Question 292 nodes
Can shared dictionaries help JSON APIs?Applying HTTP shared dictionaries to REST, GraphQL, and MCP.
Kamal with Postgres backups, replicas and self-hosted monitoring
Can Kamal handle Postgres backups/restores/replicas, and what to do for self-hosted logs and monitoring?
Quickest way to build a Bluesky feed?
Q&A: easiest path to shipping a custom Bluesky feed.
Should fake implementations live with production code?
Q&A on where to keep reusable test-only objects like a fake dice or fake time.
Are Sorbet runtime checks built on TracePoint?
How Sorbet implements runtime type checks and whether TracePoint explains its overhead.
Can parts of Eventide, Rails Event Store and Trailblazer be combined in one project?
Audience question on interoperability across the three ecosystems.
Can Trailblazer be combined with event sourcing and CQRS?
Audience question on combining Trailblazer with event sourcing and CQRS.
How does Phlex integrate with Turbo?
Audience asks how Phlex components interact with Turbo and the new Hotwire stack.
How does Ruby UI integrate with CSS and JavaScript?
Audience asks whether Ruby UI depends on Tailwind, Stimulus, or the Rails asset pipeline.
How to do DDD remotely?
Remote collaborative modeling when teams are distributed.
Is Ruby dying?
Audience question about the perception of Ruby decline.
ActiveRecord encryption vs data obfuscation for PII
Q&A on how encryption and obfuscation fit together for PII.
Adding imageflow to image_processing gem
Will imageflow (security-first) be supported as a backend?
Authorization strategies in GraphQL
How to handle authorization for flexible GraphQL queries?
Avoiding model duplication across engines
Q&A on sharing models between buyer and seller engines.
Business Motivation for UringMachine
What business need drove building UringMachine?
Can Ractors be fire-and-forget for background GUI work?
Q&A on whether Ractors can run detached while a GUI shows progress.
Can service objects and command handlers call each other?
Q&A on inter-service calls and dependency injection in layered DDD.
Can you reliably run Rails on JRuby?
Is Rails a viable target for JRuby in production?
Cross-tenant aggregation with row-level multi-tenancy
Which partitioning approach best supports staff/admin aggregated computations across tenants?
Defining scopes with relationship-based access control
Q&A: equivalent of Pundit scopes in ReBAC-based authorization.
Denormalization vs graphql-batch / graphql-preload
Did you compare denormalizing fields versus using batch/preload gems?
Difference between JRuby and TruffleRuby
How does JRuby differ from TruffleRuby on GraalVM in layman's terms?
Difference between Literal and dry-types
How does Literal differ from dry-types?
Distributed or majestic monolith?
Audience question on whether enterprise projects must be distributed.
Do events store current state, and how does replay perform?
Q&A on whether events carry latest state and on projection performance.
Do fixtures beat factories?
Q&A on whether YAML fixtures are faster than FactoryBot for tests.
Do you end up fighting Turbo when adopting these new APIs?
Q&A: interop between Hotwire/Turbo (including broadcasts) and the new web-platform APIs.
Do you unit test ViewComponents?
Audience question on whether Stankov heavily unit-tests ViewComponents.
Do You Use Falcon With Your Framework?
Does Sharon use Falcon as the web server for his framework?
Dry-types input validation vs full Sorbet typing
Why add Sorbet on top of dry-types/unit tests if input is already validated at the boundary?
Experience with mutation testing in other ecosystems?
Audience asks if Schirp has used mutation testing outside Ruby.
Experience with React comparison question
Audience question comparing ViewComponent with React and asking for Stankov's feelings.
Fault-tolerant token-by-token parsing for structured output
Can a parser validate each generated token live and retry invalid ones?
Handling third-party APIs without idempotency keys
How to guarantee single creation when neither idempotency keys nor efficient find-or-create exist?
How are fix and change different from remove and add?
Audience asks whether fix/change are just remove/add under another name.
How does MQTT compare with WebSockets?
Bidirectional MQTT vs WebSockets — are they interchangeable?
How do you handle backups with SQLite?
How does the team back up its SQLite production database?
How Kamal delivers Docker images to servers
What registry does Kamal use and how does the image reach the server?
How to handle very large event volumes over time?
Q&A on storing millions of events in write-heavy systems and whether to compact or snapshot.
How to integrate Mutant with CI
Audience asks how Schirp wires Mutant into continuous integration.
How to migrate a 5-year-old Rails+React app toward no-build?
Q&A: migration path from an established SPA to the no-build world.
How to model sensitive tasks in FGA?
Q&A: representing a sensitive-task override in the authorization schema.
In-memory filtering without a database
Did filters run in-memory or as SQL, given the large product catalog?
Instrumentation as events, independent of output
Should the codebase emit events independent of log/trace/metric destination?
Is a decider a state machine over many inputs?
Audience asks whether deciders combine state of several aggregates (campaign + donation).
Is starting with SQLite a lock-in if I later need Postgres?
Can a SQLite-on-Rails app migrate to Postgres once it gets big?
Is time travel the only way to test Time.current?
Challenge to Rodzik's claim that time-travel helpers are the only way to test current-time code.
JRuby deployment on Heroku and Railway
How hard is it to deploy JRuby apps on Heroku/Railway-style platforms?
JSON API or GraphQL — what do you recommend?
Panel question on standardized API formats and when to follow them.
Kamal Proxy multi-region load balancing roadmap
Where does Kamal Proxy stand on multi-region load balancing?
Performance of Faker on large databases during obfuscation
Q&A on slow per-row Faker generation over big tables.
Phrases to avoid in code review
Q&A on wording code-review feedback more professionally.
Reducing direct vs indirect communication friction
How to bridge direct and indirect communication styles across cultures.
Ruby developer at Google says gRPC is no longer used there
Is gRPC still used internally at Google?
Scaling receivers vs implementing idempotency
Does scaling receivers under race-condition load make idempotency harder to implement?
Scheduling granularity vs HYA
Does Geneva Drive need its own scheduler like HYA had?
Schemas for structured log attributes
How to reconcile structured logging with per-app attribute variability.
Should enterprise libraries be paid or free?
Audience question on monetization of enterprise-oriented OSS libraries.
Should OSS developers learn sales and marketing instead?
Audience comment: developers reject sales/marketing as a profession; Sidekiq-style monetization is possible.
Should the tooling or language change to support these workflows?
Audience question on whether a functional language would fit Trailblazer's style better.
Using external auth providers instead of Devise
Q&A: what about services like Auth0 instead of Devise?
Was debugging dev-environment performance worth it vs. ditching Spring/Guard?
Q&A: was the long debugging investment justified instead of dropping Spring/Guard?
Were the dynamic HTTP overrides thread-safe?
Were dynamic monkey-patches to HTTP clients safe under the app's concurrency model?
What is the Ractor implementation modeled on?
Q&A on the inspiration behind Ruby's Ractor model.
What will be the JavaScript developer's dream in a year?
Panel question about upcoming frameworks and concepts to get excited about.
What would you pick for a greenfield today?
Language choice when economic constraints are in play.
Which JavaScript fads have stood the test of time?
Panel question on long-lived JavaScript frameworks and patterns.
Why does Ruby UI avoid slots?
Audience asks why Ruby UI uses many atomic components instead of ViewComponent-style slots.
Would you recommend TypeScript or Flow?
Audience question on using typed JavaScript supersets.
Application-Level Back Pressure For Sidekiq
Can the application throttle event generation based on Sidekiq load?
Are fake-dice tests enough, or should the real dice be tested too?
Q&A on whether unit tests with fake dice need complementary end-to-end tests.
Are God objects a consequence of inheritance?
Would forbidding inheritance prevent God objects from forming?
Aren't query optimizations a better first step than caching?
Asks whether N+1 fixes/eager loading should come before caching.
Are there practical limits to the number of Ractors?
Q&A on how many Ractors can usefully run concurrently.
Are you using Crystal Ball?
Audience introduces Crystal Ball as a way to run only tests affected by changed files.
Asynchronous EventStorming Experience
Has the speaker tried asynchronous (non-overlapping) EventStorming sessions?
Book recommendation for cross-cultural communication
What resource helps with cultural differences in technical communication?
Can Grazer reference other fields for derived values like encrypted passwords?
Q&A on field-to-field references in Grazer config.
Can Mutant guide refactoring?
Audience asks whether Mutant's output can identify code to refactor.
Can Mutant list weak-spot subjects to focus on?
Audience asks whether Mutant can output a ranked list of worst-covered subjects.
Can Opal be treated as a real production technology?
Panel question on Opal's production readiness and future.
Can production observation auto-generate types?
Could tools like Coverband observe production and generate types automatically?
Can we teach the community not to shoot themselves with factories?
Q&A: should FactoryBot docs warn about association performance?
Can you do DDD without event sourcing?
Q&A on whether event sourcing is mandatory for DDD.
Centralised vs per-service event store?
Q&A on whether each microservice should own a separate event store.
Conditional DAG jumps in Geneva Drive
Should the declarative DSL allow conditional 'go to step X else step Y' jumps?
Convincing Developers to Participate in EventStorming
How to get developers to engage and be vocal in modeling workshops.
Cost control for structured logging
How to manage observability costs; rules of thumb for business-logic logging.
Could types be defined outside the class like RBS?
Have you explored RBS-style external type files for Literal?
Cross-region business reporting aggregation
How to aggregate business reports across multiple regions without violating data residency.
DDoS risk of on-the-fly processing
How to defend on-the-fly processing from attackers requesting many variants.
Deploying to a separate development server IP
How to tell Kamal to deploy to a different IP for a dev/staging machine?
Diagnostic Substitutes Are Not Mock Objects
Audience Q&A clarifying the difference between diagnostic substitutes and mocks/spies.
Distinguishing anxiety from frustration on one-on-ones
Q&A on how to identify elusive emotions when using the Knoster matrix.
Do all Red Hat developers get a Red Hat?
Whether Red Hat gives every new hire a red hat.
Does any of these techniques see real project use?
Audience asks whether she uses these tricks in production.
Does caching still help for complex apps with many models?
Asks whether caching scales beyond simple apps with few models.
Does DDD apply to startups?
Whether DDD fits startups where the business is still being discovered.
Does DDD require object-oriented programming?
Whether DDD is tied to or intersects with OOP.
Does experienced architect's intuition override precise design vocabulary?
Whether experience/intuition can replace explicit design-quality description.
Does Literal::Data raise in production on a bad prop?
Does Literal throw an exception when passed a wrong prop type in production?
Does per-field validation blow up UI complexity?
Q&A on whether accepting invalid events and validating per-field increases UI complexity.
Does randomizing structural fields break analytics?
Q&A on whether obfuscation-induced structural changes distort downstream analytics.
Does SXG break on its own due to Google changes?
Does SXG need re-measurement even without deployments because Google silently changes things?
Does the base language affect how hard automation work is?
Does a stronger base language reduce how much automation you have to build?
Does Turbo Native require separate iOS and Android code bases?
Audience question confirming hybrid still needs some native code per platform.
Do Rails adopters write handlers instead of controller actions?
Q&A on how Rails teams integrate Eventide into their MVC.
Enforcing boundaries with Packwerk given Rails dynamic calls
How to prevent cross-domain Ruby object calls that Packwerk's static analysis misses?
Engines vs namespaces for splitting a growing monolith
Q&A on when to use a Rails engine instead of a namespace for a new app inside a monolith.
EXPLAIN cost units meaning
How should the numbers in Postgres EXPLAIN cost estimates be interpreted?
GDPR and losing data intentionally
How to comply with GDPR when events are immutable?
Generate code once vs instruct LLM per request
Trade-offs between generating Ruby code once vs sending instructions to the LLM on every request.
Geneva Drive hero constraints and migrations
Do heroes need special mutations, and does the gem generate its own migrations?
Geneva Drive hero mutation and deletion
What happens if the hero record is externally mutated or deleted mid-workflow?
Has Mutant been run against popular Ruby libraries?
Audience asks whether Schirp mutation-tests popular open-source gems.
Have you considered going back to REST?
Would you drop GraphQL and return to REST APIs?
Have you heard of Stimulus?
Audience question about DHH's Stimulus JavaScript library.
Have you reported these issues upstream to Devise?
Q&A: did you open an issue on the Devise repo?
Hosting multiple pet projects on one droplet with Kamal
Can you deploy many small projects to the same Digital Ocean droplet with Kamal?
How did you migrate existing data and what was the downtime?
Process and downtime for the actual data migration.
How does error handling work?
Audience asks how errors integrate with Ruby exceptions and the UI.
How does GDPR deletion work if encrypted events are in backups?
Q&A on handling GDPR erasure when event backups still contain encrypted PII.
How does hybrid compare to React Native after a decade?
Audience question on migrating from React Native back to hybrid.
How does naming an abstraction relate to OOP design?
Audience Q relating the 'frequencies' refactor to OO-style class design.
How does tea's token scheme monetize?
Given tea's talk of blockchain/web3/NFTs and tokens, how do tokens translate to real money?
How do pure front-end developers react to Ruby views?
Audience asks how HTML/CSS/JS developers handle UIs written in Ruby.
How do routing trees find the right controller at request time?
Attendee asks about the per-request cost tradeoff of Yippee's routing-tree approach.
How do you collect information about how people monkey-patch Spree?
How does the Spree team find out what to refactor into dependency-injection points?
How do you deal with event versioning?
Q&A on changing event schemas after events have been written.
How do you determine call-graph arrow direction?
Audience Q on reading direction of arrows between view/model/controller layers.
How do you handle boilerplate of commands/handlers/events — scaffolding generators?
Audience question about boilerplate growth and code generation.
How do you order events across distributed systems?
Q&A on event ordering without relying on wall-clock time.
How do you prevent race conditions?
Audience asks how Sourced avoids concurrent workers stepping on the same command/events.
How do you search across millions of events for top items?
Q&A on searching event-sourced data efficiently.
How do you test rendered Phlex components?
Audience asks how to write automated tests / feature specs against Phlex output.
How do you test units with dependencies on other units?
Q&A on isolating tests for domains, read models and processes.
How do you test ViewComponents and Hotwire flows?
Audience question comparing testing ViewComponents vs React components.
How hard is it to keep authorization relationships in sync?
Q&A: operational cost of maintaining tuples alongside domain data.
How is MFA enabled per user — manual or automated?
Q&A: is flagging which users get MFA manual or script-driven?
How much time does maintaining the event-sourcing infrastructure take?
Audience question about time invested in custom event-sourcing plumbing.
How should POST retries handle changed source-of-truth state?
On retry, should the sender re-POST with updated parameters from the meanwhile-changed entity?
How to apply DDD when the client refuses to learn ubiquitous language?
Handling DDD when the domain-expert client is uncooperative.
How to assert on collection properties in Test Bench?
Audience asks how to do RSpec-style composable matcher assertions in Test Bench.
How to comply with right-to-be-forgotten given future quantum computers?
Q&A on GDPR erasure durability against future decryption capabilities.
How to convince a product manager at a medium-sized startup to invest in event sourcing?
Audience question about getting buy-in for event sourcing.
How to detect coupling in a large existing project?
Strategy for evaluating coupling when joining a big legacy codebase.
How to display heterogeneous documents after splitting a God model?
If invoice and receipt become separate models, how do we paginate/sort/search them together?
How to educate the team on modularity boundaries?
Getting a team aligned on what is one module vs another is hard — any advice?
How to migrate legacy data into an event-sourced system?
Q&A on importing legacy database data as the starting events of new streams.
How to obfuscate fields that drive app logic (e.g. patient age)?
Q&A on preserving domain-significant fields during obfuscation.
How to optimize tests around callbacks without refactoring?
Q&A: how to speed up tests using build instead of create when code is callback-heavy.
How to persist commands in an event-sourced system?
Q&A on whether commands themselves should be stored in the database.
How was the axiom-driven process introduced?
Audience asks whether the team adopted the process overnight or gradually.
Immediate benefit of event-driven architecture?
Audience question asking for a short pitch for event-driven architecture.
Internal state in automations
Does the read-model arrow into an automation mean the automation has internal state?
Is Active Record N+1 still a problem on SQLite?
Does reduced query overhead really eliminate N+1 concerns?
Is AOP a failed idea, and how does it relate to DDD?
Q&A on aspect-oriented programming's fate and overlap with event-driven DDD.
Is building DDD from scratch too slow for the business?
Q&A on the perceived overhead of layered DDD vs 'Rails way' delivery.
Is event sourcing more expensive long-term, and how to reduce cost?
Q&A on long-term cost of event sourcing and how to lower it.
Is Guard still useful?
Q&A remark on Guard for auto-running tests on file change.
Is Literal a good basis for a validation library?
Could Literal back a validation library that returns errors instead of raising?
Is MVC enough for a successful enterprise app?
Audience question on whether Rails MVC suffices for enterprise applications.
Isn't asking password confirmation business logic?
Audience pushback on the application-vs-business-logic distinction.
Isn't embedding the database in your app process a security regression?
Does putting SQLite inside the Ruby process sacrifice isolation for performance?
Isn't the dry-container recipe global state?
Audience question challenging dry-container's global container.
Is parsing and validation really a cycle because of coercion?
Audience asks whether coercion means parse/validate must loop.
Is rendering from cached pages with server-side includes possible in Rails?
Asks about ESI-style partial re-rendering from a cached static page.
Is that monads in your stack trace?
Audience jokingly asks whether Reform 3's stack trace shows dry-monads.
Is there a generator or scaffold for Turbo Native apps?
Audience question about scaffolding Turbo Native apps.
Is the sufficient-funds check a feature envy smell?
Q&A on whether handler-side querying of the entity violates 'tell don't ask'.
Is the WHERE-clause leak risk real with acts_as_tenant?
Does acts_as_tenant actually prevent the forgotten-WHERE-clause leak?
LLMs 10× product misfunctions too
Beyond technical bugs, AI ships features nobody asked for.
Local storage for tests
How to substitute S3 for local storage during tests.
Long-running automations and sagas
How do long-running automations that require multiple events get executed?
Measured async performance vs sequential and threads
Did switching from sequential/thread-based calls to async+fibers show measurable IO gains?
Metrics for refactoring impact on the business
Are there metrics showing a modeling refactor actually delivered business value?
Moderating the Merge Phase in Remote Workshops
How to bring subgroups back together when they seem to live in different worlds.
Move transformation vs refactoring when tests change
Audience asks about moving code as part of a refactoring when tests necessarily change.
Multi-column indexes scoped by tenant
Do row-level multi-tenant tables need indexes that include tenant_id?
Multi-shot Accept Backlog Concern
Does multi-shot accept risk accepting too many in-flight TCP connections?
NoSQL vs Postgres for an event store
Would MongoDB or DynamoDB be better than Postgres for storing versioned events?
Obfuscation and re-identification via record counts / outliers
Q&A on leakage via statistical outliers after obfuscation.
Orphan files from async uploads
How to handle users who upload a file but never submit the form.
Over-engineering risk of inventing layers
Won't developers just invent layers and classes for the sake of it?
Packaging a Rails engine that uses Webpacker
Q&A on having Webpacker both in a host app and in a packaged engine.
Passing Large Lists Of IDs To Jobs
Is it okay to pass hundreds of IDs to a single Sidekiq job?
Per-tenant read models via database views
Using custom database views/read models per tenant.
Port your Ractor talk to JRuby
Joking request to help a Sunday speaker convert a Ractor talk to JRuby.
Projection freshness and snapshots
How do you avoid reprojecting from the beginning on every decision?
Protecting against ffmpeg vulnerabilities
How do you protect against vulnerabilities in ffmpeg?
Relation to Painless Rails book
Is this approach an extension of the patterns in Painless Rails?
Risk of misidentifying domains during legacy re-architecture
Q&A on the danger of modularising too early around wrong domain boundaries.
Scaling Sidekiq Elastically
How to scale Sidekiq up and down like AWS Lambda containers?
Should 'add item' be split into add-new and increase-existing?
Should cart APIs expose separate methods for new vs existing items?
Should everyone just build an HTTP proxy class up front?
Is routing all HTTP through a single owned class worth doing from day one?
Should every project use mutation testing?
Audience asks whether mutation testing fits every project or only core domain logic.
Should touch on belongs_to be avoided in complex apps?
Asks whether touching parents via callbacks causes performance issues at scale.
Sidekiq And Redis Cluster
Why doesn't Sidekiq work with Redis Cluster?
Sidekiq Retry Race Condition Fixed In 7
How does Sidekiq 7 prevent jobs from running before the transaction commits?
Storing workflow state history
Can Geneva Drive store the full history of workflow states?
Tracking HTTP client library changes
How does WebMock keep up with changes in the HTTP-client libraries it monkey-patches?
Unlogged table durability on connection reset
Are Postgres unlogged tables wiped on connection reset as well as DB restart?
Unobvious direct-vs-indirect warnings per nation
Which nations are commonly misjudged as direct or indirect?
Using ChatGPT to soften messages for foreign colleagues
Is ChatGPT a reasonable solution for rephrasing blunt Polish messages for foreign programmers?
VNC↔WebSocket frame translation
How is VNC translated to WebSocket frames, and is it efficient in Ruby?
Was the rollout gradual and were there rollbacks?
Was the new flow rolled out gradually via feature flag, and did any rollbacks occur?
WebMock contract testing integration
Does WebMock support contract testing so mocks invalidate when upstream APIs change?
What about HTMX?
Q&A: Ryan's take on HTMX as a Hotwire alternative.
What are fixtures in Test Bench?
Audience asks to clarify 'fixture' vs the Rails fixture meaning.
What are the biggest advantages of WebAssembly?
Audience question on WebAssembly's practical benefits.
What does the application do?
Q&A: what is the product behind the case study?
What is an enterprise Rails application?
Panel opening question about the definition of 'enterprise Rails'.
What is an entity?
Q&A clarifying what Ladd means by 'entity' in event sourcing.
What licenses do Turbo Native libraries use?
Audience question about licensing of Hotwire/Turbo Native.
What's bad about JRuby?
Audience asks what the downsides of JRuby are.
What's the biggest design mistake you'd fix in Reform?
Audience asks what Nick would do differently if rebuilding Reform.
What team-level processes help avoid such bugs?
Q&A on post-mortems, PR repro requirements, and team practices.
What to do with an apparent equivalent mutation?
Audience asks how to handle a mutation you believe is equivalent and cannot kill.
When is a command considered processed?
Q&A on when to treat a command as processed in an event-driven system.
When is Geneva Drive overkill
When is a durable workflow engine not needed — YAGNI check.
When to stop automating and invest in culture?
Is there a point on the curve where culture becomes the better investment?
When to use OOP vs procedural
If service objects aren't object-oriented, when do you use OOP?
Where do calculations live in read models?
In Event Modeling, where do derived calculations like cart totals belong when multiple read models need them?
Where do you read from without event sourcing?
Q&A on where queries live when persistence is not event-sourced.
Where should events be stored?
Audience asks for storage advice — database, files, memory?
Why a limit of four login sessions?
Q&A: why was four chosen as the maximum concurrent sessions per user?
Why haven't you introduced events into Spree?
Have you considered an event-driven extension model instead of dependency injection?
Why implement the proxy in Ruby?
Audience asks why build a network proxy in Ruby rather than a faster language.
Why logs if traces are the best
Why focus on logs when Callaghan says traces are more powerful.
Why MFA via email and not authenticator app or SMS?
Q&A: rationale for picking email-based MFA.
Why Not Inject Dependencies Via Rails Configuration?
Q&A on using Rails config or IoC containers to wire operational dependencies.
Why not just install a client binary?
Why go through HTTP at all instead of a direct native VNC client?
Why not just run a local Postgres instance?
Wouldn't local Postgres on the same machine give most of the benefits?
Why not use Solid Queue for heavy migrations?
Can Solid Queue in batches handle transactional heavy data migrations?
Why server-side AI instead of in-browser WebLLM
Why run AI on the backend when modern browsers (and WebLLM, future embedded Chrome LLM) can do it client-side?
Why table swap instead of deletion?
Why create a fresh table and copy in the wanted rows instead of deleting unwanted rows in place?
Why two AnyCable server implementations
Audience asked why AnyCable ships both Go and Erlang servers.
Will JavaScript die as a development platform?
Panel question on JavaScript becoming merely a compile target.
Would you prefer types beside the code, not in it?
Sentiment check that speaker wants more automation and types outside of implementations.
Wrapper Job vs Logic In Job
Should Sidekiq jobs be thin wrappers over service objects, or hold logic directly?
Alerts on logs vs metrics at scale
Can log queries support alerting or must you switch to metrics?
Biggest table and total database size?
How large is the production SQLite database?
Concerns about letting users define weights
Audience asks whether giving users control over scoring weights caused problems in production.
Data consciously dropped vs unintentionally lost
Isn't it normal to consciously decide what to store and what to drop?
Deduplicating Jobs For The Same Record
How to avoid redundant jobs when many messages concern the same record?
Deploying a non-Rails app alongside Rails with Kamal
Can you deploy a second non-Rails service in the same Kamal-managed VM?
Does scale change the problem?
If three developers have the same problem as 500, why does scale matter?
Do users just create new shared accounts to dodge MFA?
Q&A: have you checked whether flagged users re-register with a shared email?
Do wide models with many columns hurt performance?
Follow-up asking whether 57-column models and many joins damage performance.
Do you chain/compose procs in practice?
Audience asks about functional composition of procs passed through blocks.
Do you protect against shared cookies?
Q&A: is cookie sharing between users mitigated?
Have you automated-tested Turbo Native hybrid apps?
Audience question on automated testing of hybrid apps.
Have you compared performance of these method-call styles?
Audience asks whether any technique offers speed benefits beyond fun.
How did the new flow handle different third-party request parameters?
Were recorded requests rewritten to match the new flow's different parameters?
How do you deal with GDPR in an immutable event store?
Q&A on handling regulatory requirements like GDPR with event sourcing.
How do you handle a timing-out page for expensive enterprise customers?
Real urgency — important page timing out for high-value customers threatening to leave.
How do you make the font size large in the terminal demo?
Q&A on the large timer font in the cipher-breaking terminal demo.
How do you negotiate performance with uneducated stakeholders who 'want it fast now'?
Negotiation when business has no notion of SLAs or budgets.
How hard was it to convince the client to test on production?
Was the client comfortable with on-production verification?
How is the event-write kept consistent with external operations?
Q&A on transactional integrity between event writes and external side effects.
How long ago did you migrate and have you regretted it?
Timing and regret about the migration.
How long did the rewrite take?
Total duration of the Lemonade quoting rewrite project.
How often should projections be rebuilt?
Does every event have to trigger a projection, and how fresh must read models be?
How to handle a 20+ filter index endpoint with big datasets?
Final audience question on combinable filters and large datasets.
How to triage performance problems?
Which performance issues to fix first.
How were manual tweaks benchmarked
How did Nebulab decide that a coefficient tweak was an improvement?
Including Unavailable Domain Experts
How to handle people who can't attend full sessions due to schedules.
Is layered DDD really junior-friendly?
Q&A challenging the claim that DDD structure helps juniors.
Meaning of the projection sequence number
Why does the projection code reference an event's position in the stream?
Optimization algorithm with a score function framing
Audience frames the problem as set-level optimization over SKUs; was it considered?
PII privacy when logging domain objects
How to handle PII when logging domain objects to a vendor.
Preserving user options when AI-prefills a constrained form
Forms give users the structure/options they can pick — how do you preserve that with AI?
Projections affecting many streams at once
How are projections handled when one action affects multiple order streams?
Real large apps using this pattern
Any large Rails app applying this layered approach?
Should I have told my partner about the rumor?
Was it right to withhold an unproven rumor about a subcontractor from a new partner?
Specifying types at the method boundary with pattern matching
Can Literal check parameter types on ordinary methods?
Turning Workshop Enthusiasm Into Real Results
How to avoid the post-workshop letdown where nothing ships.
Walking users through nested conditional forms with AI
How do you guide users through deeply nested conditional forms so they know the AI understands them correctly?
What about CPU, memory, IO, network, and large DOM profilers?
Tools for lower-level and front-end performance concerns.
What alternative strategies were considered?
Were other verification strategies considered besides production replay?
What are typical Rails performance problems?
Opening panel question on the most common performance problems in Rails applications.
What counts as 'shit in production'?
Asks whether 'shit' means bugs or a broader category of damage.
What is the biggest challenge with this approach?
Q&A on the biggest day-to-day pain of event-sourced services.
What's the biggest issue with complex data pages — data, SQL, or rendering?
Where the cost really lies when pages are slow and complex.
What's the most complex performance problem you've solved?
War stories from panelists.
When is hybrid a bad idea?
Audience question on use cases where hybrid shouldn't be chosen.
When should performance optimization end?
Heuristics for stopping optimization — 50% good-enough vs endless perfection.
Which tools do you use to monitor, triage, and fix performance problems?
Asking for concrete tooling for performance work.
Why AI backend instead of optimizing the UI into smaller steps
Why not split the 20-field form into multi-step pages of 2–5 fields instead of adding AI?
Will we all be exclusively JavaScript developers in two years?
Panel question on JavaScript dominance over backend languages.
Worth writing a paper about debugging like a scientist?
Audience suggests publishing the method as a paper.