# results Published articles for results. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Running out of runway: Migrating Temporal Cloud's billing store to ClickHouse DevFeed: [Running out of runway: Migrating Temporal Cloud's billing store to ClickHouse]() Original publisher: [Read original article]() Author: Paul Oh Published: 2026-08-11T00:00:00Z Content type: article Language: en Sources: [Temporal Blog]() Topics: [clickhouse](), [Cloud](), [data](), [Databases]() Tags: [clickhouse](), [cloud](), [control-plane](), [cost](), [data](), [database](), [databases](), [parallel](), [results](), [table](), [temporal-voices]() ### AI overview Temporal Cloud migrated its billing store to ClickHouse after its previous warehouse had limited capacity. The migration provided three to five years of projected headroom, reduced steady-state costs by about 70%, and delivered usage data to customers roughly 30 minutes sooner. Temporal validated billing accuracy by running both databases in parallel for six weeks. ### Source excerpt How Temporal Cloud migrated its billing store to ClickHouse, gaining years of headroom, cutting steady-state costs by about 70%, and improving data freshness. ## Cost-effective orchestration with Fable 5 DevFeed: [Cost-effective orchestration with Fable 5]() Original publisher: [Read original article]() Author: Jeff Morhous Published: 2026-07-21T11:37:27Z Content type: tutorial Language: en Sources: [The AI-Augmented Engineer]() Topics: [Fable](), [Claude Code](), [Orchestration]() Tags: [better](), [claude-code](), [cost](), [effective](), [fable](), [how-to](), [results]() ### AI overview This tutorial presents Fable 5 in Claude Code as an orchestration approach and claims it can produce better results at lower cost. The supplied text does not provide implementation details or supporting measurements. ### Source excerpt Fable 5 in Claude Code is effective. I'll show you how to use Fable as an orchestration to get better results at a fraction of the cost ## Keycloak Community Survey Finds Demand for Official Helm Charts and More Robust GitOps Flows DevFeed: [Keycloak Community Survey Finds Demand for Official Helm Charts and More Robust GitOps Flows]() Original publisher: [Read original article]() Author: Keycloak Cloud Native team Published: 2026-03-18T00:00:00Z Content type: article Language: en Sources: [Keycloak Blog]() Topics: [Keycloak](), [Helm charts](), [argocd](), [GitOps](), [configuration]() Tags: [adoption](), [argocd](), [challenges](), [community](), [configuration](), [containerized](), [deployments](), [gitops](), [helm-charts](), [idm](), [kerberos](), [keycloak](), [ldap](), [openid-connect](), [results](), [saml](), [sso](), [survey]() ### AI overview Keycloak's survey of more than 360 respondents examined how the community deploys and manages Keycloak. Respondents expressed demand for official Helm Charts and more robust GitOps workflows, while adoption of the Keycloak Operator was reported as low because its setup is difficult to understand and its functionality is perceived as insufficient. ### Source excerpt Earlier this year, we conducted a survey to better understand how the community deploys and manages Keycloak in the real world. With over 360 respondents (!!), the results provide a better understanding of the community's needs and wishes. What we learned The survey results highlight a community that is deeply invested in the containerized ecosystem but is facing specific challenges in day-to-day operations. Key Findings: The Helm Charts: A significant portion of the community explicitly requested official Helm Charts, particularly for managing Keycloak installation and managing deployments via tools like ArgoCD. Operator: Adoption of the Keycloak Operator is not very high in the community, which is mainly caused by the fact that the Operator setup is hard to grasp, and do not provide the required functionality. GitOps & Configuration: There is a strong desire for more robust GitOps flows, especially when it comes to managing Keycloak resources like Realms and Clients. The whole Keycloak team is aware of these results and is discussing the next steps. Thank you to everyone who shared their feedback in this survey! More details: You can browse the detailed breakdown of the survey results in the slides below: ## word2vec-style vector arithmetic on docs embeddings DevFeed: [word2vec-style vector arithmetic on docs embeddings]() Original publisher: [Read original article]() Published: 2025-10-29T00:00:00Z Content type: article Language: en Sources: [technicalwriting.dev]() Topics: [Embeddings](), [Documentation](), [Angular](), [Supabase](), [Testing](), [Database]() Tags: [angular](), [database](), [embeddings](), [experiment](), [gemini](), [models](), [research](), [results](), [technical](), [vector]() ### AI overview This article investigates whether word2vec-style vector arithmetic works with document embeddings in technical writing. It describes experiments that subtract and add word or concept vectors to document vectors, using documentation from Supabase and the EmbeddingGemma model, with different task types and comparative verification. ### Source excerpt Does vector("king") - vector("man") + vector("woman") ~= vector("queen") work in technical writing contexts? ## Why Software Engineers Should Emphasize the Impact of Their Work DevFeed: [Why Software Engineers Should Emphasize the Impact of Their Work]() Original publisher: [Read original article]() Author: Tiger Abrodi Published: 2024-06-09T12:46:03Z Content type: opinion Language: en Sources: [Saiyan Growth Letter]() Topics: [Software](), [bus factor](), [lead time](), [Learning]() Tags: [bus-factor](), [lead-time](), [learning](), [promotion](), [results](), [software-engineer](), [team]() ### AI overview The article argues that software engineers should track and communicate the impact of their work, not only the tasks they completed. It uses organizing a weekly learning session as an example, linking that activity to a larger bus factor, shorter lead time, faster delivery, and happier customers. ### Source excerpt Input never matters. Outcome does. ## Dysfunctional without knowing: 5 signs that your team might have a problem DevFeed: [Dysfunctional without knowing: 5 signs that your team might have a problem]() Original publisher: [Read original article]() Author: tobi@techunicorn.builders (Tobias Mende) Published: 2023-02-25T06:00:00Z Content type: opinion Language: en Sources: [Tobias Mende]() Topics: [trust](), [function]() Tags: [accountability](), [collaboration-culture-psychological-safety-remote-work-technical-leadership-engineering-excelle](), [leadership](), [opinion](), [results](), [teams](), [trust]() ### AI overview The article describes five signs that a team may be dysfunctional without recognizing it: absence of trust, fear of conflict, lack of commitment, avoidance of accountability, and inattention to results. It explains how these problems build on one another and can be overlooked when a team appears harmonious or productive. ### Source excerpt Dysfunctional without knowing: 5 signs that your team might have a problem In the last 10+ years, I have worked in and with many teams. And I was fortunate to experience working in many fantastic teams. Teams where people cared for each other, trusted each other and were committed to deliver the best results together. People, where conflicts and disagreements were beneficial and important to achieve better results. ## Airflow Survey 2022 DevFeed: [Airflow Survey 2022]() Original publisher: [Read original article]() Author: Apache Airflow Published: 2022-06-17T00:00:00Z Content type: article Language: en Sources: [Apache Airflow Blog]() Topics: [airflow](), [data](), [Documentation](), [migration](), [Monitoring](), [Stack Overflow]() Tags: [airflow](), [apache-airflow](), [community](), [contributions](), [data](), [documentation](), [migration](), [monitoring](), [results](), [stack-overflow](), [survey](), [users]() ### AI overview The 2022 Airflow User Survey collected 210 responses over two weeks. Respondents were primarily data engineers, and Airflow was widely used in larger companies. Most participants were willing to recommend Apache Airflow, documentation and Stack Overflow were important information sources, and most deployments used Airflow 2. Migration from Airflow 1 was mainly constrained by resources and capacity. ### Source excerpt Airflow User Survey 2022 This year's survey has come and gone, and with it we've got a new batch of data for everyone! We collected 210 responses over two weeks. We continue to see growth in both contributions and downloads over the last two years, and expect that trend will continue through 2022. The raw response data will be made available here soon, in the meantime, feel free to email john.thomas@astronomer.io for a copy. TL;DR Overview of the user Like previous years, more than half of the Airflow users are Data Engineers (54%). Solutions Architects (13%), Developers (12%), DevOps (6%) and Data Scientists (4%) are also active Airflow users! There was a slight increase in the representation of Solutions Architect roles compared to results from 2020 and 2019 . Airflow is used and popular in bigger companies, 64% of Airflow users work for companies with 200+ employees which is an 11 percent increase compared to 2020. 62% of the survey participants have more than 6 Airflow users in their company. More Airflow users (65.9%) are willing to recommend Apache Airflow compared to the survey results in 2020 and 2019. There is a general positive trend in a willingness to recommend Airflow, 93% of surveyed Airflow users are willing to recommend Airflow ( 85.7% in 2019 and 92% in 2020 ), only 1% of users are not likely to recommend (3.6% in 2019 and 3.5% in 2020). Airflow documentation is a critical source of information, with more than 90% (15% increase compared to results from 2020) of survey participants using the documentation. Airflow documentation is also one of the top areas to improve! What's interesting, also Stack Overflow usage is critical, with about 60% users declaring to use it as a source of information (24% increase compared to results from 2020). Deployments 85% of the Airflow users have between 1 and 7 active Airflow instances. 62.5% of the Airflow users have between 11 and 250 DAGs in their largest Airflow instance. 75% of the surveyed Airflow users have be ## PinePhone community poll results DevFeed: [PinePhone community poll results]() Original publisher: [Read original article]() Published: 2022-01-31T00:00:00Z Content type: article Language: en Sources: [Community blog on PINE64]() Topics: [Pinephone](), [data]() Tags: [community](), [community-edition](), [geography](), [keyboard](), [pinephone](), [pinephone-pro](), [results]() ### AI overview Pine64 reviews results from a January poll about how people use their PinePhones. The poll received 3,079 self-reported responses, representing less than 5% of PinePhone owners, so the findings are not representative of the entire community. The article discusses respondent geography and broader trends among active community members. ### Source excerpt Throughout January we ran a poll asking people about how they use their PinePhones. Some poll results are quite predictable while others yielded unexpected results. Before we start, let me put on my academic hat for just one minute. This is a small, self-reported and not representative sample. A total of 3079 respondents took part in the poll - which amounts to less than 5% of all PinePhone owners. I'd therefore hesitate to extrapolate any of the results to the entire PinePhone community. I find it more likely that this sample reflects the most active community members and people following our project. It is also worth mentioning that some questions posed in the poll weren't exactly well phrased, and that the poll itself did not follow any established conventions. But this was never meant to be a comprehensive study - it is instead, at best, a peak at the overarching trends of the most active portion of the community. To this end, when describing the results I'll use descriptive generalizations, such as 'more than half' or 'the majority', rather than exact numbers or percentages I feel that this is more in line with the nature of the data and the general spirit of the poll. For what it's worth, I'll also share some of my thoughts and insights as we review the results. Exact numbers and percentages are, however, included in the graphs. As always, I encourage you to share your thoughts about the results in the comments section. Thank you to all those who took the time to fill in the questionnaire! Geography A third of all respondents are from the United States. In Europe, Germany, the United Kingdom (UK) and France have the highest respondent counts. The number is by far the greatest in Germany - more than twice that of the United Kingdom. In terms of per capita, it would appear that we also have a large active community in the Netherlands, Sweden, Austria and Czechia. Adjusted for population size, the response counts in these countries are actually on par or larger t ## Building a new snapshot fuzzer & fuzzing IDA DevFeed: [Building a new snapshot fuzzer & fuzzing IDA]() Original publisher: [Read original article]() Author: Axel "0vercl0k" Souchet Published: 2021-07-15T15:00:00Z Content type: article Language: en Sources: [Diary of a reverse-engineer]() Topics: [Fuzzing/Fuzz testing](), [C++](), [Windows](), [App](), [Emulator](), [test-coverage](), [bug](), [cross-platform]() Tags: [bochs](), [bochscpu](), [bug-bounty](), [c-plus-plus](), [code-coverage](), [cross-platform](), [emulator](), [fuzzing](), [ida](), [kvm](), [misc](), [results](), [snapshot-fuzzing](), [whv](), [winhv]() ### AI overview This article describes the development of a Windows snapshot-based fuzzer designed to target user-mode and kernel applications and services. It explains the planned use of hypervisor and KVM APIs, instrumentation, code coverage, crash detection, mutation, cross-platform support, and fast restoration, with IDA selected as a development target. ### Source excerpt Introduction It is January 2020 and it is this time of the year where I try to set goals for myself. I had just come back from spending Christmas with my family in France and felt fairly recharged. It always is an exciting time for me to think and plan ... ## JetBrains' first Kotlin Multiplatform survey | #TLTouchPoints - Touchlab DevFeed: [JetBrains' first Kotlin Multiplatform survey | #TLTouchPoints - Touchlab]() Original publisher: [Read original article]() Published: 2021-02-24T18:51:11Z Content type: article Language: en Sources: [Touchlab | Enterprise Mobile Innovation & Development]() Topics: [Kotlin Multiplatform](), [jetbrains]() Tags: [adopting](), [insights](), [kotlin](), [kotlin-multiplatform](), [management](), [results](), [survey](), [touchlabshare]() ### AI overview A Touchlab #TLTouchPoints episode summarizes key insights from JetBrains' first Kotlin Multiplatform survey and discusses implications for management considering KMP. ### Source excerpt JetBrains recently released the first ever Kotlin Multiplatform survey. In this episode of #TLTouchPoints Justin summarizes the key insights from the results and implications for management considering KMP. ## Porting Syzkaller to ReactOS (Evaluation-1) - GSoC 2020 DevFeed: [Porting Syzkaller to ReactOS (Evaluation-1) - GSoC 2020]() Original publisher: [Read original article]() Published: 2020-06-30T00:00:00Z Content type: article Language: en Sources: [Front Page on ReactOS Website]() Topics: [ReactOS](), [Fuzzing/Fuzz testing](), [boot](), [Google](), [coding]() Tags: [boot](), [coding](), [crashes](), [free](), [fuzzing](), [google](), [logging](), [open-source](), [os](), [react](), [reactos](), [results](), [ssh](), [win32](), [winapi]() ### AI overview A GSoC 2020 progress report describes porting Syzkaller to ReactOS. The author reports booting ReactOS with UBSan enabled, getting Syz-stress and Syz-executor working, adding initial kernel32 function definitions, and fuzzing ReactOS with about 107 definitions, which produced approximately four crashes. Planned work includes SSH access, rebasing, shared-memory and KCOV support, and automated virtual-machine fuzzing. ### Source excerpt Hey, I am Suraj K Suresh (@Freakston), In this post I'm going to be listing out what I've completed in Coding period-1 of GSoC. Status The initial part of setting up Syzkaller and being able to fuzz ReactOS is done. The link to grab ReactOS compatible syzkaller is as follows: Github. The detailed instruction on setting it up can be found here. Work-done Boot ReactOS with UBSan enabled. The branch which has UBSan enabled can be found in extravert34's fork of ReactOS. ## Threat Modeling Building Blocks DevFeed: [Threat Modeling Building Blocks]() Original publisher: [Read original article]() Author: Adam Published: 2019-09-04T00:00:00Z Content type: article Language: en Sources: [Shostack & Friends Blog]() Topics: [data](), [Requirements]() Tags: [building](), [comparison](), [data-protection](), [models](), [requirements](), [results](), [reuse](), [tasks](), [techniques](), [things]() ### AI overview The article presents threat modeling as a collection of interconnected tasks and discusses research on breaking threat modeling into smaller, assessable building blocks. It highlights comparisons involving privacy and data protection, requirements, STRIDE, CAPEC, CWE, OWASP, data-flow-diagram variants, and CARiSMA extended. ### Source excerpt Threat modeling isn't one task -- its a collection of tasks that build on each other to produce more valuable insights. ## OKRs aren't going to fix your communication issues DevFeed: [OKRs aren't going to fix your communication issues]() Original publisher: [Read original article]() Author: Map Published: 2019-03-30T20:55:56Z Content type: opinion Language: en Sources: [Craig Kerstiens]() Topics: [meetings]() Tags: [behavior](), [change](), [communication](), [distributed](), [early-stage](), [founders](), [management](), [meetings](), [notes](), [okrs](), [philosophies](), [priority](), [recap](), [results](), [team]() ### AI overview The article argues that OKRs do not by themselves resolve startup communication problems. Teams must first identify the problem they are trying to solve and explicitly communicate agreed goals, especially as organizations grow and people miss meetings. ### Source excerpt Talking with a startup a few days ago they asked for my opinions on OKRs. I have slightly mixed opinions on them overall and started to disclose some of those. Though in sharing some of this I had a few immediate realizations that might be broadly applicable. The crux of his question was, at what stage should we put them in place. I've seen a few companies try to put in some form of OKR, and most were met with pretty mixed results. The reason is that OKRs need to change something about your behavior otherwise why put them in place... either change something about the goals you would otherwise have or the methods at which you went about achieving them. Stepping back a bit, my first question and a very focusing question on almost any situation to ask is "What problem are we trying to solve?" In our conversation he actually paused a bit. As he paused a bit longer it was clear that question had not been fully asked or answered. The first and most common case I see with startups trying to put in place OKRs, v2moms, management by objectives is that the team is not aligned and focused on the same goals. But my follow-on question is consistently, have you communicated what you decided you goals were. Startups tend to go through some distinct growing phases. The early stages all the founders are in a room together building out the product. When you get the first few engineers you expand out a little, but still in a single co-working conference room easily. Eventually you need a real office. At the real office stage you start to have an all hands where, this is probably gathered around a large lunch table at first. At all hands no one takes meeting minutes and sends out a recap, instead people take some notes and you assume everyone was present. But, at about 20 people you have at least one person that misses the weekly team meeting and misses something key. In a 1:1 you catch it that it was talked about as a priority... but they weren't there. This very subtle change I've seen l ## Post-Mortem: Applied Image Optimization DevFeed: [Post-Mortem: Applied Image Optimization]() Original publisher: [Read original article]() Author: andrew@nystudio107.com (Andrew Welch) Published: 2016-12-07T18:11:00Z Content type: article Language: en Sources: [nystudio107 | Articles on modern web development.]() Topics: [Optimization](), [Image](), [Website](), [benchmarking]() Tags: [applying](), [benchmarking](), [image](), [loading](), [optimization](), [performance](), [post-mortem](), [post-mortems](), [project](), [results](), [shows](), [speed](), [strategies](), [website]() ### AI overview This project post-mortem describes applying image optimization techniques to improve website loading performance. Lazy-loading below-the-fold images, responsive server-side image processing, progressive JPEGs, and HTTP/2 reduced load time from 107.8 seconds to 2.8 seconds, Start Render from 29.8 seconds to 1.4 seconds, and Visually Complete from 35.5 seconds to 1.7 seconds. ### Source excerpt This project post-mortem shows the results of applying image optimization strategies to speed up the loading of a website ## Ethical Obligations in Internet Operations - Survey Results DevFeed: [Ethical Obligations in Internet Operations - Survey Results]() Original publisher: [Read original article]() Published: 2015-10-12T17:28:55Z Content type: article Language: en Sources: [Signs of Triviality]() Topics: [Internet](), [context]() Tags: [context](), [internet](), [operations](), [results](), [survey](), [talk](), [velocity]() ### AI overview Survey results on ethical obligations in Internet operations, presented in the context of a talk at Velocity NY 2015. ### Source excerpt Results from my survey on Ethical Obligations in Internet Operations in the context of my talk at Velocity NY 2015 ## Go: Slice search vs map lookup DevFeed: [Go: Slice search vs map lookup]() Original publisher: [Read original article]() Author: Graham King Published: 2015-06-30T06:09:24Z Content type: article Language: en Sources: [Graham King]() Topics: [Go Language](), [Data structures](), [Benchmark]() Tags: [array](), [benchmark](), [code](), [data-structure](), [go](), [performance](), [results](), [slices](), [software]() ### AI overview A comparison of Go map lookups and slice searches finds that slices can be faster for very small collections: up to about five items for key-value lookups and ten items for integer-set lookups. Maps are generally the straightforward choice beyond those break-even points. ### Source excerpt Maps vs. slices: Choose wisely for optimal performance. ## When flying from Oslo to London, you don't want to go via Moscow DevFeed: [When flying from Oslo to London, you don't want to go via Moscow]() Original publisher: [Read original article]() Author: Per Jørgen Walstrøm Published: 2015-06-15T12:47:32Z Content type: article Language: en Sources: [Finn.no]() Topics: [Flight](), [Query (disambiguation)]() Tags: [query](), [results](), [search](), [travel]() ### AI overview The article explains how FINN Flight Search addresses implausibly long itineraries, such as routes from Oslo to London with distant layovers. It reviews a simple duration threshold, discusses using quartiles and the upper-fence rule to identify outliers, and notes that Solr requires a separate query to calculate quartiles. ### Source excerpt If you want to travel from Oslo to London, you may end up getting results which include layovers in far-away places. Do you really want to spend a night at the airport in Moscow if you only want to go from Oslo to London? Probably not, but nonetheless, some of the results we receive at the FINN Flight Search are those kind of travels. Those travels may be really cheap, ending up at the top of our result list as we sort by price, but they are mostly garbage and we do not want our users to be confused by those outliers. We want them to go away! The old solution We have tried to fix this before. With limited success. A solution we used a couple of years back was something like this: x = 3 * Average of the three shortest flights We removed results with duration > x. This works fine on short travels, like a two-hour flight from Oslo to London. With the given formula, all trips longer than 6 hours to London will be removed. That's ok. However, it doesn't work too well on longer flights. If you want to go from Oslo to Bangkok, you would probably spend at least 11 hours in the air. In that case, it would be too conservative to just remove the flights that are 33 hours or more. A theoretical solution The correct way to remove the outliers is by finding the quartiles. When the quartiles are found, the upper fence is defined by Upper fence = Q1 + 1.5 * IQR Where IQR is the Interquartile Range; Q3 - Q1. We are using Solr at FINN, but the stats query in Solr does not give us the quartiles out of the box, so we would need to do a separate query to calculate the quartiles. Should be easy enough. What is too long? So what is a too long flight? That's not an easy question to answer. Racking our brains (and doing some guesswork), we came up with the following suggestions on what the upper fence on duration to different destinations should be. The durations are per leg: OSL - LON 300 mins = 5 hrs. Given the minimum flight time OSL - LON = 115 OSL - BKK (Bangkok) 1000 mins = 17 hrs. Mi ## The Odds of Finding a Set in The Card Game SET DevFeed: [The Odds of Finding a Set in The Card Game SET]() Original publisher: [Read original article]() Published: 2010-04-28T00:00:00Z Content type: article Language: en Sources: [Peter Norvig]() Topics: [math](), [Simulation]() Tags: [card](), [color](), [game](), [number](), [results](), [set](), [simulation]() ### AI overview An analysis of why the odds of finding no SET in the card game SET change during play. A simulation shows that the chances differ substantially between a freshly dealt layout and layouts reached after SETs are repeatedly removed and replaced, suggesting the instruction booklet's odds do not describe the course of an actual game. ### Source excerpt Correcting an error in the instruction booklet for the card game SET ## Soft of the Year Award DevFeed: [Soft of the Year Award]() Original publisher: [Read original article]() Published: 2008-12-19T00:00:00Z Content type: news Language: en Sources: [Front Page on ReactOS Website]() Topics: [ReactOS](), [Software]() Tags: [free](), [mail](), [open-source](), [os](), [react](), [reactos](), [results](), [software](), [win32](), [winapi]() ### AI overview ReactOS won second place in the Technologies category of the 2008 Soft of the Year national Russian software award, based on people's voting conducted by Mail.Ru. ### Source excerpt "Soft of the Year 2008" Award "Soft of the Year 2008" is a national Russian software award given annually to the best computer software. Winners in seven different categories are selected by results of people's voting (carried out by Mail.Ru). ReactOS won second place in the "Technologies" category. Thanks to everyone who supported us! ## Warning Signs in Experimental Design and Interpretation DevFeed: [Warning Signs in Experimental Design and Interpretation]() Original publisher: [Read original article]() Published: 2007-07-09T00:00:00Z Content type: opinion Language: en Sources: [Peter Norvig]() Topics: [test]() Tags: [essay](), [experiment](), [results]() ### AI overview This essay explains why statistically significant experimental results do not necessarily establish that a treatment will work for an individual. It identifies warning signs in experimental design and interpretation that can reduce a study's credibility, while noting that one warning sign does not invalidate a study. ### Source excerpt An update to my essay on experiment design; it now stresses warning signs rather than clear-cut mistakes and urges the reader to consider the factors that dilute the credibility of an experiment. ## Evaluating Extraordinary Claims: Mind Over Matter? Or Mind Over Mind? DevFeed: [Evaluating Extraordinary Claims: Mind Over Matter? Or Mind Over Mind?]() Original publisher: [Read original article]() Published: 2007-07-09T00:00:00Z Content type: opinion Language: en Sources: [Peter Norvig]() Topics: [Simulation](), [test](), [Software]() Tags: [analysis](), [results](), [simulation]() ### AI overview An essay evaluates claims about the efficacy of intercessory prayer by examining experimental design and simulating fMRI results. It reports that random simulations found no significant differences, while simulations incorporating short-term participant anxiety sometimes produced significant results, illustrating how small samples and confounding factors can affect findings. ### Source excerpt A companion to my essay on experiment design, this essay uses as an example the literature on the efficacy of intercessory prayer on medical conditions.