — are easily forged. This treats the problem as one requiring certainty of identity rather than a risk-based or probabilistic approach.

The report does not examine whether a verification system based on registered commercial entities, established industry relationships, or tiered authentication would face the same categorical limitation. The assumption that any authorization check must resolve identity with near-certainty, and that nothing less would be practically useful, is not stated explicitly but shapes how the conclusion is framed.

The decentralized architecture of the internet is assumed to be both technically fixed and normatively authoritative

Conclusion 5 rests on the assumption that the decentralized, ground-up structure of the internet is a fixed technical parameter. Requiring pre-authorization for file sharing, the report argues, would change the internet’s architecture in a way that would severely degrade its function or render it inoperable. The report states that such a requirement “would completely change the nature of the Internet” and that “it is not even clear that the World Wide Web could technically continue to function in such a top-down model.”

This conclusion assumes that any authorization requirement would have to operate globally, across all internet traffic and all file-sharing utilities. The possibility that a targeted authorization requirement — one applying to a specific platform or service without altering the internet’s general architecture — would face the same systemic consequences is not addressed.

The report’s framing collapses the question of Napster-specific obligations into a question about internet-wide architectural change.

Napster’s comparators are assumed to have equivalent legal and functional standing

Conclusions 1 and 2 both rely on the assumption that the listed consumer technologies — cassette decks, VCRs, CD burners, FTP, email, Gnutella — are functionally equivalent comparators to Napster in the relevant respect. The report presents the comparator list as establishing that Napster is not meaningfully different from existing, widely accepted technologies.

This requires that the legally and technically relevant features of Napster are shared with its comparators. Whether Napster’s centralized index function, the scale of its user base, the speed of access it enables, or the structure of its network effects distinguish it from the listed technologies in ways that might be relevant to the legal analysis is not examined. The comparator framework assumes equivalence at the level of the file-sharing function without analyzing whether other dimensions of the comparison carry analytical weight.

The report also assumes that the widespread commercial availability of the comparator technologies is itself evidence of their legal permissibility, and by extension, evidence that Napster occupies the same legal space. This bridging assumption connects commercial availability to legal status without explicit analysis.

Legitimate uses are assumed to be sufficiently numerous to establish non-infringing purpose

Conclusion 7 presents a list of use cases — space-shifting, preview use, instructor use, format conversion — as evidence that Napster supports a variety of legitimate uses it cannot distinguish from infringing ones. The embedded assumption is that the existence of these use cases is legally and factually relevant to the question of Napster’s liability.

For that assumption to hold, the described use cases must represent actual, non-trivial uses of the system, not merely theoretical possibilities. The report presents them as illustrative examples rather than as empirically documented usage patterns. Whether these uses constitute a meaningful proportion of actual Napster activity is not addressed. The conclusion treats their existence as sufficient without establishing their prevalence.

The ID/password system is assumed to represent the ceiling of technically feasible user identification

Conclusion 8 rests on the assumption that ID/password authentication is the best available mechanism for user identification given the constraints of internet-scale systems. The report evaluates alternatives — IP addresses, biometrics, smart cards, public key cryptography, credit cards — and concludes that each is inferior or impractical. Each alternative is evaluated individually against a standard of universal, reliable, and scalable identification. Whether a layered approach combining multiple mechanisms could achieve higher identification reliability than any single mechanism alone is not examined.

The report also treats the state of these technologies at the time of writing as a fixed constraint. Whether any of the mechanisms evaluated — particularly public key infrastructure or biometric systems — were on a development trajectory that would alter the comparison within a relevant time frame is not addressed.

Bot detection and user blocking are assumed to be categorically different technical problems

Conclusion 9 depends on the assumption that the technical difference between real-time bot detection and prospective user blocking is significant enough that the ability to do one provides no meaningful information about the ability to do the other. The report frames this as a timing difference: bots are caught in the act while their IP address is known, whereas blocked users must be identified at login after the relevant infringement has already occurred.

IP address variability is treated as the decisive technical factor separating the two problems. The report does not examine whether a registry-based or credential-based blocking mechanism, independent of IP address, could bridge this gap. The operative assumption is that the technical architecture relevant to bot detection and the technical architecture relevant to user blocking are sufficiently distinct that success in one domain carries no implications for capability in the other.

The recording industry’s failure to adopt rights-marking standards is assumed to be relevant to Napster’s obligations

Running through Conclusion 6 and the broader structure of Conclusions 3 through 5 is a structural assumption that the recording industry’s own technical decisions are causally relevant to the question of what Napster can or should be expected to do. The report assumes that RIAA’s failure to adopt a viable watermarking standard over a twenty-year period means that the rights-identification infrastructure necessary to enable Napster-side screening simply does not exist, and that this absence is attributable to the recording industry rather than to Napster.

This frames a counterfactual: had RIAA adopted a workable standard in the late 1980s, Napster would today have the technical capability to screen content. The report presents this counterfactual as establishing that the absence of screening capability is not a result of choices Napster made. Whether Napster’s own design decisions contributed to or independently produced the screening gap is not analyzed within this frame.

Section 5: Early Vulnerability Signals

The Comparative Products Framework Rests on Asserted Equivalence

Conclusion 1 builds its central claim through a catalog of products: cassette decks, VCRs, CD burners, ripping software, and others. The report presents this catalog as demonstrating that music sharing is a longstanding, widely accepted consumer practice. What the report does not do is demonstrate functional equivalence between those devices and Napster in any analytically developed way.

The list of analog and digital devices is offered as evidence of cultural and commercial normalcy. But the structural question — whether Napster’s architecture differs in kind, not just degree, from those devices — goes unanalyzed. The report notes that Napster uses the Internet to facilitate sharing, then treats that distinction as immaterial. The analytical step that bridges consumer copying devices to a networked file-sharing index operating at scale is not visible in the report.

The comparison argument depends on the premise that Napster’s function is relevantly similar to home taping. If the scale, simultaneity, and searchability of Napster distinguish it from a cassette deck in legally or technically material ways, the comparison framework does not carry the weight placed on it. The report asserts comparability; it does not demonstrate it.

The Technical Infeasibility Claims Depend on Scope Assumptions That Are Not Established

Conclusions 3 and 4 rest on the claim that copyright identification and authorization verification are technically infeasible for Napster. The report evaluates three approaches — file name matching, checksum identification, and pre-authorization collection — and concludes that each fails. This is the most structurally consequential claim in the report.

The infeasibility argument for file names rests on the observation that file names are user-generated, ambiguous, and easily manipulated. Both points are supported with examples. But the argument reaches a universal conclusion from observations about specific failure modes. The report does not analyze whether partial identification, combined with other signals, could achieve a meaningful screening function — and the gap between “file names are unreliable in isolation” and “file name-based filtering is categorically infeasible” is not bridged by any analysis in the report.

The checksum argument relies on the observation that the same underlying recording can produce different checksums depending on encoding decisions, compression settings, and analog intermediary steps. The report supports this with the Metallica letter data: 2,280,474 items and 470,846 distinct checksums. But what percentage of those checksums a reference library of known checksums for commercially released recordings would eliminate goes unexamined. The argument moves from “checksums are not perfectly reliable” to “checksums cannot function as a filtering mechanism,” and that transition is not analytically supported. The Metallica data shows variability; it does not establish that a checksum-based filtering system operating at a threshold below perfect accuracy is impossible or useless.

The pre-authorization argument concludes that requiring rights-holder authorization would be technically infeasible and would transform the Internet’s architecture. The first part connects to the authentication problems analyzed in Conclusion 4. The second part — that authorization requirements would fundamentally restructure the Internet — is a structural policy claim rather than a technical finding. The report does not show the analytical path from “authorization is difficult to verify cryptographically” to “the Internet cannot function in a top-down authorization model.” Those are different claims, and only the first is technically grounded.

The RIAA Watermarking Narrative Carries Normative Weight the Report Frames as Technical

Conclusion 6 is the report’s most historically developed section. It documents the trajectory of RIAA’s copy protection efforts from 1980 through 1999, characterizes those efforts as marked by poor engineering decisions, and concludes that the absence of widespread watermarking technology results from those failures. The implicit structural claim is that the recording industry had the technical means to build a marking infrastructure and failed to deploy it through its own engineering and management deficiencies.

That framing carries significant weight in the report’s overall architecture, positioning the absence of marking technology as a circumstance attributable to the recording industry rather than to the inherent difficulty of the technical problem. The report does not analyze, however, whether commercially viable and technically adequate watermarking technology was, at any specific point in the timeline, actually deployable at scale across the full recording market. Characterizing RIAA’s decisions as “poor engineering” and “bungles” is a normative assessment. The technical record the report cites — Copycode failing NBS tests, BBN being rejected by SDMI — shows that specific proposals were found inadequate. It does not establish that an adequate proposal was available and simply not adopted.

The report cites SDMI’s selection of Musicode as a transitional technology as evidence that watermarking could have been deployed earlier. But the report does not analyze whether Musicode, or any equivalent technology, was technically mature enough for deployment in the late 1980s or early 1990s, when the report implies it should have been adopted. The argument compresses a twenty-year technical development timeline into evidence of avoidable delay. Whether that compression is analytically sound is a question the report does not address.

The Legitimate Use Examples in Conclusion 7 Are Asserted Without Empirical Support

Conclusion 7 identifies several scenarios in which Napster use with copyrighted material would be legitimate: space-shifting, previewing, distribution by artists, and access for instructional purposes. The report presents these as illustrations that Napster has “a variety of uses, many of which appear to be perfectly legitimate.” The word “appear” is the expert’s own framing. The conclusion is asserted normatively, not demonstrated analytically.

No usage data, user survey, or empirical analysis appears in the report to establish what proportion of Napster activity falls into the categories identified as legitimate. The examples are constructed hypothetically. The argument that Napster cannot distinguish legitimate from infringing use is a technical claim about system architecture that the report supports adequately. The argument that legitimate uses constitute a meaningful component of Napster activity is a factual claim about user behavior that the report does not support with evidence.

The ID/Password Analysis Evaluates Alternatives Without Establishing the Mechanism’s Own Effectiveness

Conclusion 8 defends Napster’s ID/password system as reasonable and superior to alternatives. The report evaluates IP address blocking, biometrics, smart cards, names, public key cryptography, and credit cards, and identifies problems with each. The conclusion is that ID/password authentication is preferable to those alternatives.

The structure of this argument is comparative elimination: ID/password wins because the alternatives are shown to be more problematic. Missing from that analysis is any examination of whether the ID/password system itself prevents circumvention by users who have already been blocked. The report acknowledges that a user can erase or manually edit the registry to defeat the blocking mechanism, then characterizes this as requiring “a fair level of technical expertise.”

Whether that technical threshold is meaningful in a user population that has already demonstrated facility with ripping software, MP3 encoding, and peer-to-peer networking is not analyzed.

The argument that the registry-based blocking system is effective depends on an assumption about the technical sophistication of the typical blocked user. That assumption receives no support from data or analysis in the report.

The Bot Detection Argument Introduces a Comparative Capability Claim That Creates Tension With Other Conclusions

Conclusion 9 addresses Napster’s capacity to detect and block bots. The report explains that Napster can identify bots while their activity is in progress and block their IP addresses in real time, then argues this is not contradictory to Napster’s inability to block known infringing users — because user blocking depends on authentication at log-on while bot detection operates during active session monitoring.

The analytical distinction between these two mechanisms is described at a high level. The report does not document the technical specifics of how Napster’s bot detection functions: what signals trigger detection, how reliably the detection operates, or what error rates exist. The argument that bot detection capability is categorically different from user filtering capability relies on a technical description that remains at the level of general explanation.

More structurally, the concession that Napster has the technical capability to identify and block certain categories of automated access creates tension with the broader infeasibility claims elsewhere in the report. The report draws a clear analytical line between bot detection and user filtering. Whether that line is as clean as presented, and whether the capacity demonstrated by bot detection carries implications for other monitoring capabilities, are questions the report does not fully resolve.

The Architecture of Conclusion 5 Rests on a Policy Claim Presented as a Technical Constraint

Conclusion 5 claims that requiring pre-authorization for file sharing would transform the Internet from a decentralized architecture into a centrally controlled distribution system. The report frames this as a technical consequence of the authorization requirement.

The claim that authorization would change the Internet’s fundamental architecture is presented as a technical finding, but the reasoning is primarily structural and policy-oriented. The report does not model what an authorization-checking mechanism would actually require in terms of technical infrastructure, nor does it demonstrate that no technically viable intermediate architecture exists between full decentralization and complete centralized control. The argument moves from “pre-authorization is difficult and poses authentication problems” to “pre-authorization would destroy the Internet’s decentralized character” without an analytical path connecting those two positions.

That transition from technical analysis to architectural policy argument carries the conclusion’s structural load. Whether it would withstand scrutiny as a technical expert opinion, rather than as commentary on Internet design philosophy, depends on how rigorously the Daubert gatekeeping inquiry distinguishes empirical technical analysis from normative structural claims.

Section 6: Litigation Impact Overview

The Tygar report engages the central causation structure of the plaintiffs’ copyright infringement claims at multiple levels simultaneously. The core factual issues in A&M Records v. Napster turn on whether Napster had knowledge of infringing activity, whether it materially contributed to that activity, and whether it had the technical capability to prevent or limit infringement. Tygar’s opinion addresses all three issues directly and, in doing so, reframes the factual predicate for each.

Knowledge and Identification Capability

Plaintiffs’ contributory infringement theory depends in part on establishing that Napster knew infringing material was being distributed through its system. Tygar’s Conclusion 3 positions the knowledge issue as a technical impossibility rather than a factual question about what Napster actually knew. The report argues that no practicable algorithm exists for identifying whether a given audio file is copyrighted, then methodically eliminates the three most plausible technical mechanisms — filename matching, checksum comparison, and pre-authorization screening — as each inadequate or infeasible.

That construction shifts the frame from what Napster chose to do to what Napster could have done. The reasoning implies that the absence of filtering reflected genuine technical limitation rather than a decision to permit infringement. The Metallica letter and checksum data — 470,846 distinct checksums for what Metallica claimed were its recordings — illustrate the checksum fragmentation problem at scale. If that illustration holds under scrutiny, it bears directly on the knowledge element by suggesting that even a good-faith filtering effort would have been technically unreliable.

The report does not establish what Napster’s actual knowledge was at any specific point in time. It argues structural incapacity — that the system lacked the architecture to know.

Whether that argument displaces the knowledge inquiry or merely reframes it is a central tension the report creates for the plaintiffs’ case.

Material Contribution and the Comparative Technology Framework

Conclusion 1 places Napster within a long catalog of consumer devices that allow users to reproduce and share music: cassette decks, CD burners, DAT recorders, VCRs, ripping software, and portable MP3 players. Conclusion 2 extends that comparative framework to file-sharing technologies, including email, FTP, the World Wide Web, search engines, Gnutella, and Freenet. Together, these two conclusions build a contextual argument that Napster represents a point on a continuum of widely accepted technologies rather than a novel infringing mechanism.

The Sony Corp. v. Universal City Studios framework — the substantial noninfringing uses doctrine — provides the obvious doctrinal context for this positioning, even though the report does not cite it by name. Conclusion 7 supplies that layer explicitly, listing seven categories of arguably legitimate Napster uses: space-shifting, time-shifting, format-shifting, promotional distribution by artists, academic use, preview before purchase, and portability. The report uses these categories to establish that Napster cannot distinguish infringing from non-infringing use — not only as a factual claim, but as an argument that the capability to facilitate harm does not, by itself, resolve the question of liability.

This structure interacts directly with the material contribution element of contributory infringement. Plaintiffs must show that Napster’s operation substantially assisted the infringing activity. By situating Napster’s functionality as derivative of and continuous with technologies that have not attracted liability, the report frames the material contribution question as one that requires a principled distinction Napster’s counsel can argue is unavailable.

Vicarious Liability and Practical Control

Vicarious liability requires both financial benefit from infringement and the practical ability to supervise or control infringing activity. Conclusion 4 addresses the control question by arguing that Napster cannot authenticate whether a party asserting authorization over a file actually holds the relevant rights. Conclusion 5 extends this by arguing that any pre-authorization requirement would transform the decentralized Internet architecture into a centralized gating mechanism, structurally incompatible with how the network operates.

Conclusion 9 addresses a potential inconsistency in this control argument. Plaintiffs could point to Napster’s documented ability to detect and block automated bots as evidence of technical control capability. Tygar resolves this by drawing a structural distinction: bots are detected in real time while their IP address is active, whereas user blocking must occur at login against a list that cannot be reliably keyed to identity. The IP address unreliability analysis in Conclusion 8 supports this, demonstrating that IP-based identification is insufficient for reliable user blocking. The shoplifter analogy in Conclusion 9 translates the technical point into accessible terms for a non-technical finder of fact.

Whether this distinction fully answers the control question depends on what the record shows. The report establishes the structural argument, but the factual record concerning what filtering capabilities Napster was actually asked to implement, and what it chose or declined to do, lies outside the report itself.

RIAA’s Role in the Current State of Technology

Conclusion 6 introduces a causation strand that runs in a different direction from the rest of the report. The watermarking analysis argues that the absence of widespread rights-marking technology in MP3 files is attributable to a series of engineering and policy failures by the RIAA and the recording industry, beginning with the Copycode debacle in the late 1980s and continuing through the SDMI process. The conclusion holds that had the RIAA adopted a viable watermarking standard in the early 1990s, MP3 players and services like Napster could have been built to screen for rights information embedded in the recordings themselves.

This argument places primary responsibility for the current technological gap on the recording industry’s own decisions, and within the litigation it complicates the causation picture for plaintiffs. If the conditions that make infringement difficult to filter are traceable to industry failures in rights-marking infrastructure, the degree to which Napster bears responsibility for the absence of that infrastructure becomes a contested factual and legal question.

The argument carries a practical limitation the report acknowledges only indirectly: watermarking protects only recordings that carry the watermark. It would not address bootleg recordings or content predating any adopted standard.

The report does not quantify what proportion of the allegedly infringing Napster traffic involved watermark-eligible recordings versus content that would remain unaddressed even under a fully deployed watermarking regime. That gap may affect the weight the watermarking argument carries once the factual record is fully developed.

Interaction with Daubert Admissibility

The report is structured as nine independent conclusions, each addressing a distinct technical or factual issue. Several conclusions — particularly those concerning file name ambiguity, checksum fragmentation, ID and password authentication, and bot detection — rest on computer science methodology and technical analysis within the expert’s disclosed specialization in software engineering, computer security, and cryptography.

Other conclusions involve more hybrid analysis. The watermarking history in Conclusion 6 draws largely on news articles, trade press coverage, patent summaries, and the expert’s personal attendance at a 1993 conference. The consumer product comparisons in Conclusions 1 and 2 rest primarily on publicly available commercial materials. The legal characterization of Section 512(a) within the Conclusion 2 discussion represents a legal opinion embedded in what is presented as technical analysis. The methodological foundation is not uniform across all nine conclusions, and those that depend more heavily on external sources than on the expert’s own technical analysis occupy different evidentiary ground under Daubert.

The report’s most technically grounded conclusions — the checksum analysis, the authentication comparison, the bot detection distinction — are its most durable portions from an admissibility standpoint. The watermarking history, the consumer product catalog, and the policy discussion in Conclusion 5 function primarily as contextual framing and do not carry the same methodological weight.

Summary of Litigation-Facing Issues

The opinion engages four pressure points likely to remain active throughout the litigation: the technical feasibility of copyright identification, the scope of Napster’s practical control over user activity, the comparative responsibility arising from the recording industry’s rights-marking failures, and the characterization of Napster as a technology with substantial legitimate uses. Each maps onto a distinct element of either contributory or vicarious infringement doctrine. The report does not resolve any of those issues as factual matters. It frames each as a technical question whose answer limits or forecloses liability. Whether the factual record supports those framings is the question the plaintiffs’ case must answer.