Abstract:
When creating large DTDs, authors are required to learn a large number of attributes and elements, and where they can be positioned in order to know how to use the DTD. By reducing the number of choices that the author has to make at any point in the DTD by group related elements beneath newly introduced elements, the learning requirements of the author can be reduced.
Problem:
In a complex DTD with many elements and attributes, a user of a document can be overwhelmed with the number of choices that have to be made. With many options users have a difficult time knowing all of the elements available and the order that these appear in.
Context:
In large, general-purpose DTDs where many logical units are presented, and the documents are expected to be authored by humans.
Forces:
When the underlying order of the data to be represented, or to make DTDs applicable in many situations, large numbers of logical units need to appear in the DTD. However, presenting large numbers of attributes and elements makes the DTD more difficult to learn by authors of documents.
Solution:
Present only a limited number of choices to the author at any point within the document. If the author has the choice of insert 100 different elements at some point in the document, it can be difficult to know how to proceed. By grouping elements into related sets, the number choices to be made at any one point can be reduced. The author has to choose between only a few higher level groups instead of many more detailed elements.
Examples:
Here is a DTD that presents several choices to the author:
<!ELEMENT Doc
( Para
| OrderedList
| UnorderedList
| Figure
| Artwork )+>
The author here is present with a choice of 5 different elements to choose from after creating the Doc element. This choice could be limited by introducing new elements and grouping some of the existing elements together as children of the new elements like this:
<!ELEMENT Doc (Para | List | Illustration )+)>
<!ELEMENT List (OrderedList | UnorderedList )>
<!ELEMENT Illustration (Figure | Artwork )>
Discussion:
Once the Choice Reducing 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 element types in the DTD. The number of elements in the above example went from 6 to 8 when the Group Choice pattern was applied. This can make the overall complexity of the DTD higher. It can also increase the size of documents.
The Choice Reducing Pattern hides choices from the authors of documents by introducing new elements that hide the details of a piece of the document that the author might not care about at that point in time. By introducing a single higher level element, the number of choices the author needs to make is lower, thereby reducing the learning requirements of the author.
Related Patterns:
This Choice Reducing Container is similar to the
Optional Container Element in that they both try to make the DTD easier to use by the author by introducing new element types. Existing elements are grouped together as children of the new element. One difference between the two is the context they are used in. The Choice Reducing Container relates elements that are all choices in a content group while the
Optional Container Element relates optional elements from the content group.
The motivation for the two is also different. In the Choice Reducing Container, the amount of information at any one point is reduced, but it is not intended to reduce the overall complexity of the document for a single author. The
Optional Container Element hides parts of the structure permanently from the view of individual authors. It is likely that some authors will never use certain parts of large DTDs.
References:
See
Structuring XML Documents, Section 4.2.1 - Limiting Choices.