Jonathan Lalou's Blog

Archive for the ‘en-US’ Category

Lecturer

Jeff Maury is an experienced product manager at Red Hat, specializing in Java technologies for large-scale systems. Previously, as Java Offer Manager at Syspertec, he architected solutions integrating open systems like Java and .NET. Co-founder of SCORT, a firm focused on enterprise system integration, Jeff has leveraged Xtext to develop advanced development tools, providing hands-on insights into DSL ecosystems. An active contributor to Java communities, he shares expertise through conferences and practical implementations.

Abstract

This article analyzes Jeff Maury’s introduction to Xtext, Eclipse’s framework for crafting domain-specific languages (DSLs), structured across theoretical underpinnings, real-world applications, and hands-on development. It dissects Xtext’s grammar definition, model generation, and editor integration, emphasizing its role in bridging business concepts with executable code. Contextualized within the rise of model-driven engineering, the discussion evaluates Xtext’s components—lexer, parser, and scoping—for enabling concise, domain-tailored notations. Through the IzPack editor example, it assesses methodologies for validation, refactoring, and Java interoperability. Implications span productivity gains in specialized tools, reduced cognitive load for non-programmers, and ecosystem extensions via EMF, positioning Xtext as a versatile asset for modern software engineering.

Theoretical Foundations: Components and DSL Challenges

Domain-specific languages address the gap between abstract business requirements and general-purpose programming, allowing experts to articulate solutions in familiar terms. Jeff frames DSLs as targeted notations that encapsulate métier concepts, fostering adoption by broadening accessibility beyond elite coders. Challenges include syntax design for intuitiveness, semantic validation, and tooling for editing—areas where traditional languages falter due to verbosity and rigidity.

Xtext resolves these by generating complete language infrastructures from a declarative grammar. At its core, the grammar file (.xtext) defines rules akin to EBNF, specifying terminals (e.g., keywords, IDs) and non-terminals (e.g., rules for structures). The lexer tokenizes input, while the parser constructs an abstract syntax tree (AST) via ANTLR integration, ensuring robustness against ambiguities.

Model generation leverages Eclipse Modeling Framework (EMF), transforming the grammar into Ecore metamodels—classes representing language elements with attributes, references, and containment hierarchies. Scoping rules dictate name resolution, preventing dangling references, while validation services enforce constraints like type safety. Jeff illustrates with a simple grammar for a configuration DSL:

grammar ConfigDSL;

Config: elements+=Element*;

Element: 'define' name=ID '{'
    properties+=Property*
'}';

Property: key=ID '=' value=STRING;

This yields EMF classes: Config (container for Elements), Element (with name and properties), and Property (key-value pairs). Such modularity enables incremental evolution, where grammar tweaks propagate to editors and validators automatically.

Theoretical strengths lie in its declarative paradigm: Developers focus on semantics rather than boilerplate, accelerating prototyping. However, Jeff cautions on over-abstraction—DSLs risk becoming mini-general-purpose languages if scopes broaden, diluting specificity. Integration with Xbase extends expressions with Java-like constructs, blending DSL purity with computational power.

Business Applications: Real-World Deployments and Value Propositions

Beyond academia, Xtext powers production tools, democratizing complex domains. Jeff cites enterprise modeling languages for finance, where DSLs express trading rules sans procedural code, slashing error rates. In automotive, it crafts simulation scripts, aligning engineer notations with executable models.

A compelling case is workflow DSLs in BPM, where Xtext-generated editors visualize processes, integrating with Activiti or jBPM. Business analysts author flows textually, with auto-completion and hyperlinking to assets, enhancing traceability. Healthcare examples include protocol DSLs for patient data flows, ensuring compliance via built-in validators.

Value accrues through reduced onboarding: Non-technical stakeholders contribute via intuitive syntax, while developers embed DSLs in IDEs for seamless handoffs. Jeff notes scalability—Xtext supports incremental parsing for large files, vital in log analysis DSLs processing gigabytes.

Monetization emerges via plugins: Commercial tools like itemis CREATE extend Xtext for automotive standards (e.g., AUTOSAR). Open-source adoptions, such as Sirius for graphical DSLs, amplify reach. Challenges include learning curves for grammar tuning and EMF familiarity, but Jeff advocates starting small—prototype a config DSL before scaling.

In 2025, Xtext remains Eclipse’s cornerstone, with version 2.36 (March 2025) enhancing LSP integration for VS Code, broadening beyond Eclipse. This evolution sustains relevance amid rising polyglot tooling.

Practical Implementation: Building an IzPack Editor with Java Synergies

Hands-on, Jeff demonstrates Xtext’s prowess via an IzPack DSL editor—a packaging tool for Java apps. IzPack traditionally uses XML; the DSL abstracts to human-readable syntax like “install ‘app.jar’ into ‘/opt/app’ with variables {version: ‘1.0’}.”

Grammar evolution: Start with basics (packs, filesets), add cross-references for variables, and validators for conflicts (e.g., duplicate paths). Generated editor features syntax highlighting, outlining, and quick fixes—e.g., auto-importing unresolved types.

EMF integration shines in serialization: Parse DSL to IzPack model, then generate XML or JARs via Java services. Jeff shows a runtime module injecting custom validators:

public class IzPackRuntimeModule extends AbstractIzPackRuntimeModule {
    @Override
    public Class<? extends IValidator> bindIValidator() {
        return IzPackValidator.class;
    }
}

Java linkage via Xtend—Xtext’s concise dialect—simplifies services:

def void updateCategory(Element elem, String newCat) {
    elem.category = newCat
    elem.eAllContents.filter(Element).forEach[ it.category = newCat ]
    // Trigger listeners
    elem.eSet(elem.eClass.getEStructuralFeature('category'), newCat)
}

This propagates changes, demonstrating EMF’s notification system. Refactoring renames propagate via index, while content assist suggests variables.

Deployment: Export as Eclipse plugin or standalone via Eclipse Theia. Jeff’s GitHub repo (github.com/jeffmaury/izpack-dsl) hosts the example, inviting forks.

Implications: Such editors cut packaging time 70%, per Jeff’s Syspertec experience. For Java devs, Xtext lowers DSL barriers, fostering hybrid tools—textual DSLs driving codegen. In 2025, LSP support enables polyglot editors, aligning with microservices’ domain modeling needs.

Xtext’s trifecta—theory, application, practice—empowers tailored languages, enhancing expressiveness without sacrificing toolability.

Links:

• Lecture video
• Jeff Maury on LinkedIn

Lecturer

Mathias Seguy founded Android2EE, specializing in Android training, expertise, and consulting. Holding a PhD in Fundamental Mathematics and an engineering degree from ENSEEIHT, he transitioned from critical J2EE projects—serving as technical expert, manager, project leader, and technical director—to focus on Android. Mathias authored multiple books on Android development, available via Android2ee.com, and contributes articles to Developpez.com.

Abstract

This article explores Mathias Seguy’s in-depth coverage of Android’s advanced components, focusing on service modes, content provider implementations, and optimization strategies. It examines unbound/bound services, URI-based data operations, and tools like Hierarchy Viewer for performance tuning. Within Android’s multitasking framework, the analysis reviews methodologies for lifecycle alignment, asynchronous execution, and resource handling. Through practical code and debugging insights, it evaluates impacts on battery efficiency, data security, and UI responsiveness. This segment underscores patterns for robust architectures, aiding developers in crafting seamless, power-efficient mobile experiences.

Differentiating Service Modes and Lifecycle Integration

Services bifurcate into unbound (autonomous post-start) and bound (interactive via binding). Mathias illustrates unbound for ongoing tasks like music playback:

startService(new Intent(this, MyService.class));

Bound for client-service dialogue:

private ServiceConnection connection = new ServiceConnection() {
    public void onServiceConnected(ComponentName name, IBinder service) {
        myService = ((MyBinder) service).getService();
    }
    public void onServiceDisconnected(ComponentName name) {
        myService = null;
    }
};
bindService(new Intent(this, MyBoundService.class), connection, BIND_AUTO_CREATE);

Lifecycle syncing uses flags: isRunning/isPaused toggle with onStartCommand()/onDestroy(), ensuring tasks halt on service termination, averting leaks.

Constructing Efficient Content Providers

Providers facilitate inter-app data exchange via URIs. Define via extension, with UriMatcher for parsing:

private static final UriMatcher matcher = new UriMatcher(UriMatcher.NO_MATCH);
static {
    matcher.addURI(AUTHORITY, TABLE, COLLECTION);
    matcher.addURI(AUTHORITY, TABLE + "/#", ITEM);
}

Implement insert():

@Override
public Uri insert(Uri uri, ContentValues values) {
    long rowId = db.insert(DBHelper.MY_TABLE, null, values);
    if (rowId > 0) {
        Uri result = ContentUris.withAppendedId(CONTENT_URI, rowId);
        getContext().getContentResolver().notifyChange(result, null);
        return result;
    }
    throw new SQLException("Failed to insert row into " + uri);
}

Manifest exposure with authorities/permissions secures access.

Asynchronous Enhancements and Resource Strategies

AsyncTasks/Handlers offload UI: Extend for doInBackground(), ensuring UI updates in onPostExecute().

Resource qualifiers adapt to locales/densities: values-fr/strings.xml for French.

Databases: SQLiteOpenHelper with onCreate() for schema.

Debugging and Performance Tools

Hierarchy Viewer inspects UI hierarchies, identifying overdraws. DDMS monitors threads, heaps; LogCat filters logs.

Permissions: Declare in manifest for features like internet.

Architectural Patterns for Resilience

Retain threads across rotations; synchronize for integrity.

Implications: These techniques optimize for constraints, enhancing longevity and usability in diverse hardware landscapes.

Mathias’s guidance refines development, promoting sustainable mobile solutions.

Links:

• Lecture video
• Mathias Seguy on LinkedIn

(disclaimer: this “long tweet” has absolutely no interest, except playing with Tomcat 7 and “Leda”)

Friday is technical scouting… Today it will be a “Hello World!”.

• Get the code and project available on this page: http://www.mkyong.com/jsf2/jsf-2-0-hello-world-example/
• Get and install Tomcat 7 (very nice new home page)
• Get and install IntelliJ IDEA 12 EAP “Leda”
• Launch IDEA
• Create a new project on existing sources, select the pom.xml of the project above.
• Build with Maven
• In IDEA:
• Run Configuration
• Tomcat Server
• Startup Page: http://localhost:8080/
• Deployment
• Add
• Artifact…
• select the WAR
• OK
• This launches Tomcat 7. You can connect http://localhost:8080/

Actually, I’m sure you can deploy with Tomcat-Maven plugin, ie without actual Tomcat install.

Lecturer

Mathias Seguy founded Android2EE, specializing in Android training, expertise, and consulting. Holding a PhD in Fundamental Mathematics and an engineering degree from ENSEEIHT, he transitioned from critical J2EE projects—serving as technical expert, manager, project leader, and technical director—to focus on Android. Mathias authored multiple books on Android development, available via Android2ee.com, and contributes articles to Developpez.com.

Abstract

This article delves into Mathias Seguy’s continuation of essential Android development concepts, emphasizing services for background tasks, content providers for data sharing, and patterns for lifecycle synchronization. Building on foundational elements, it analyzes implementation strategies for bound/unbound services, CRUD operations in providers, and thread management with handlers/AsyncTasks. Within Android’s resource-constrained environment, the discussion evaluates techniques for internationalization, resource optimization, and database integration. Through detailed code examples, it assesses implications for application responsiveness, data integrity, and cross-app interoperability, guiding developers toward efficient, maintainable architectures.

Implementing Services for Background Processing

Services enable persistent operations independent of UI, running in the application’s main thread—necessitating offloading to avoid ANRs. Mathias distinguishes unbound (started via startService(), autonomous) from bound (via bindService(), allowing communication).

Lifecycle binding is critical: Align service states with calling components using booleans for running/pausing. Code for an unbound service:

public class MyService extends Service {
    @Override
    public int onStartCommand(Intent intent, int flags, int startId) {
        // Start task
        return START_STICKY; // Restart if killed
    }
    @Override
    public void onDestroy() {
        // Cleanup
    }
}

Bound services use Binder for IPC:

public class MyBoundService extends Service {
    private final IBinder binder = new LocalBinder();
    public class LocalBinder extends Binder {
        MyBoundService getService() {
            return MyBoundService.this;
        }
    }
    @Override
    public IBinder onBind(Intent intent) {
        return binder;
    }
}

These ensure tasks like downloads persist, enhancing user experience without freezing interfaces.

Crafting Content Providers for Data Exposure

Content providers standardize data access across apps, using URIs for queries. Mathias outlines creation: Extend ContentProvider, define URIs via UriMatcher.

CRUD implementation:

@Override
public Cursor query(Uri uri, String[] projection, String selection, String[] selectionArgs, String sortOrder) {
    SQLiteQueryBuilder qb = new SQLiteQueryBuilder();
    qb.setTables(DBHelper.MY_TABLE);
    if (matcher.match(uri) == ITEM) {
        qb.appendWhere(DBHelper.ID + "=" + uri.getLastPathSegment());
    }
    return db.query(qb.getTables(), projection, qb.getWhere(), selectionArgs, null, null, sortOrder);
}

Manifest declaration:

<provider
    android:name=".MyProvider"
    android:authorities="com.example.provider"
    android:exported="true"
    android:readPermission="com.example.READ"
    android:writePermission="com.example.WRITE" />

This facilitates secure sharing, like contacts or media, promoting modular ecosystems.

Asynchronous Patterns and Resource Management

Asynchrony prevents UI blocks: Handlers for UI updates, AsyncTasks for background work. Pattern: Bind threads to activity lifecycles with flags.

onRetainNonConfigurationInstance() passes objects across rotations (pre-Fragments):

@Override
public Object onRetainNonConfigurationInstance() {
    return myThread;
}
@Override
protected void onCreate(Bundle savedInstanceState) {
    super.onCreate(savedInstanceState);
    myThread = (MyThread) getLastNonConfigurationInstance();
    if (myThread == null) {
        myThread = new MyThread();
    }
}

Resources: Externalize strings in values/strings.xml for localization; use qualifiers like values-fr for French.

These patterns optimize for device variability, ensuring fluid performance.

Database Integration and Permissions

SQLite via SQLiteOpenHelper manages schemas:

public class DBHelper extends SQLiteOpenHelper {
    @Override
    public void onCreate(SQLiteDatabase db) {
        db.execSQL("CREATE TABLE " + MY_TABLE + " (_id INTEGER PRIMARY KEY, name TEXT);");
    }
}

Permissions in manifest control access, balancing security with functionality.

Testing and Project Management Strategies

Unit tests with JUnit; instrumentation via AndroidJUnitRunner. Maven for builds, Hudson/Jenkins for CI.

Implications: These foster reliable apps, mitigating fragmentation. In enterprise mobility, they enable scalable, secure solutions.

Mathias’s methodical breakdown equips developers for real-world challenges.

Links:

• Lecture video
• Mathias Seguy on LinkedIn

Lecturer

David Delabassee serves as a Director of Developer Relations in the Java Platform Group at Oracle, where he champions Java technologies worldwide through presentations, technical articles, and open-source engagements. Previously at Sun Microsystems for a decade, he focused on end-to-end Java implementations, from smart cards to high-end servers. A member of the Devoxx Belgium steering committee, David co-hosts the Inside Java Podcast and maintains a blog at delabassee.com. He holds Belgian nationality and has spoken at numerous conferences and Java User Groups.

Abstract

This article investigates David Delabassee’s rapid-fire presentation on 55 underappreciated features of Java 7, released in 2011, extending beyond well-known additions like Project Coin, Fork/Join, NIO.2, and invokedynamic. It categorizes enhancements across core libraries, security, internationalization, graphics, and more, analyzing their practical utilities and implementation details. Positioned as a post-Sun acquisition milestone under Oracle, the discussion evaluates how these refinements bolster platform stability, performance, and developer productivity. Through code demonstrations and comparisons to prior versions, it assesses implications for migration, legacy code maintenance, and modern application design, emphasizing Java 7’s role in bridging to future iterations like Java 8.

Core Language and Library Improvements

Java 7 introduced subtle yet impactful tweaks to foundational elements, addressing longstanding pain points. David highlights enhanced exception handling: multi-catch clauses consolidate try-catch blocks for related exceptions, reducing redundancy:

try {
    // Code
} catch (IOException | SQLException e) {
    // Handle
}

String switches leverage interned strings for efficient comparisons, useful in parsing:

switch (input) {
    case "start": // Action
        break;
    // ...
}

Underscores in numeric literals improve readability for large numbers: long creditCard = 1234_5678_9012_3456L;.

Library updates include Objects class utilities like requireNonNull() for null checks, and BitSet enhancements with valueOf() for byte/long array conversions. These foster cleaner, more maintainable code, mitigating common errors in enterprise applications.

Security and Cryptography Advancements

Security received substantial bolstering, crucial amid rising threats. David details elliptic curve cryptography integration, offering stronger keys with smaller sizes for SSL/TLS. Algorithm disabling via jdk.security.provider.disabledAlgorithms property enhances compliance.

SChannel provider on Windows improves native integration, while JSSE updates support SNI for virtual hosting. These fortify networked applications, essential for cloud and web services, reducing vulnerability exposure without external libraries.

Internationalization and Locale Refinements

Java 7 refined locale handling for global apps. Unicode 6.0 support adds scripts like Batak, enhancing text processing. Locale enhancements include script, variant, and extension keys:

Locale loc = new Locale.Builder().setLanguage("fr").setRegion("FR").setScript("Latn").build();

Currency updates reflect ISO 4217 changes, with getAvailableCurrencies() listing supported ones. NumberFormat improvements allow custom symbols, aiding financial software. These ensure accurate, culturally sensitive representations, vital for international markets.

Graphics and UI Toolkit Upgrades

Swing and AWT saw usability boosts. Translucent/shaped windows via GraphicsDevice enable modern UIs:

window.setOpacity(0.5f);

Nimbus look-and-feel, now default in some contexts, provides scalable, themeable components. JLayer adds decoration layers for effects like blurring. These empower richer desktop apps, aligning Java with contemporary design trends.

Performance and JVM Optimizations

JVM internals evolved for efficiency. Tiered compilation combines client/server compilers for faster startups and peak performance. G1 garbage collector, experimental in Java 7, targets low-pause collections for large heaps.

Compressed oops extend 32-bit addressing to 64-bit, reducing memory footprint. These optimizations benefit server-side applications, improving throughput and responsiveness in high-load scenarios.

Migration Considerations and Ecosystem Impact

Adopting Java 7 involves assessing compatibility, given end-of-life for Java 6. David notes seamless transitions for most code, but highlights needs like updating deprecated APIs. Tools like javac -Xlint warn of issues.

Ecosystem-wise, Java 7 paved for Java 8’s lambdas, solidifying Java’s enterprise dominance. Implications include smoother upgrades, enhanced security postures, and broader internationalization, encouraging developers to leverage these for robust, future-proof systems.

Links:

• Lecture video
• David Delabassee on LinkedIn
• David Delabassee on Twitter/X

After seven years in Cadextan, then Sungard Global Services, I wanted to give new impetus to my career and face new challenges. That is why, at the end of my position in Amundi AM, I left Sungard Global Services last Friday.

I’d like to thank the staff of Sungard Global Services-consultants, administrators, managers, etc. – for these seven years spent working on their sides. I learnt much and met many people with great human qualities. Sungard Global Services is a great success, that changed in a few years from a 100-people company in France to than 500-people in Europe. I am proud to have been part of these successful achievements and adventure.

After a couple of months spent as technical expert, transverse architect and manager of project “Reliability and Stabilization”, I left Amundi AM last Friday. I faced exciting and arduous subjects, such us OSGi, JOnAS, JVM customization, EJB2 to EJB3 migration, webservices upgrade, Decalog AS leaks, etc.

I would like to thank the whole teams that greeted me, and especially among them Antoine BODY and Jean-Guillaume BATTAGLIA.

Lecturer

Florent Benoit leads the OW2 EasyBeans open-source project and contributes significantly to the OW2 JOnAS application server. An expert in OSGi and Java EE, he provides architectural guidance on major Bull projects. Member of the Java EE 6 expert group for EJB 3.1 specifications, Florent holds a Master’s in Computer Engineering from Joseph Fourier University, Grenoble. He speaks at open-source conferences like JavaOne and Solutions Linux. Alexis Gaches specializes in automating software development lifecycles. Joining the Jazz movement in 2008, he architects Jazz solutions for IBM Rational, collaborating with French enterprises on agile practices for application management.

Abstract

This article assesses Florent Benoit and Alexis Gaches’s overview of IBM’s Jazz platform, aimed at streamlining collaborative software development from requirements to deployment. It dissects tools for requirements management, architecture modeling, implementation, building, testing, and project oversight. Positioned as a response to fragmented processes, the analysis reviews integration mechanisms, open-source alignments, and deployment options. Through demonstrations, it evaluates benefits for agility, traceability, and efficiency, alongside implications for organizational adoption and tool interoperability in diverse environments.

Rationale and Architecture of Jazz Platform

Jazz addresses silos in development by promoting unified collaboration. Florent outlines its genesis: enhancing processes across lifecycle stages—requirements, design, coding, builds, tests, management. Core philosophy: Tools should interconnect, enabling traceability from user stories to code commits.

Architecture leverages Eclipse for IDE integration, with Rational Team Concert (RTC) as hub. RTC supports SCM, work items, builds via Jazz Team Server. Open Services for Lifecycle Collaboration (OSLC) standardizes integrations, allowing third-party tools like Jira to link.

Alexis emphasizes agility: Iterative planning, dashboards for metrics, reducing manual handoffs.

Key Tools and Functionalities

Requirements Composer manages specs, linking to work items. Quality Manager handles testing, integrating with RTC for defect tracking.

Implementation uses Eclipse with RTC plugins for code management, supporting SVN/Git via bridges. Builds automate via Ant/Jenkins, with traceability to changesets.

Demonstrations showcase scenarios: From story creation to code delivery, highlighting real-time updates and approvals.

Deployment options: On-premise or cloud (JazzHub), with free tiers for small teams/academia.

Integration with Open-Source and Legacy Systems

Jazz embraces open-source: Eclipse foundation, OSLC for extensibility. Migrations from ClearCase/SVN use connectors, preserving history.

Challenges: Cultural shifts toward transparency; tool learning curves. Benefits: Reduced cycle times, improved quality via automated traceability.

Future Directions and Community Engagement

IBM’s openness: Public development on jazz.net, inviting contributions. Academic JazzHub fosters education.

Implications: Enhances enterprise agility, but requires commitment. In global teams, it bridges geographies; for startups, free tools lower barriers.

Jazz exemplifies integrated ALM, driving efficient, collaborative delivery.

Links:

• Lecture video
• Alexis Gaches on LinkedIn

Case

My previous subject was the following: considering a WSDL that was generated by Axis 2, develop a webservice with CXF and/or SpringWS in “contract first” (ie the contract seen as an API is given, Java code comes afterwards) approach.

It cannot work easyly, unless you are ready to spend time, money and energy in creating adapters and adding layers in your application architecture.

CXF

With CXF, the first step is to generate Java code with the wsdl2java embeded in Maven CXF plugin. The output is explicit:

[java]Failed to execute goal org.apache.cxf:cxf-codegen-plugin:2.6.1:#wsdl2java on project PocEjb3: Rpc/encoded wsdls are not supported with #CXF[/java]

(cf my tweet)

RPC vs Encoded

Actually, the SOAP envelope has two ways to transport the message: RPC and Document. More detail is available here: Which style of WSDL should I use?
Axis 2 generates RPC/encoded webservice ; this method is deprecated, CXF does not support it.

Spring WS

I then tried to use Spring WS. I thought “Spring WS is only contract-first, it should work”. The first pitfall I fell on was that SpringWS bases on XSD files, not on actual WSDL. For instance, if you follow SpringWS tutorial, you will find such a block:

[xml]    <sws:dynamic-wsdl id="holiday" portTypeName="HumanResource" locationUri="/holidayService/"
targetNamespace="http://mycompany.com/hr/definitions">
<sws:xsd location="/WEB-INF/hr.xsd"/>
</sws:dynamic-wsdl>
[/xml]

To workaround, you can specify use tag, such as:

[xml]<sws:static-wsdl id="precomputed-holiday" location="/WEB-INF/precomputed-holiday.wsdl"/>[/xml]

Anyway, the next issue appears: SpringWS expects to deal with document webservices, ie blocks like:

[xml]<soapenv:Envelope xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/" xmlns:sch="http://mycompany.com/hr/schemas">
<soapenv:Header/>
<soapenv:Body>
<sch:HolidayRequest>
<!–You may enter the following 2 items in any order–>
<sch:Holiday>
<sch:StartDate>?</sch:StartDate>
<sch:EndDate>?</sch:EndDate>
</sch:Holiday>
<sch:Employee>
<sch:Number>?</sch:Number>
<sch:FirstName>?</sch:FirstName>
<sch:LastName>?</sch:LastName>
</sch:Employee>
</sch:HolidayRequest>
</soapenv:Body>
</soapenv:Envelope>>

[/xml]
whereas Axis-generated WSDL look like:
[xml]<soapenv:Envelope xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/" xmlns:int="http://interfaces.api.lalou.jonathan">
<soapenv:Header/>
<soapenv:Body>
<int:sayHello soapenv:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/"/>
</soapenv:Body>
</soapenv:Envelope>

[/xml]

In other words, I have just fallen on the exact same issue as with CXF: SpringWS cannot support RPC/encoded WSDLs.

Lecturer

Kito Mann leads as Principal Consultant at Virtua, Inc., focusing on enterprise architecture, training, and mentoring in JavaServer Faces (JSF), HTML5, portlets, Liferay, and Java EE. Editor-in-chief of JSFCentral.com, he co-hosts the Enterprise Java Newscast and hosts the JSF Podcast series. Author of “JavaServer Faces in Action” (Manning), Kito participates in JCP expert groups for CDI, JSF, and Portlets. An international speaker at events like JavaOne and JBoss World, he holds a BA in Computer Science from Johns Hopkins University.

Abstract

This article probes Kito Mann’s exploration of common pitfalls in JavaServer Faces (JSF) development, juxtaposed with recommended strategies for optimal performance and maintainability. It scrutinizes real-world antipatterns, from hardcoded IDs and database accesses in getters to broader issues like inconsistent standards and improper API usage. Embedded in JSF’s component-based framework, the analysis reviews techniques for dependency injection, state management, and view optimization. Via code illustrations and case studies, it evaluates consequences for scalability, team onboarding, and application longevity, advocating principled approaches to harness JSF’s strengths effectively.

Common Pitfalls in Component and Bean Management

JSF’s strength lies in its reusable components and managed beans, yet misuse breeds inefficiencies. Kito identifies hardcoding IDs in backing beans as a cardinal error—components autogenerate IDs, risking conflicts. Instead, employ bindings or relative references.

Database operations in getters exacerbate performance: invoked multiple times per request, they overload servers. Solution: Fetch data in lifecycle methods like init() or use lazy loading:

@PostConstruct
public void init() {
    users = userService.getUsers();
}

Lack of standards fragments codebases; enforce conventions for naming, structure. Wrong APIs, like FacesContext in non-UI layers, violate separation—inject via CDI.

Optimizing State and View Handling

State management plagues JSF: View-scoped beans persist unnecessarily if not destroyed properly. Kito advises @ViewScoped with careful serialization.

Large views bloat state; mitigate with for partial renders or for modularization. c:if toggles subtrees but beware quirks—prefer rendered attributes unless tree pruning is essential.

Dependency lookups in getters repeat calls; leverage CDI injection:

@Inject
private UserProvider userProvider;

This ensures singletons are fetched once, enhancing efficiency.

Enhancing Performance Through Best Practices

Ajax integrations demand caution: Overuse swells requests. Optimize with execute/render attributes.

Navigation rules clutter; use implicit navigation or bookmarkable views with GET parameters.

Testing antipatterns include neglecting UI tests—employ JSFUnit or Selenium for comprehensive coverage.

Implications: These practices yield responsive, scalable apps. By avoiding antipatterns, teams reduce debugging, easing onboarding. In enterprise contexts, they align JSF with modern demands like mobile responsiveness.

Kito’s insights empower developers to refine JSF usage, maximizing framework benefits.

Links:

• Lecture video
• Kito Mann on LinkedIn