Skip to content

R0045/2026-03-29

Research R0045 — Prediction Verification
Mode Query
Run date 2026-03-29
Queries 7
Prompt Unified Research Standard v1.0-draft
Model Claude Opus 4.6

This run investigated seven queries related to enterprise technology predictions from the early 2000s, focusing on the transition from proprietary Unix/commercial software to Linux/open source, the industry analyst consensus, key incidents at OSCON 2004, the timeline of Wall Street's open source adoption, and the chilling effect of the SCO lawsuits.

Queries

Q001 — Unix Server Market Share — Sun dominated; Linux crossed over ~2010-2012

Query: What was the enterprise Unix server market share in mid-2001, specifically Sun Solaris vs Linux on Intel?

Answer: Sun held 54-68% of Unix/RISC market in 2001. Linux was ~$236M growing at 79% annually. Revenue crossover occurred ~2011-2012.

Hypothesis Status Probability
H1: Sun dominated, Linux grew, ~2010 crossover Supported
H2: Slower transition Eliminated
H3: Faster transition Partially supported

Sources: 3 | Searches: 2

Full analysis

Q002 — Web Server Market Share — Apache 63% by hostnames; measurement-dependent

Query: What was the web server market share in mid-2001, specifically Apache vs commercial alternatives?

Answer: Apache held 63.02% by hostnames (18.5M sites) vs IIS at 20.38%. But by physical servers, Windows led at 49.2% vs Linux at 28.5%.

Hypothesis Status Probability
H1: Apache dominant (63% hostnames) Supported by hostname count
H2: Not dominant in enterprise Partially supported by server count
H3: Measurement-dependent Supported

Sources: 3 | Searches: 2

Full analysis

Q003 — Open Source Enterprise Consensus — Cautious: viable for infrastructure, unproven for enterprise apps

Query: In 2001, what was the general industry and analyst consensus about whether open source software was a strategic risk or strategic advantage?

Answer: Gartner confirmed open source worked at infrastructure level but called results "inconclusive" for databases and middleware. Forrester identified lack of support as top barrier. The 2001 IT recession accelerated adoption. By 2004, IDC called Linux "clearly mainstream."

Hypothesis Status Probability
H1: Primarily a risk Partially supported
H2: Strategic advantage Eliminated
H3: Viable for infrastructure, unproven for enterprise Supported

Sources: 3 | Searches: 2

Full analysis

Q004 — OSCON 2004 Panel — Documented but differs from query framing

Query: Was the OSCON 2004 incident where Phil Moore from Morgan Stanley challenged panelists documented?

Answer: Yes. Doc Searls covered it in Linux Journal #7730 (November 2004), "We're Going to Be a 90% Linux Shop." However, Moore was a speaker at a breakout session, not an audience member challenging a hostile panel.

Hypothesis Status Probability
H1: Documented as described (panel challenge) Eliminated
H2: Not documented Eliminated
H3: Documented but details differ Supported

Sources: 1 | Searches: 2

Full analysis

Q005 — Financial Institutions Open Source — Three-phase adoption pattern

Query: What was the timeline of major financial institutions publicly adopting and contributing to open source?

Answer: Three phases: (1) Quiet adoption 2001-2010 (Morgan Stanley on Linux since 2001), (2) Early contributions 2012-2017 (Goldman Sachs GS Collections 2012), (3) Institutional embrace 2018+ (FINOS launched April 2018).

Hypothesis Status Probability
H1: Late 2000s contributions Partially supported
H2: Post-2015 resistance Eliminated
H3: Three-phase adoption pattern Supported

Sources: 3 | Searches: 2

Full analysis

Q006 — Doc Searls Article — Found: Linux Journal #7730, November 2004

Query: Search for a Doc Searls article in Linux Journal (approximately 2004) that quoted W. Phillip Moore from Morgan Stanley.

Answer: Article identified: "We're Going to Be a 90% Linux Shop," Linux Journal #7730, November 1, 2004. ACM DOI: 10.5555/1029015.1029027. Moore was a session speaker, not an audience challenger.

Hypothesis Status Probability
H1: Article exists and identified Supported
H2: Article not found Eliminated
H3: Exists but details differ Supported

Sources: 1 | Searches: 2

Full analysis

Q007 — SCO Lawsuit Chilling Effect — Real but informal; no systematic speech restrictions found

Query: Did the SCO legal threats cause a documented chilling effect on corporate employees speaking publicly about open source adoption?

Answer: SCO sent letters to 1,500 companies, demanded $1,399/CPU licensing, and was Microsoft-funded ($6-16M+). Phil Moore (Morgan Stanley) explicitly cited SCO as limiting his public speech. But systematic corporate speech restrictions were not documented. The net effect was paradoxically positive: code scrutiny strengthened Linux's legal standing.

Hypothesis Status Probability
H1: Systematic chilling effect with speech restrictions Partially supported
H2: No significant effect Eliminated
H3: Real but informal chilling effect; no systematic restrictions Supported

Sources: 3 | Searches: 2

Full analysis


Collection Analysis

Cross-Cutting Patterns

Pattern Queries Affected Significance
The 2001 IT recession as adoption accelerator Q001, Q003, Q005 The dot-com bust drove enterprises to Linux/open source as a cost reduction strategy, accelerating a transition that might otherwise have taken longer
Measurement methodology determines narrative Q001, Q002 Both the server market share (units vs revenue) and web server market share (hostnames vs physical servers) tell different stories depending on metric choice
SCO litigation's paradoxical effect Q004, Q006, Q007 SCO created real fear but ultimately strengthened Linux's legal position; Phil Moore's SCO-constrained speech at OSCON is the connecting thread between Q004, Q006, and Q007
Gap between quiet adoption and public contribution Q004, Q005, Q006 Financial institutions adopted open source years before they contributed publicly — Morgan Stanley used Linux from 2001 but Goldman Sachs' first open-source release was 2012, FINOS launched 2018
Analyst caution lagging market reality Q001, Q003 By 2002, Gartner was still calling open source database results "inconclusive" while enterprises were already deploying Linux at scale; by 2004, IDC acknowledged "clearly mainstream"

Collection Statistics

Metric Value
Queries investigated 7
Answered with high confidence 4 (Q001, Q002, Q004, Q006)
Answered with medium confidence 3 (Q003, Q005, Q007)
H3 (nuanced) supported 6 of 7 queries
H1 supported 1 of 7 queries (Q001)

Source Independence Assessment

The evidence base draws from several independent source categories: (1) analyst firm data (IDC, Gartner Dataquest) for market share numbers, (2) Netcraft surveys for web server data, (3) trade press coverage (Linux Journal, The Register, LWN.net) for event reporting, (4) corporate announcements (FINOS, Goldman Sachs, Morgan Stanley) for open source contribution timelines, and (5) Wikipedia/retrospective analysis for SCO litigation history. The primary independence concern is that market share data concentrates in IDC and Gartner — but this reflects the domain reality, not a search failure.

Collection Gaps

Gap Impact Mitigation
Archived Gartner/Forrester/IDC reports from 2001-2003 Cannot directly quote analyst predictions from the period; relying on secondary reporting Used LWN.net and press coverage that quoted analyst reports
Full text of Linux Journal #7730 Only summaries and extracted quotes available via web The key quotes and facts are captured; full context would add nuance
Enterprise-specific web server deployment data Cannot separate enterprise Apache usage from hosting provider usage Acknowledged the measurement methodology distinction as a key finding
Systematic documentation of SCO's chilling effect on employee speech Only one documented case (Phil Moore) of a named individual citing SCO as constraining speech Acknowledged this as a finding gap; systematic restrictions may have existed but were not documented
Exact quarter of Linux-Unix server revenue crossover Cannot pinpoint more precisely than 2011-2012 Sufficient for the query's needs; trajectory is clear

Collection Self-Audit

Domain Rating Notes
Eligibility criteria Pass Evidence types well-defined before searching; consistently applied across all 7 queries
Search comprehensiveness Some concerns Limited by availability of archived analyst reports from 2001-2004; many original sources are paywalled or no longer online
Evaluation consistency Pass Same scoring framework applied across all sources; same hypothesis structure (H1/H2/H3) for all queries
Synthesis fairness Pass H3 (nuanced) outcome in 6 of 7 queries reflects genuine complexity, not default to middle ground; each H3 is substantively different

Resources

Summary

Metric Value
Queries investigated 7
Files produced ~175
Sources scored 16
Evidence extracts 20
Results dispositioned 28 selected + 112 rejected = 140 total
Duration (wall clock) 20m 6s
Tool uses (total) 111

Tool Breakdown

Tool Uses Purpose
WebSearch 18 Search queries across all 7 queries
WebFetch 12 Page content retrieval for key sources
Write ~35 File creation for query outputs
Read 4 Reading methodology and output format specs
Edit 0 No file modifications needed
Bash ~15 Directory creation, batch file generation

Token Distribution

Category Tokens
Input (context) ~350,000
Output (generation) ~80,000
Total ~430,000