Object-Based Development
Spectra applications consist of ContentObjects and ContentObject Types. Object-Based Development is the process of creating and maintaining these application components. Object-Based Development is not just a methodology for developing application code; it's also a methodology for creating your data model and data storage.
The core of Object-Based Development is the ContentObject Type. The ContentObject Type defines what will be stored in each ContentObject (properties), what can be done with each ContentObject (methods), and how each possible action that can be performed on a ContentObject relates to the code you write (handlers). The Object-Based Development model can be much more flexible and maintainable than standard ColdFusion development techniques, as shown in Table 3.2. Object-Based Development also allows you to take full advantage of all the services Spectra provides.
Table 3.2 Differences Between ColdFusion and Spectra Applications
|
Task |
ColdFusion Approach |
Spectra Approach |
|
Rendering of the data model |
A database schema is created in an RDBMS |
Property definitions, ContentObject Types, and type properties are created. |
|
Providing data access |
SQL queries are generated and encapsulated in CFQUERY tags. |
Data access is provided by the system. |
|
Modifying the data model |
Changes are made to the database schema and the SQL queries. |
Type properties are modified as necessary. Complex changes can be made without database modifications. |
Differences from Object-Oriented Development
Although objects, properties, and methods are common terms in object-oriented programming, Object-Based Development has key differences from this model. There are no class libraries or inheritance, and object instances vary in their behavior and life cycle. Object-Based Development isn't meant as a replacement--it's an encapsulation of the best practices in developing application logic and presentation layers for Web applications. Object-oriented development (using Java or C++) complements object-based Spectra applications very well, and Spectra and ColdFusion both fully support integration.
The ContentObject API (COAPI)
The core of Spectra and Object-Based Development is the ContentObject API (COAPI). This set of tags and services comprises the data access and application organization of Spectra. The COAPI tags reside in the /Program Files/Allaire/Spectra/Customtags/COAPI/ directory as well as the /Program Files/Allaire/Spectra/Customtags/tier0/ directory. The COPAI directory contains system-level tags such as ContentObject storage and retrieval and Process Logic Path instantiation. The tier0 directory contains the tags you are most likely to use, such as those used in the creation of ContentObject Types and ContentObjects, meta data services, security, and other services.
The COAPI is also the point of integration with external systems such as user databases, external databases, and legacy applications. All other Spectra systems and services are built atop the COPAI and use the tags and services it provides.
The Application Applications are sets of resources available to your Spectra applications. The resources you define in an application are the Security Context, the Spectra Database (ContentObject Database), and the Logging Object:
The Security Context is the user databases, rules, and policies that form the basis of the security model for your application. This is covered in more detail in Chapter 16, "Using Spectra Security Options."
The ContentObject Database is where all your application data will reside, including ContentObject Types, Process Logic Paths, and Site Layout Model data. Spectra manages the data access for this database. You will create one for each of your applications, as described in Chapter 6, "Introduction to Spectra Programming."
The Logging Object defines what data from your application will be logged and where the log will reside. This is defined in the System Administration portion of the Webtop and is described in Chapter 17, "Site Reporting."
ContentObject Types ContentObject Types are the ContentObject definitions. ContentObject Types define what data is contained in ContentObjects of that type, as well as what actions can be performed on those ContentObjects. A ContentObject Type consists of a set of type properties, a set of methods, a handler root, and a label. The ContentObject Type becomes a template for all the ContentObjects of its type, otherwise known as its instances.
Properties Properties are the place where data is stored in a ContentObject. For instance, if you wanted to store and retrieve articles in your application, normally you would create a database table with columns for each component of an article. These components could be Tile, Body, and Author. In Spectra, the creation of a database table isn't necessary; your data is stored in the ContentObject. Three components make this possible: the Property Definition, the Type Property, and the Property itself.
NOTE
You aren't limited to data storage in properties. You can also store data in external databases. This is covered in Chapter 6. n
The Property Definition is a global definition of what the property will eventually be. Property Definitions contain a data type, a label, and validation information to be used by Spectra's default handlers. For the article example above, you would create Property Definitions for each component of the article. They would vary significantly from each other in their use in an application. A title, for example, is most likely a short string, whereas a body would probably be a very long set of text. In creating Property Definitions, you also consider how they will be used in all your ContentObject Types. You may have both articles and products as ContentObject Types, but both will have a body property. To enable your users to search both the body property of all the articles and the body property of all the products, you would create a Property Definition like the one shown in Table 3.3 and relate it with both ContentObject Types. This related Property Definition is called a Type Property.
Table 3.3 Example Property Definitions
|
Label |
Data Type |
Input Type |
Validate |
|
Title |
char |
text |
required |
|
Body |
longchar |
textarea |
none |
|
DueDate |
datetime |
text |
date |
|
OrderStatus |
char |
select(provide options) |
|
Each ContentObject Type contains a set of Type Properties. Each Type Property can be multilingual, can contain multiple values, and can be indexed by the Verity engine or stored in a normalized table. Type Properties can even contain other ContentObjects. Type Properties are limited only by the data type defined in the related Property Definition. The Type Property is a template for what can be stored in any ContentObject of that type. When a ContentObject of that type is created and the actual text of an article is stored with its title and author, a property is created.
Methods and Handlers All the actions that can be performed on a ContentObject are called methods. For the article example, you would create, edit, and display each ContentObject. You would create a method for each action. Each method has a label and an associated handler. Handlers are ColdFusion templates that you create. So, to allow users to edit your ContentObjects, you code a handler that contains a form for modifying the ContentObject and an action to persist the changes.
The system provides special handlers to enable you to build your application more quickly and also to take care of common tasks. The most commonly encountered are the create, edit, and display handlers. If you don't create named methods for these, Spectra will create default handlers. Once you create named methods for these tasks, your new handlers will override the default system handlers. One exception to this is the create handler; if you create a named method for this, Spectra will execute your code sequentially after the system create handler (the constructor). This allows you to perform complex initialization of a ContentObject if you need to. Table 3.4 lists the default handlers in Spectra.
Table 3.4 Spectra's Default Handlers
|
Handler |
Function |
|
create |
Creates a new ContentObject. Can't be overridden. |
|
edit |
Renders a form for editing any ContentObject. |
|
delete |
Deletes a ContentObject. |
|
display |
Displays ContentObject properties and values. |
|
editattributes |
Allows for changing the Active and Published status for a ContentObject. |
|
editmetadata |
Allows categorization of a ContentObject. |
|
editsecurity |
Manages the permissions for all methods for a ContentObject. |
|
get |
Retrieves a ContentObject. |
|
preview |
Previews a ContentObject. |
UML and Documentation After you create your ContentObject Type, you can use several strategies for documentation. UML provides a good set of tools for documenting your ContentObject Types because of their similarity to objects in object-oriented languages. However, if you are unfamiliar with UML, you can use techniques derived from UML, as shown in Table 3.5.
NOTE
UML is the acronym for the Unified Modeling Language,an industry-standard language for documenting and visualizing software systems. A good primer on UML is Sams Teach Yourself UML in 24 Hours (ISBN 0-6723-1636-6).
Table 3.5 Example of a Documented Object
|
Object |
Type Properties |
Methods |
|
Article |
|
|
|
|
Title |
displayTeaser |
|
|
Body |
displayFull |
|
|
Author |
|
|
|
Byline |
|
ContentObjects
ContentObjects are instances of ContentObject Types. In object-based development, ContentObjects are really sets of data that have been constrained by a ContentObject Type. These data sets can then be acted on or invoked by using the defined methods of the ContentObject Type. Put simply, ContentObjects are to ContentObject Types as rows are to tables in a database.
Invocation Passing the data from a ContentObject to a method is the process of invocation. Invocation occurs throughout the lifecycle of a ContentObject. Any time you want to do something with a ContentObject, you invoke it with a method, as shown in Figure 3.1. If you want to edit ContentObject x, you would invoke object x by using the edit method. The data contained in the ContentObject would be passed to the handler you associated with the edit method (or the default handler), and the result would be output to the page.
){kind=link}
Figure 3.1 Object invocation consists of four steps.
Remote Invocation ContentObjects can also be invoked remotely, allowing interaction with other Spectra servers or applications. In fact, any application on virtually any platform can invoke a ContentObject remotely. This is the basis of Application Syndication, which allows Spectra application functionality to be shared securely with other Web applications.
Embedded Objects A Property Definition can have a data type that's actually a ContentObject Type. When this Property Definition is related to a ContentObject Type, any ContentObjects of this type will contain an Embedded Object. One property will fully contain another object. In our article example above, you could now place the author's name, email address, and biography in each article, all contained within a single property.
Embedded Objects can be shared or standalone. Shared Embedded Objects are objects that exist on their own and are embedded into multiple ContentObjects. When the master object is updated, all embedded versions are also updated. Standalone Embedded Objects exist only within a single ContentObject and must be accessed through the embedding object.
TIP
For both shared and standalone Embedded Objects, the Embedded Objects have an objectID or unique identifier of {embedding objectid}.{property}.