Don't trust the claim.
Test it.

But Is It True? Lab tests claims about technology, software, products, services and more. We document our methods, preserve the evidence and publish what actually happened.

What Is But Is It True? Lab?

Companies, products, software, books and people make claims every day. Some are accurate. Some need context. Some sound impressive until the numbers are examined.

Instead of guessing, we test the claim.

Show the Claim

Start with what was actually said rather than changing the claim into something easier to prove or disprove.

Test It

Define a repeatable test, record the conditions and measure what actually happens.

Show the Evidence

Publish enough methodology and evidence that conclusions can be challenged, reproduced or corrected.

Our Current Projects

We currently have two active technology research projects. One examines programming-language and database performance. The other tests whether we can build and operate a practical, independently hosted web search engine using open-source software.

PROJECT 01 — TESTING

Programming Languages & Databases

Are programming languages that are commonly described as faster actually faster when real applications retrieve information from databases?

We're testing multiple languages against different database systems under controlled workloads, measuring throughput, response times, errors and stability.

  • The first test round has been completed.
  • A second round of testing has been performed.
  • Haskell has been discontinued from future rounds.
  • C# has been added for the final three database comparisons.
  • Additional investigations include performance differences at higher concurrency.
View Benchmark Results
PROJECT 02 — WORKING PROTOTYPE

MyGlobalWebsite Search Engine

Can we build a functioning web search engine using free, open-source technology, without depending on another commercial search engine's results?

Our experimental search engine is now accessible at myglobalwebsite.com. Its initial implementation uses Apache Nutch to crawl websites and Apache Solr to index and search their contents.

  • Independent website crawling and parsing.
  • Solr full-text indexing and relevance scoring.
  • Custom HTML, CSS and JavaScript search frontend.
  • Nginx and HTTPS on the development server.
  • First successful test: 19 indexed documents from four websites.
Try MyGlobalWebsite

Both projects are experimental. Current results demonstrate what has worked in our testing environment; they are not claims of production-scale performance.

What We're Building

But Is It True? Lab is more than a place to publish individual tests.

Our long-term goal is to build a directory, search, testing and research platform designed to help people examine claims made by businesses, websites, products and services.

Instead of simply showing what someone says, we want to help answer the next question:

But is it true?

As the platform develops, we want to connect information with claims, supporting evidence, our own testing where appropriate, and eventually community and business responses.

MyGlobalWebsite is one part of that research. It allows us to study how web crawling, indexing, ranking and retrieval actually work by operating our own search engine.

The programming-language and database benchmarks are another part. They help us evaluate technologies before choosing how to implement larger platform features.

Project Update: Programming Language Benchmarks

We began by comparing implementations of the same database-backed API in different programming languages.

Each implementation must perform comparable work and return the correct data. We record requests per second, response times, errors and the concurrency levels at which problems develop.

Second Test Round

We have now conducted our second round of testing. This gives us additional data to compare with the original results and helps identify behavior that was not visible in the first round.

We are particularly interested in whether apparent differences are caused by the language, the database, the client libraries, the implementation or the benchmark environment.

Changes to the Language Comparison

Haskell has been discontinued from the remaining benchmark rounds. It remains part of the historical record of the project. We are not removing its earlier work or results.

C# has been added to the comparison for the final three databases. This is an addition to the remaining testing plan, not a claim that those comparisons are already complete.

The active comparison has therefore changed from the original thirteen-language lineup. Historical results will continue to identify the actual languages included in each test.

Databases Under Investigation

Our testing has included Apache Cassandra, PostgreSQL and MariaDB, with additional database comparisons and compatibility testing planned or underway.

We also investigated CockroachDB and YugabyteDB. CockroachDB's licensing restrictions affected our high-concurrency preflight experiment, while the YugabyteDB setup produced problems that require further investigation before a fair comparison.

We retain these findings because failed or discontinued experiments are part of the evidence.

Read the Actual Results

We publish benchmark measurements rather than selecting a favorite language based on reputation. Individual results should always be interpreted together with the workload, database, configuration and error counts.

View Benchmark Test

Why This Matters

We don't expect a single programming language or database to be the best choice for every component. A high-throughput search API may need a different architecture from a reporting interface, data-processing service or crawler.

We are willing to use multiple languages and database systems when the evidence supports it. Performance, reliability, resource consumption, development complexity and maintainability will all matter.

Project Update: MyGlobalWebsite

We now have a working web search prototype that crawls websites, indexes their content and returns keyword searches through a public website.

Visit: https://myglobalwebsite.com

What We're Using

Apache Nutch

Crawls websites, discovers links, checks robots.txt rules and parses content for indexing.

Apache Solr

Stores the indexed documents and performs full-text searches with relevance scoring.

Nginx & Frontend

Serves our custom search page and connects search queries to Solr without exposing its administration interface.

Our First Successful Search

We began with four websites we control. The Nutch-to-Solr pipeline successfully indexed 19 documents.

A search for inspection returned four matching documents. Solr reported 55 milliseconds of server-side query processing for that request.

This is proof that the pipeline works with a tiny index. It is not enough evidence to establish performance with millions or billions of documents.

What We're Testing Next

Before attempting large-scale crawling, we are building and refining the public search interface. A later trial may expand the index to approximately 10,000 pages to examine relevance, stability, indexing behavior and resource usage.

We also intend to investigate duplicate URLs, image indexing, crawling rules and how easily the architecture can grow across servers.

We're Testing This Website Too

But Is It True? Lab isn't beginning with a finished platform. We're documenting the process of building it.

That includes what we started with, what we changed, why we changed it, what worked and what didn't.

Starting Environment

Development Computer

Our documented initial development computer is a Mac mini with an Apple M1 processor and 8 GB of RAM.

Code Editor

We use Visual Studio Code for development, including remote editing of files on servers.

AI Development Tool

We use ChatGPT with a paid account for architecture discussions, code generation, troubleshooting, documentation and analysis.

AI-generated suggestions are not automatically treated as correct. Code and configurations must work in the actual environment.

Internet Connection

Our office uses T-Mobile Business Internet. The connection is part of the real working environment, not a dedicated data-center link.

Public But Is It True? Lab Server

The original public website was deployed on a multi-purpose SSD Nodes server.

  • Operating system: Ubuntu 24.04
  • CPU: 12 vCPUs
  • Memory: 64 GB RAM
  • Storage: 1,200 GB NVMe

This public server hosts other websites and services. It uses Docker containers with Nginx handling external web routing.

MyGlobalWebsite Development Server

The search engine is currently being developed on a separate multi-purpose server with 96 GB RAM. Nutch and Solr run in Docker containers, while Nginx runs directly on Ubuntu.

Our Initial Website Build

1. Register the Domain

We registered butisittruelab.com, configured DNS and verified the required web and email records.

2. Create an Isolated Web Service

We placed the original public frontend in its own Docker container running Nginx.

3. Prove the Site Works Locally

We tested the web service on the server before connecting the public domain.

4. Connect the Domain

Nginx was configured to route requests to the frontend. We tested the configuration before reloading it.

5. Enable HTTPS

We installed a TLS certificate so the public website could use HTTPS.

6. Build the First Public Page

The first version was deliberately built with HTML, CSS and JavaScript, without pretending incomplete interactive features already worked.

7. Test the Technology

We began programming-language and database benchmarks to help make future architecture decisions based on measured results.

8. Launch the Search Experiment

We created the MyGlobalWebsite prototype, installed Nutch and Solr, configured crawling and indexing, and built a public search page accessible over HTTPS.

Hosting Disclosure

We have used SSD Nodes for other websites and services for approximately three years. In our documented experience, two server connection problems were resolved by rebooting. That is our experience, not a guarantee of anyone else's results.

This type of hosting requires managing operating systems, security, updates, backups and services yourself. It may not be appropriate for users without server administration experience.

Affiliate Disclosure

We have an affiliate relationship with SSD Nodes. If someone signs up through our referral link, we may receive compensation.

We were using SSD Nodes before starting But Is It True? Lab. The affiliate relationship does not purchase a favorable rating or conclusion.

SSD Nodes Referral Link

Future reviews of SSD Nodes will also disclose the affiliate relationship.