Validation
The process of checking data, behavior, systems, models, or results against defined criteria to assess their correctness, suitability, or acceptability.
Machine Learning Platform Engineering
Model validation is the process of evaluating a trained machine learning model on data or tests intended to estimate its performance and suitability.
- How it is used
- Places model checks between training and deployment decisions within reproducible ML pipelines.
- How deep it goes
- Works through a KFP component that runs YOLO validation and logs metrics; leaves custom validation logic open to evolving requirements.
- Examples
- Uses YOLO object-detection metric logging, PVC-mounted model weights, and a movie-recommender training workflow.
- What you can do after
- Build it from scratch as a KFP validation component that loads a trained YOLO model and records validation metrics.
Read how Machine Learning Platform Engineering explains validation →
ASP.NET Core in Action, 3e
Validation is the application framework process that checks model data against rules to determine whether it is acceptable.
- How it is used
- Checks user-supplied request values before application logic uses them, including for security, formatting, and business rules.
- How deep it goes
- Works through attributes, minimal API filters, and Razor Pages failure handling, then compares custom attributes with FluentValidation.
- Examples
- Uses a user-details form, a minimal API user endpoint, and currency-conversion rules to demonstrate valid and invalid input.
- What you can do after
- Apply it via DataAnnotations and a minimal API filter, or check ModelState.IsValid in a Razor Pages handler.
Pro ASP.NET Core 7, Tenth Edition
Validation is the process of checking data against rules to determine whether it is acceptable for processing.
- How it is used
- Checks submitted form and API data before actions use it or store it, while providing feedback when it fails.
- How deep it goes
- Works through attribute rules, ModelState branching, error display, and API responses; identifies custom and client-side validation without detailing their implementations.
- Examples
- Uses an RSVP form, a product API POST, and a database-backed product create form.
- What you can do after
- Build it from scratch with validation attributes, ModelState checks, form feedback, and a bad-request response for invalid API input.
Read how Pro ASP.NET Core 7, Tenth Edition explains validation →
Build AI-Enhanced Web Apps
Validation is the checking of input against permitted formats, values, or constraints before it is accepted or processed.
- How it is used
- Protects AI request processing by checking submitted data and tests whether generated text meets specified constraints.
- How deep it goes
- Works through threat-model questions, server-versus-client responsibilities, a Zod request schema, and a constraint test; defers fuller production response verification.
- Examples
- Uses a length-limited prompt endpoint, a mocked word-count-and-keyword test, and file-upload type and size checks.
- What you can do after
- Apply it via Zod to check incoming prompts and practice/perform constraint checks in generated-text tests.
The Art of AI Product Development
Data validation is the process of checking data for correctness, consistency, completeness, and conformity to specified conditions.
- How it is used
- Checks AI opportunities, training data, application inputs and outputs, and prototypes where mistakes could undermine users or customer trust.
- How deep it goes
- Provides risk-based choices and concrete checks for data, queries, schemas, and user reactions, without a detailed implementation procedure.
- Examples
- Uses corrupted recommendation training data, manipulative prompts, schema-checked model outputs, and prototypes that simulate AI failures.
- What you can do after
- Practice/perform risk-matched checks before training and deployment, and use user feedback and production signals to investigate failures.
Read how The Art of AI Product Development explains validation →
Pro Angular 16
Data validation is the process of checking input data against specified rules for correctness, completeness, or acceptability.
- How it is used
- Checks product and checkout form input before accepting a product or saving an order.
- How deep it goes
- Works through field rules, status, error messages, whole-form checks, summaries, and reset; identifies custom rules without implementing them here.
- Examples
- Uses a product-entry form with name, category, and price fields, plus an order checkout form.
- What you can do after
- Apply it via Angular forms to reject invalid submissions and show field-level or form-wide feedback.
Machine Learning System Design
A procedure for partitioning or otherwise structuring data and evaluation conditions to estimate a machine learning system's performance under expected operating conditions.
- How it is used
- Estimates predictive power on unseen data while reflecting the knowledge available during real operation.
- How deep it goes
- Works through split assumptions, rolling windows, gaps, and nested evaluation, while leaving the best search-query split ratio open.
- Examples
- Contrasts rolling four-week retail forecasts with user-grouped search queries and deterministic split buckets.
- What you can do after
- Build it from scratch: choose repeatable splits, reproduce inference conditions, and separate tuning from final performance estimation.
Read how Machine Learning System Design explains validation →
Build Financial Software with Generative AI (From Scratch)
Validation is the process of checking data or inputs against defined correctness and acceptance rules.
- How it is used
- Guards ACH APIs, records, generated files, and UI values against malformed or unacceptable data.
- How deep it goes
- Works through regex constraints, database-backed field validation, HTTP error handling, and positive and negative tests; recovery from rejected files is deferred.
- Examples
- Uses file_id patterns, file-ID modifiers, bank routing numbers, type-6 ACH entry counts, and Base64 image strings.
- What you can do after
- Apply it via Pydantic and FastAPI to constrain fields, return handled validation errors, and test accepted and rejected inputs.
Read how Build Financial Software with Generative AI (From Scratch) explains validation →
Django in Action
Validation is the process of checking whether a value satisfies specified correctness or format constraints.
- How it is used
- Uses field and form checks to keep invalid values from entering application processing.
- How deep it goes
- Explains field-based checks, is_valid(), cleaned_data, error redisplay, and redirects without detailing validation internals.
- Examples
- Demonstrates EmailField format checking and a RiffMates comment form handled by one GET-and-POST view.
- What you can do after
- Apply it via Django forms to accept valid submissions, redisplay invalid input with errors, and redirect after success.
Learn Go with Pocket-Sized Projects
Input validation is the process of checking submitted data against required formats, types, and value constraints before processing it.
- How it is used
- Protects domain entities, API boundaries, HTTP handlers, and table-driven tests from invalid or incorrectly asserted data.
- How deep it goes
- Works through normalization, rejection, defaults, typed errors, transport mapping, parsing, and focused tests; it stops short of a validation library implementation.
- Examples
- Uses money conversion tests, habit creation, whitespace-only names, missing defaults, gRPC status codes, and HTTP form parsing.
- What you can do after
- Build it from scratch by validating domain entities, returning typed errors, testing outcomes, and mapping failures at HTTP or gRPC boundaries.
Read how Learn Go with Pocket-Sized Projects explains validation →
Kafka for Architects
Message validation is the checking of a message's structure and content against required validity rules before it is accepted or processed.
- How it is used
- Protects event-driven consumers by checking message conformance before publication and planning for invalid messages that still reach topics.
- How deep it goes
- Explains the client-side validation flow, broker limitation, and consumer failure paths without a low-level implementation walkthrough.
- Examples
- Uses schema-ID retrieval and a missing-field message that repeatedly fails deserialization until the consumer handles or bypasses it.
- What you can do after
- Diagnose/debug with it when invalid messages disrupt consumers, and establish handling policies for dead-letter topics and offset recovery.
Starting Data Analytics with Generative AI and Python
Validation is the process of checking whether an analytical or model-generated result is accurate, relevant, plausible, and fit for its intended purpose.
- How it is used
- Uses checks to protect data quality and assess whether AI-generated results support responsible analytical or business decisions.
- How deep it goes
- Provides procedural steps, test-case guidance, and reporting requirements without developing formal statistical validation theory.
- Examples
- Uses malformed dates, cross-table foreign-key violations, and AI-generated financial-product pricing code with edge-case tests.
- What you can do after
- Diagnose data and model-output problems through constraints, sanity checks, expert review, test cases, and documented verification.
Read how Starting Data Analytics with Generative AI and Python explains validation →
Learn Haskell by Example
Validation is the process of checking data against specified conditions to determine whether it is acceptable or well-formed.
- How it is used
- Transforms untrusted parsed PNM data into image data only after format, size, and color constraints pass.
- How deep it goes
- Works through per-pixel conversion, bounds checks, helper predicates, and top-level error propagation with Either and mapM.
- Examples
- Uses bitmap, graymap, and RGB pixmap pixels, including invalid constructors and values exceeding a maximum.
- What you can do after
- Apply it via JuicyPixels pixel types to reject malformed PNM parse results before later image processing.
ScyllaDB in Action
Data validation is the process of checking that migrated data is complete and accurate relative to its source.
- How it is used
- Uses dual-reading to verify that a live migration remains invisible to upstream applications.
- How deep it goes
- Defines completeness and correctness, then details comparisons, retries, logging, and mismatch classification.
- Examples
- Uses source-versus-destination reads and millisecond-versus-microsecond timestamps to classify mismatches.
- What you can do after
- Diagnose migration mismatches by comparing dual-reads, retrying transient differences, and investigating persistent ones.
Generative AI in Action
Validation is the process of checking generated software or content against functional, quality, security, and compliance requirements.
- How it is used
- Checks generated code, production data, model behavior, and LLM outputs before or during enterprise deployment.
- How deep it goes
- Provides review criteria and production monitoring measures, plus security safeguards, without specifying a complete validation pipeline.
- Examples
- Uses generated-code review criteria and an LLM-generated SQL injection scenario to show what needs checking.
- What you can do after
- Practice/perform validation by reviewing generated code, monitoring data and models, and checking inputs and outputs at security boundaries.
Classic Computer Science Problems in Swift: Essential techniques for practicing programmers
Validation is the evaluation of a trained model on examples held apart from the data used to adjust its parameters.
- How it is used
- Evaluates classification networks on samples excluded from parameter training.
- How deep it goes
- Demonstrates dataset splitting, repeated training, output interpretation, and accuracy reporting without explaining validation internals.
- Examples
- Uses iris and wine datasets with held-out subsets, interpreted class outputs, and reported correct counts and percentages.
- What you can do after
- Evaluate a trained Swift neural network with validate and quantify classification accuracy on held-out samples.
F# in Action: A Revised Edition of Get Programming with F#
Validation is the process of checking data against rules that determine whether it is acceptable for a domain or operation.
- How it is used
- Guards JSON and HTTP input before it enters a rich domain model.
- How deep it goes
- Works through Result-returning field checks and combination, while deferring all-error accumulation to FsToolkit Validation.
- Examples
- Uses raw customer and guitar records, plus a dataset parsing and analysis exercise.
- What you can do after
- Apply it via Result-returning F# functions to convert raw input into trusted domain values and report failures.
Read how F# in Action: A Revised Edition of Get Programming with F# explains validation →
Coding with AI
Validation is the checking of input or application state against expected rules and constraints.
- How it is used
- Applies validation to user input, application behavior, and AI-generated code before accepting results.
- How deep it goes
- Demonstrates requirements checks, form handling, error cases, testing, and human review, but does not detail validation implementation internals.
- Examples
- Uses a Flask quiz form, stored answers, question navigation, database testing, and review of an overengineered SQLAlchemy proposal.
- What you can do after
- Diagnose missing inputs, implementation failures, and unsuitable AI suggestions before deploying or adopting them.
The Well-Grounded Python Developer: How the pros use Python and Flask
Validation is the process of checking input data against required presence, format, and value constraints.
- How it is used
- Protects authentication, registration, and comment-submission workflows from missing, malformed, or unacceptable input.
- How deep it goes
- Explains field validators, custom database-backed checks, error display, and conditional persistence without cataloging broader validation strategies.
- Examples
- Uses login, new-user registration, duplicate-email detection, and hierarchical comment creation as concrete validation scenarios.
- What you can do after
- Apply it via Flask-WTF validators and validate_on_submit() before rendering errors or committing accepted records.
Read how The Well-Grounded Python Developer: How the pros use Python and Flask explains validation →
Grokking Machine Learning
A model-development procedure that evaluates candidate models or settings on held-out data without using the final test data for those decisions.
- How it is used
- Uses held-out validation data to choose models or settings while reserving the test set for a final check.
- How deep it goes
- Separates data into training, validation, and testing roles, gives common split ratios, and defers further procedural detail.
- Examples
- Compares polynomial regression models of degrees 1, 2, and 10, and gives 60-20-20 and 80-10-10 dataset splits.
- What you can do after
- Apply the three-way split to choose a model with validation data and reserve the test data for final evaluation.
Build an AI Agent (From Scratch)
Data validation is the process of checking data against specified structural, type, and constraint requirements.
- How it is used
- Checks agent tool inputs before actions and agent outputs against explicit, programmable requirements.
- How deep it goes
- Shows concrete schema constraints and early rejection, but does not implement a runtime validator.
- Examples
- Uses reservation fields and ranges, a light-state enum, refund identifiers, and output-format constraint checks.
- What you can do after
- Operate/configure it by constraining tool arguments, checking inputs before requests, and testing outputs against precise rules.
Read how Build an AI Agent (From Scratch) explains validation →
100 SQL Server Mistakes and How to Avoid Them
Input validation is the process of checking supplied data against allowed formats, values, or constraints before processing it.
- How it is used
- Protects web-login queries by checking or safely handling user-supplied values before database processing.
- How deep it goes
- Demonstrates the vulnerability and injection result, then stops short of defining a validation checklist or validation rules.
- Examples
- Uses a C# Web Forms login, a Users table, a normal credential query, and the payload ' OR 1=1--.
- What you can do after
- Apply it by replacing concatenated login input with stored-procedure parameters and testing an injection payload.
Read how 100 SQL Server Mistakes and How to Avoid Them explains validation →
Functional Design and Architecture: Examples in Haskell
Validation is the systematic checking of input or data against rules before it is accepted for further processing.
- How it is used
- Protects web-service business logic by rejecting invalid requests at the API boundary.
- How deep it goes
- Descends from immediate handler failures to applicative, nested, path-aware validation, without detailing library internals.
- Examples
- Uses nonempty request fields, pointed validators, nested structures, and path-specific validation errors.
- What you can do after
- Apply it via Haskell handlers and applicative validators to reject bad requests and report nested failures.
Read how Functional Design and Architecture: Examples in Haskell explains validation →
Terraform in Depth: Infrastructure as Code with Terraform and OpenTofu
Validation is the process of checking configuration inputs against required formats, values, or dependencies before processing them.
- How it is used
- Catches Terraform input and provider-configuration errors before they lead to misconfiguration or client creation.
- How deep it goes
- Explains variable conditions and error messages, then gives a provider procedure for collecting diagnostics before returning.
- Examples
- Uses integer-range checks, an instance-count input, and an unknown provider host to show distinct failures.
- What you can do after
- Apply it via Terraform validation blocks and provider diagnostics to report actionable input errors before using configuration values.
Simple Object-Oriented Design: Create clean, maintainable applications
A validation mechanism checks requested values or state changes against validity constraints.
- How it is used
- Protects objects and client-facing operations from invalid data and state changes.
- How deep it goes
- Explains ownership, service extraction, setter design, and two validation approaches, but does not detail their implementations.
- Examples
- Uses training-course offerings, participant limits, enrollment, database-dependent checks, and protected versus unprotected baskets.
- What you can do after
- Apply validation by placing local checks in entities and complex business checks in services or dedicated validation classes.
Building Web APIs with ASP.NET Core
Data validation is the process of checking incoming data for validity, accuracy, meaningfulness, and security before processing it.
- How it is used
- Protects web API endpoints by checking incoming HTTP data before controller processing.
- How deep it goes
- Explains model binding, annotations, ModelState, cross-validation, and automatic versus manual processing.
- Examples
- Uses invalid pageIndex input and a DomainDTO rule requiring an allowed Id or Name.
- What you can do after
- Apply it via ASP.NET Core model binding, validation attributes, IValidatableObject, and ModelState handling.
Read how Building Web APIs with ASP.NET Core explains validation →
Go in Action, Second Edition
Validation is the process of checking that input data satisfies required structural, type, and semantic constraints.
- How it is used
- Validates JSON configurations, template-specific schemas, command-line values, environment values, and file-based settings before application use.
- How deep it goes
- Descends from parsing and type checks to range checks and custom URL validation through UnmarshalJSON.
- Examples
- Uses configuration JSON, template dispatch, Host URL checks, bounded port parsing, Boolean parsing, and formatted configuration values.
- What you can do after
- Builds validation from scratch with encoding/json, custom UnmarshalJSON methods, and strconv parsing functions.
AI Agents in Action, Second Edition
Output validation is the process of checking generated or returned data against defined conditions before it is accepted or delivered.
- How it is used
- Checks agent inputs, outputs, and tool results before they are accepted or used in further actions.
- How deep it goes
- Demonstrates string and length checks with tripwires, while allowing more complex logic without detailing it.
- Examples
- Uses a forbidden research title, string comparisons, length checks, and a typed research output model.
- What you can do after
- Apply guardrails to check agent inputs and outputs, trigger failures, and validate tool results before further actions.
Read how AI Agents in Action, Second Edition explains validation →
Rust Servers, Services, and Apps
Input validation is the process of checking supplied data against format, length, type, or business constraints before processing it.
- How it is used
- Checks registration and sign-in form data before processing and guides users when entries are invalid.
- How deep it goes
- Distinguishes simple browser checks from complex handler checks, without detailing their implementation.
- Examples
- Uses tutor registration fields, sign-in credentials, field-length checks, and username availability as examples.
- What you can do after
- Apply basic checks in HTML and complex checks in handlers, then return invalid forms with suitable feedback.
Read how Rust Servers, Services, and Apps explains validation →
.NET MAUI in Action
Validation is the process of checking user-provided data against validity rules and informing the user when the data is invalid.
- How it is used
- Improves UI usability by checking entered data, communicating failures, and stopping invalid form submission.
- How deep it goes
- Explains validation rules, feedback, and implementation choices, but stops short of detailed bindable-property implementation.
- Examples
- Uses a stock-count stepper, nonnegative range checks, warning labels, approved lists, and email-address checks.
- What you can do after
- Apply validation through warning labels, control-state checks, or .NET MAUI Community Toolkit behaviors.
Related concepts
- evaluation datasetuses · An evaluation dataset is a collection of examples or test cases, often with expected or judged outcomes, designated to assess a model or system against defined criteria.
- hallucinationaddresses · A hallucination is model-generated content that is false, fabricated, or unsupported by relevant evidence but presented as plausible or factual.
- handlerused by · A handler is a callable or software component that processes an event, request, message, or other input and performs associated processing.
- enumused by · An enum is a type or language construct whose values are selected from a finite set of named alternatives, which may carry associated data.
- Validationuses · Validation is a computation or value type representing either a successful result or failure information while supporting successful processing and failure propagation or accumulation.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.