XMLPatterns.com

Home

Introduction

Patterns
Categorized
Alphabetical
Chronological

Books

Links

About
Optional Container Element
Printable Version

Abstract:

When creating large DTDs with many logical units authors might be required to learn a large number of these logical units to know how to use the DTD. By hiding the details of optional parts of the DTD beneath optional elements, some of this complexity can be reduced.


Problem:

In a DTD with many logical units, an author of a document can be overwhelmed with the number of choices that have to be made.


Context:

In large, general-purpose DTDs where many logical units are presented, and more than one of the units is optional. The document will be authored by humans.


Forces:

In order to make DTDs applicable in many situations, large numbers of logical units need to appear in the DTD. However, presenting large numbers of logical units makes the DTD more difficult to learn by authors of documents.


Solution:

Hide related parts of a document that may never be needed by an author of a document beneath an element type. Any details of the structure of a document that are optional in a document can be grouped together as attributes and children elements of a single element. If an author never uses this branch of the document, none of the logical units within that branch need to be known to the other.


Examples:

Lets create a DTD that allows an author to create notes that can be a note to oneself, or an email. The DTD might look like this:
                            

<!ELEMENT Note (NoteToSelf | Email )>
<!ELEMENT NoteToSelf (Body)>
<!ELEMENT Body (#PCDATA)>
<!ELEMENT Email (Recipient+, CC*, BCC*, Body)>
<!ELEMENT Recipient (#PCDATA)>
<!ELEMENT CC (#PCDATA)>
<!ELEMENT BCC (#PCDATA)>

                            
                          
If the author of these documents only had to be create notes to himself, he would never need to be aware of the elements that are required in the Email branch of the DTD. He would create his document like this:
                            

<Note>
<NoteToSelf>
<Body>This is my note</Body>
</NoteToSelf>
<Note>

                            
                          
Although the DTD contains 7 elements, the author would only have to use three elements: Note, NoteToSelf and Email. The author would never be required to use the Email element and would not even have to be aware of the existence of the remaining three elements.


Discussion:

Once the Optional Group Pattern is applied, the learning requirements placed on the author of documents is reduced. Application of this pattern can however lead to an increase in the overall number of logical units in the DTD. The above note DTD could have been written as follows:
                            

<!ELEMENT Note (Recipient?, CC?, BCC?, Body)>
<!ELEMENT Body (#PCDATA)>
<!ELEMENT Recipient (#PCDATA)>
<!ELEMENT CC (#PCDATA)>
<!ELEMENT BCC (#PCDATA)>

                            
                          
This actually reduces the number of logical units in the DTD from 7 to 5, but makes things more complicated for users who just want to write notes to themselves, because they now have to be aware of the Recipient, CC, and BCC elements. The Optional Group hides complexity from the authors of documents by introducing optional elements that hide the details of a piece of the document that the author might not care about. By introducing a single higher level element, the details of many logical units may be hidden from the author, thereby reducing the learning requirements of the author.


Related Patterns:

This Optional Group is similar to the Choice Reducing Container in that they both hide logical units of a document from the author by introducing new element types. Existing elements are group together beneath the new element. The major difference between the two is the context they are used in. The Choice Group relates elements that are all choices in a content group while the Optional Group relates optional elements from the content group.


Known Uses:

This DTD fragment was taken from XBEL:
                            

<!ELEMENT xbel (title?, info?, desc?, (%nodes.mix;)*)>
<!ELEMENT info (metadata+)>
<!ELEMENT metadata EMPTY>
<!ATTLIST metadata owner CDATA #REQUIRED>

                            
                          
When generating an XBEL document, the info element is not required. This means that if the author was never interested in creating an info element, he would not need to know about the info element, the metadata element, or the owner attribute, thus eliminating 3 logical units from what the author needs to learn.

Contributions:

Thanks to Pankaj Kamthan from Concordia University for fixing up a bit of XML in the pattern.


Visit our sponsors: (Interested in advertising on XMLPatterns.com?)