Feature Descriptions (and samples)
The following brief feature descriptions describe the features in the table above, and provide links to live examples and documentation.
Client Features (Summary): SmartClient LGPL and Smart GWT include all in-browser components, including all UI components and all client-side databinding systems, with the sole exception of the UI components involved in the Analytics module (CubeGrid and Chart) and communications code designed specifically for interactions with the SmartClient Server Framework. A complete system ready for binding with any REST or WSDL services or with GWT-RPC.
Server Features (Summary): Server features are based on the SmartClient Java Server Framework, a set of standard Java .jars, servlets, jsp tags, tools and other resources that work with any standard J2SE or J2EE servlet engine or application server and with any standard Java IDE. The SmartClient Server Framework can be installed into existing projects and is compatible with popular Java frameworks such as Struts or Spring.
Calendar: Provides out of the box drag and drop calendaring interface similar to Google Calendar, including day/month/week views and automatic persistence to any data provider. Calendar can also visualize any dataset that contains date/time information. (Examples: SmartClient; SmartGWT.)
Pluggable Data Model: UI components support a "pluggable" data model that provides both a description of the data (fields, types, validation rules, etc) and operations to fetch and modify data. This "pluggable" data model approach means user interfaces can be re-used with any data provider without change. Screen designs such as master-detail navigation can be easily captured as custom components that are then usable with any data model — you can even swap the data model on the fly. (Examples: SmartClient; SmartGWT.)
Advanced Filtering: Allows end users to define advanced search criteria, such as fields being greater or less than target values or matching a pattern. Advanced mode provides power equivalent to a SQL "where clause". Searches are easy to save and restore.
Adaptive Filter & Sort: Combines large dataset handling with adaptive use of client-side filtering and sorting: to handle datasets that are too large to download completely in advance, filtering and sorting initially take place on the server. When all matching records have been loaded client-side, SmartClient automatically & transparently switches to client-side filtering and sorting. This provides immediate responsiveness and dramatically reduced server load. Works with any server platform; Java Servers with SmartClient Pro or better have built-in support for server-side filtering for SQL and Hibernate. (Adaptive Filtering examples: SmartClient; SmartGWT. Adaptive Sort examples: SmartClient; SmartGWT.)
Server Data Binding: Provides a simple Java API for handling requests for data from SmartClient visual components. Greatly accelerates integration with Java-based servers as compared to building REST or WSDL services for SmartClient components to consume. Client-to-server communications are handled automatically and transparently via a high speed, compressed protocol. Sophisticated client-server protocol handles data paging, server-side sort, automatic cache updates, validation and other error handling, optimistic concurrency (aka long transactions) and binary file uploads. Requests can be routed to Java classes via simple XML declarations, and required parameters to the method are fulfilled automatically, similar to Spring's "dependency injection". (SmartClient custom DataSource example here. DMI examples: SmartClient; SmartGWT.)
Server-side Validation: Both client and server validation are driven from a single XML declaration, with a built-in set of validators similar to XML Schema or JPA restrictions. (Examples: SmartClient; SmartGWT.) Custom client and server-side validators can be added, and a powerful expression language can turn complex cross-object validations into one-line declarations. (Examples: SmartClient; SmartGWT.) Server contact is automatically managed for validators that require it. (Examples: SmartClient; SmartGWT.)
Declarative Security: Attach role-based security declaratively to DataSource operations. UI components are security-aware and automatically disable UI for disallowed functions. Zero code integration with J2EE Security, Spring Security, or JAAS. Documentation can be found here and here
UI Generation: Automatic generation of grid, form, detail and tiled views from Java Beans or SQL tables. Unlike "scaffolding" approaches that leave you maintaining generated code, generation is dynamic, and you can customize any detail of the generated components without losing automatic generation. Without this capability, every screen of your application duplicates the information you already have in your server data model, leading to highly redundant, brittle applications. See the QuickStart Guide, Server Framework Chapter, first section.
UI Auto-mapping: Intelligent 2-way binding between persistent Java objects and UI components handles arbitrarily nested objects, Java Enums and Generics, avoids transmitting unused data to the browser without the need to write separate DTOs (Data Transfer Objects), and even allows bi-directional XPath-based mapping between UI fields and Java objects, when loading and saving data. (Data Selection example: SmartClient; SmartGWT. Saving nested objects example: SmartClient; SmartGWT.)
Basic Connectors (SQL, JPA & Hibernate): A one-line declaration enables all CRUD operations (create, retrieve, update, delete) against any SQL table, SQL View, JPA entity or Hibernate entity. No server code is required. Java business logic may be added to modify requests before they are executed by the connector, modify responses, or take any other action. For an overview, see the QuickStart Guide, Server Framework Chapter. Note: Basic connectors enable query-by-example search; for Advanced Search, purchase Power Edition or better.
- Hibernate Connector in action: SmartClient, SmartGWT examples.
- Auto-derived DataSources in SmartClient: Hibernate and JPA examples.
- Auto-derived DataSources in Smart GWT EE: Hibernate and JPA examples.
Queuing: Queuing allows multiple data load or save requests from different UI components to be transparently combined into a single HTTP request with guaranteed in-order execution. Queuing makes it easy to reuse data operations across different screens and in different combinations without needing to write screen-specific code to combine operations together. Queuing also makes it simple to define transaction boundaries when data in different UI components need to be saved together. (Examples: SmartClient; SmartGWT.)
HTTP Proxying: The HTTP Proxy allows an application to access to web services hosted on remote servers, which are normally inaccessible to web applications due to the same origin policy. Once installed, the HTTP Proxy provides transparent proxying of requests without further configuration; public sites may wish to whitelist specific services that may be accessed via proxy, for security reasons. (Examples: SmartClient; SmartGWT.)
Excel Export: Allows any grid component to export its current dataset as an Excel spreadsheet or CSV file without the need to write any server code. Exported data respects the user's current column order, sort direction, search criteria, and shown vs. hidden fields. Excel spreadsheets include correct data types (numbers, dates) and custom data hiliting rules. Works with any DataSource type — Hibernate, SQL, or custom. (Examples: SmartClient; SmartGWT.)
SQL/HQL Templating: Allows the SQL or HQL queries automatically generated by the SQL and Hibernate connectors to be easily customized directly in the DataSource definition, by overriding specific clauses from the automatically generated query, without the need to Java code. SQL Templating makes it trivial to use SQL joins while retaining built-in features like advanced search (SmartClient, SmartGWT examples). Even complex reporting queries can still leverage search, data paging, and other built-in features (SmartClient, SmartGWT examples).
Automatic Transactions: When using SQL and Hibernate connectors, any batch of save operations submitted to the server is automatically combined into a database transaction. Since UI interactions such as Mass Update and Multi-Row Drag and Drop automatically submit a batch of save operations to the server, these and many other UI operations are automatically transactional in Power Edition and above.
Batch Uploader: An end user interface for batch upload of data as a pre-built, customizable component. User can upload data, have it validated, make corrections, and save the data to a target DataSource. Supports a variety of upload formats and any data store that can be connected to SmartClient. (Examples: SmartClient; SmartGWT.)
- Background download of the SmartClient runtime and application resources while users are interacting with a login/registration page, splash screen or other non-SmartClient content
- Automatic compressed delivery of all of compressible content, including browser-specific workarounds where required.
Analytics Module: Specialized components for viewing, navigating and editing very large, multi-dimensional data sets (also known as data cubes) for business intelligence, analytics and reporting applications. (Examples: Basic Cube; Cube load on demand; Cube with charting.)
Real-time Messaging Module: Real-time data delivery via HTTP streaming (aka "Comet") for monitoring, command & control, trading desk and real-time collaboration applications involving time-sensitive, continuously updated data. SmartClient Messaging supports data delivery rates sufficient for continuous full-screen refresh and works through firewalls, HTTP proxies and other intervening network devices that can cause issues with other HTTP Streaming / Comet implementations. Supports codeless integration with JMS (Java Message Service) and a custom connector architecture for other message sources.
Deeper Discussion (on question marked items)
- Offline or intermittently connected operation is impossible. If the network is unavailable the application becomes inert.
- Common application scenarios in which logic must run for every mouse movement (custom drag and drop) or every keystroke (rapid data entry) cannot be achieved with acceptable user experience & server load.
- Because of frequent server trips needed to run application logic, the application will always feel sluggish relative to client-centric solutions like SmartClient & Smart GWT, reducing user productivity. This is especially true when latency is high (mobile) or the server is heavily loaded. While this problem can be partly mitigated in hand-crafted demos, in actual application development this problem gets worse as more customizations are added, requiring more and more frequent server contact.
- Scalability issues because of greatly increased server load, due to:
- many additional server requests
- server-side memory footprint from having the entire UI component model in server memory, which can hamper virtualization efforts
- replicating the server-side state between servers in a high-availability clustered environment (NOTE: sticky sessions doesn't solve this problem, as replication to at least one slave is still required to avoid data loss on server failure)
Because of these severe limitations, Isomorphic does not regard server-centric technologies as suitable choices for modern web applications. They are also extremely poor choices for enterprise-wide technology standardization: since there are so many scenarios in which an alternative technology must be used, the benefits of standardization (standard architecture, common tools, single set of skills to train and retain) are lost.
Server Data Binding: As compared to our completely free offerings, server-oriented technologies make certain trivial server binding scenarios easier, such as hardcoded / tiny datasets. However they do not handle advanced data binding scenarios with the simplicity available with SmartClient Pro. (Master-Detail examples: SmartClient; SmartGWT.)
Ext Price: Timely access to Ext patches requires a support subscription, otherwise, patch availability may be delayed by several months, as only major dot releases are available to licensees who have not have purchased support.
In contrast, patches to SmartClient and Smart GWT are immediately posted publicly, and licensed users are given special treatment on the SmartClient Forums.
For this reason Ext's combined license plus support price is more directly comparable to the license price alone for SmartClient / Smart GWT. Isomorphic then offers further support services, such as SLA-backed support plans with response time and resolution effort guarantees appropriate for mission-critical applications.
Vaadin pricing: Features provided for free in our open source product are commercial add-ons in Vaadin, including Automated Testing support ($1200/developer), mobile support ($750/developer), and a Calendar component ($150/developer).
Furthermore, by the time you have also added Vaadin's basic JPA support ($300/developer), you are now at price which exceeds our Power license with capabilities far less than our Pro license, and with the architectural drawbacks explained above.